Home

MSc Thesis: The Design, Implementation and Analysis of a

1. D dt PONS ur ax 0 5 ee dy 0 5 EE az 0 5 T p 0 5 ree q 0 5 ee r 0 5 ei T 3 deges Although it is not expected that this configuration will be best suited for RMC s 3DOF it serves as a Starting point from which more appropriate coefficients will eventually be derived All three flight profiles were played back as per the experimental procedure To begin this test run the author sat in the simulator and flew several flight manoeuvres to collect subjective information on the performance of the configuration selected Overall the motion simulation for this test run was felt to be clearly inadequate and even detrimental to the overall flight simulation Pilot observations for this test run are summarized in Table 17 Table 18 shows the response variables obtained for each run It is important to note that not all response variables provide relevant information for each flight profile Indeed depending on the manoeuvre some motion channels are not significantly stimulated As such a color code is used to identify response variables with a high relevance to the study of each motion channel Table 17 Pilot Observations Test Run 1 Motion Channel Observations Large heave overshoot are present after a heave motion i e After a heave up the Translational platform will travel down forcefully without any reason Tilt Coordination Tilt Coordination is nearly non existent
2. Request a subscribtion to a simulation start event j hr SimConnect_SubscribeToSystemEvent hSimConnect EVENT_SIM_PAUSE Pause hr SimConnect_RequestDataOnSimObject hSimConnect REQUEST_PDR DEFINITION_PDR SIMCONNECT_OBJECT_ID_USER SIMCONNECT_PERIOD_VISUAL_FRAME SIMCONNECT_DATA_REQUEST_FLAG_CHANGED SIMCONNECT_DATA_REQUEST_FLAG_TAGGED HI if FSXConnectRequest pe Tn the FSXDisconnect method the Simconnect connection is terminated void FSXConnect FS XDisconnect HRESULT hr Handle return object hr SimConnect_Close hSimConnect close the connection to SimConnect connected false Set the connection status flag to false FSXDisconnectRequest Tn the simulation method a call is made to SimConnect Tt is important to add this as the last parameter in order to receive it the callback This method returns the state of the connection void FSXConnect simulation if quit SimConnect_CallDispatch hSimConnect MyDispatchProcPDR this Simulation step HI if else if quit connected false else simulation px The isConnected method returns the status of the FSX connection return amp lt true amp gt if FSXConnect is connected to FSX amp lt false amp gt if there is no connection bool FSXConnect isConnected return connected isConnected Mainwindow h pee Mainwindow h IDE Visual C 2008 Express
3. Table 35 Complete Impact and Training Potential Ratings Pilot Motion Simulation Impact Flight Training Potential 1 1 75 1 2 2 2 3 2 1 4 1 5 2 5 1 5 2 6 1 1 7 1 1 8 2 2 9 1 5 1 5 10 1 1 5 11 2 1 12 1 1 Average 1 52 1 42 Deviation 0 43 0 47 Ratings from 1 to 5 where 1 is excellent and where 5 is very poor 175 Annex J Ethics Board Documentation 176 177 The conduct of a pilot validation exercise with human participation required the authorization from RMC s Research Ethics board In this annex the submission for ethics review the informed consent letter and final authorization letter are included Se Si x y SUBMISSION FOR ETHICS REVIEW OF Rs RESEARCH INVOLVING HUMAN SUBJECTS REMINDER If you are conducting research in an institution agency other than RMC you may be required to get ethical approval from other boards before proceeding with your research Approval may also be required for the area in which the research will occur e g Nunavut Indian reserves international research Please ensure that you have all required approval prior to conducting your project Investigators and co investigators must ensure that all personnel involved in their research projects are aware of and adhere to the ethical principles outlined in the Tri Council Policy statement For bilingual projects all relevant documentation must be
4. Turn Profile Test Run 1 Overshoot ja Overshoot 3 10 15 20 25 30 Time s Figure 65 p Angular Rates Turn Profile vs Test Run 1 5 3 3 Solving the Problem of 2 Order Filter Overshooting Through testing of different filter dampening ratio and scaling coefficients with 2 order high pass filter equations it was found that overshooting can never be totally eliminated What is specifically referred to 2 order filter overshooting can be seen in Figure 66 where unfiltered and filtered signal are compared and where overshooting is present on almost all filtered signal peaks At first a method using a switch based on the derivative of unfiltered signals as a criterion to zero filtered signals was investigated The first approach consisted in monitoring the unfiltered signal derivative as criteria to zero the filtered signal The hypothesis is that when the unfiltered signal derivative is within a certain interval of zero no filtered signals should be let through This makes sense because signals with a very shallow slope with a derivative 0 are of too 120 low frequency to be simulated and represent areas where there should be no motion in the filtered signal a m s Unfiltered Filtered 0 5 10 15 20 25 30 Time s Figure 66 2 Order High Pass Filter Overshooting Thus where shallow unfiltered slopes are present there should be less or no overshoot allowed After a series of t
5. A 3 5 N a 4 N P 2 a amp a en O 5 10 lt 12 0 5 10 15 20 25 30 Time s Figure 57 Plot of Climb Profile Motions against Time 105 The difference between motions signals for each flight profiles can be seen in Figure 57 Figure 58 and Figure 59 Q Aircraft Motions m s or deg s 0 5 10 15 20 25 30 Time s Figure 58 Plot of Standard Rate Turn Profile Motions against Time As a Starting point each motion channel is first configured using the optimal scaling factors and filtering equations Table 2 as reported by 12 For each test run the recorded flights are played back through a modified version of the classical WF algorithm and the motion response of RMC s 3DOF motion platform is recorded using an inertial measurement unit More information on the flight recording and playback modifications of the Simulink classical WF algorithm can be found in Annex G Following a sequential approach after each test run the performance of each channel is analyzed based on pilot feedback motion plots and the response variables representing the mean absolute error plus two standard deviations The motion channels parameters are then modified in order to solve problems and increase the motion 106 simulation fidelity There are no hard limits on the number of tests runs required to achieve the EN best performance 2 Q 0 2 Ss 2 by
6. Os By HP Bsu i Bsu tx Bs a t s T X L L SCALE FILTER a b I IEEE AEEA OEE eek ee AE A Ts Ts Ts Figure 4 Classical WF Algorithm 6 In this study aircraft accelerations are expressed using the metric system in mes and the angular rates are expressed in deges The outputs of the classical WF algorithm are the 16 simulator displacement S where S d d d and the simulator Euler angles Bs where Bs 6 p In this formulation dx dy and dz are the displacement of the motion platform for each axis The variables 0 and w represent the motion platform Euler angles with regards to the earth reference frame F7 When combining S and Bs it is possible to obtain the final position of the motion platform with respect to F From this information we can derive the appropriate actuator lengths speeds and accelerations required to achieve this position In the sections to follow each channel and its respective signals of the algorithm are discussed in details Finally the term Box followed by a number used in the following sections refers to the corresponding function block of the classical WF algorithm as illustrated in Figure 4 2 3 1 Translational Motion Channel The translational motion channel consisting of Boxes one to four of Figure 4 transforms the aircraft accelerations Fa into simulator translational displacement S In Box one accelerations are first scaled do
7. Pointer to a Galil object for RMC s 3DOF Galil g px String used to differentiate between 3 and 6DOF string platformType pee The speed of the actuators for user inputs 1 string speed ex String containing the controller ip adress string ipAddress Ja String array for the lengths string 1 3 ex String array for the velocities string v 3 pee String array for the accelerations string a 3 pee Value containing the number of actuators int number_of_actuator pee Value containing the conversion factor from metric to 3 6DOF 210 int conversion_ factor pee Flag that is used to prevent any motion when set to amp lt true amp gt bool prevent_movement string int2string int i string make3DOFCommand string prefix double v endif MODEL_H Model cpp pe Model cpp IDE Visual C 2008 Express Edition author 2Lt Jean Le Bouthillier Latest Revision 11 May 2010 ek For the Galil C API http www galilmc com support manuals galiltools library html library Version ek Description This class is responsible for interacting with RMC s 3DOF or 6DOF motion platforms AJl motion platform specific operations are happening in this class This model is able to detect which motion platform it is connected to and self configure appropriately Other classes can ask this model to process pre programmed motions i e park m1 up
8. The request is passed to the mediator in this method eh void MainWindow on_pitchBackward_released Feedback m gt execute stopMovement Send request to mediator on_pitchBackward_released pe Event handler for the Roll Left button As long as this button is pressed the platform will roll left The request is passed to the mediator in this method al void MainWindow on_rollLeft_pressed Feedback m gt execute rollLeft Send request to mediator on_rollLeft_pressed px Event handler for the Roll Left button When the button is released the motion stops The request is passed to the mediator in this method void MainWindow on_rollLeft_released Feedback m gt execute stopMovement Send request to mediator on rollLeft_released pe Event handler for the Roll Right button As long as this button is pressed the platform will roll left The request is passed to the mediator in this method af 201 void MainWindow on_rollRight_pressed Feedback m gt execute rollRight Send request to mediator on_rollRight_pressed ee Event handler for the Roll Right button When the button is released the motion stops The request is passed to the mediator in this method void MainWindow on_rollRight_released Feedback m gt execute stopMovement Send request to mediator on_rollRight_released pe Event handler
9. on_Extend_clicked pe Event handler for the Extend button The request is passed to the mediator in this method void MainWindow on_center_clicked Feedback m gt execute center Send request to mediator 199 on_center_clicked pe Event handler for the Park button The request is passed to the mediator in this method ef void MainWindow on_park_clicked Feedback m gt execute park Send request to mediator on_park_clicked px This method enables of disables the manual motion platform controls as warranted by the state of execution of the program gt i e the manual controls are disabled during flight simulation void Main Window enableManualControls bool b if b true If asked to enable manual controls ui gt speedSlider gt setEnabled true ui gt pitchForward gt setEnabled true ui gt pitchBackward gt setEnabled true ui gt rollLeft gt setEnabled true ui gt rollRight gt setEnabled true ui gt heaveUp gt setEnabled true ui gt heaveDown gt setEnabled true ui gt mlup gt setEnabled true ui gt m1down gt setEnabled true ui gt m2up gt setEnabled true ui gt m2down gt setEnabled true ui gt m3up gt setEnabled true ui gt m3down gt setEnabled true ui gt extend gt setEnabled true ui gt center gt setEnabled true ui gt park gt setEnabled true ui gt erasePosition gt setEnabled true ui gt center gt setEnable
10. speed speed Actuator 1 up actuators 2 and 3 down g gt command BGX YZ Begin motion on all actuators feedback Pitching forward Feedback message else if else if uInput rollLeft g gt command JG speed speed Actuator 2 up and actuator 3 down g gt command BGYZ Begin motion on actuators 2 and 3 feedback Rolling left Feedback message else if else if uInput rollRight g gt command JG speed speed Actuator 2 down and actuator 3 up g gt command BGYZ Begin motion on actuators 2 and 3 feedback Rolling right Feedback message W else if else if uInput heaveUp g gt command JG speed speed speed All actuator up g gt command BGX YZ Begin motion on all actuators feedback Heaving up Feedback message else if else if uInput heaveDown g gt command JG speed speed speed All actuator down g gt command BGXYZ Begin motion on all actuators feedback Heaving down Feedback message else if else if uInput actuator Up g gt command JG speed Actuator 1 up g gt command BGX Begin motion on actuator 1 feedback Actuator 1 moving up Feedback message W else if else if uInput actuator Down g gt command JG speed Actuator 1 down g gt command BGX Begin motion on actua
11. 0 0724 Turn Descent 0 0220 0 0002 0 0546 In light of the results collected during this test run it is felt that the system reached an optimal level of tuning and that the experimental study of the system reached a point where additional iteration of the experimental method would not yield substantial improvements In addition as pilot observations indicated that all motion channels create a realistic and proportional sensation there is no reason to modify any channels further 5 3 5 Conclusion Several problems were identified during the experimental test runs The combined method of pilot observation response variable measurement and plot investigation used in this study was effective at identifying the problems of the classical WF algorithm In most cases solutions were quickly found to correct pilot complaints The experimental test runs led to the modification of the classical WF algorithm in two areas First a switching mechanism based on the polarity of unfiltered accelerations was implemented to manage 2 order filter overshooting The results of this implementation had a positive effect on the overall motion simulation Next a 133 gain coefficient was implemented to solve the problem of an underpowered tilt coordination This implementation also led to a positive effect on the overall motion simulation An iterative process was followed to fine tune the classical WF algorithm and led to a
12. 134 5 4 1 Procedure Several pilots with various degrees of experience were approached with regards to the flight simulator and their possible participation in a pilot rating exercise A procedure for the conduct of the pilot validation exercise was developed and will be discussed in this section The protocol and nature of this validation exercise were approved by the Research Ethics Board of RMC An informed consent to proceed with the validation exercise was collected from each participant The Research Ethics Board proposal and approved protocol information can be found in Annex J All volunteer pilots were first introduced to the flight simulator through a short briefing highlighting the main features of RMC s 3DOF flight simulator Then the conduct of the pilot validation exercise was discussed Table 26 Pilot Validation Exercise Procedure Step Description Participant arrival Introduction to the flight simulator and 1 re introduction to the pilot validation procedure 2 Briefing Explanation of the rating questionnaire as well the manoeuvres used for rating 3 Initial run The participant performs a circuit without motion to get used to the controls the airplane and the manoeuvres 4 Validation run The participant performs the same circuit with full motion to evaluate the flight simulator 5 Rating The participant fills the questionnaire to rate the flight simulator 6 Debriefing The particip
13. pe Event handler for the Yaw right button As long as this button is pressed the platform will yaw right The request is passed to the mediator in this method void MainWindow on_yawRight_pressed Feedback m gt execute yawRight Send request to mediator on_yawRight_pressed px 203 Event handler for the Yaw right button When the button is released the motion stops The request is passed to the mediator in this method fal void MainWindow on_yawRight_released Feedback m gt execute stopMovement Send request to mediator on_yawRight_released pe Event handler for the Motor 1 Up button As long as this button is pressed actuator will move up The request is passed to the mediator in this method void MainWindow on_mlup_pressed Feedback m gt execute actuator Up Send request to mediator on_mlup_pressed px Event handler for the Motor 1 Up button When the button is released the actuator stops The request is passed to the mediator in this method idl void MainWindow on_mlup_released Feedback m gt execute stopMovement Send request to mediator on_m1up_released pe Event handler for the Motor 1 Down button As long as this button is pressed actuator will move down The request is passed to the mediator in this method af void MainWindow on_m1down_pressed F
14. while checkConnections pe The Feedback method is used to quickly append a text to the user feedback window The method accepts an std string and performs the change to QString ek 197 kk ai param s the text to be appended to the user feedback window void MainWindow Feedback string s if feedback true Check if the feedback option is enabled ui gt feedback gt appendPlainText S2Q s n Carriage return and append new text Feedback kk Event handler for the Start FSX Connection button The mediator attempts a connection to FSX when the user clicks the button The connection request is passed to the mediator in this method void MainWindow on_FSXConnect_clicked m gt execute FSXConnect Feedback Attempting connection to FSX checkConnections Check connections on_FSXConnect_clicked ee Event handler for the Stop FSX Connection button The mediator disconnects from FSX when the user presses the button The disconnect request is passed to the mediator in this method void MainWindow on_FSXDisconnect_clicked m gt execute FSXDisconnect Feedback Disconnecting from FSX Feedback message checkConnections Check connections on_FSXDisconnect_clicked px Event handler for the Begin Motion Simulation button When the user presses this button the motion simulation begins As suc
15. 21 In this study it was decided to use the visual frame rate as the polling frequency for the data The reason for this is that it provides better control due to the synchronization of visual frames with motion frames Indeed it is not possible to control the exact frequency when using simulation steps and the Hz frequency is simply too low to produce a high quality motion cueing Table 7 Information Requested from FSX Using FSXConnect Name Type Function Unit ACCELERATION BODY X Data x axis acceleration mes ACCELERATION BODY Y Data y axis acceleration mes ACCELERATION BODY Z Data z axis acceleration mes ROTATION VELOCITY BODY X Data p angular velocity deges ROTATION VELOCITY BODY Y Data q angular velocity deges ROTATION VELOCITY BODY Z Data r angular velocity deges PLANE PITCH DEGREES Data Plane pitch degree deg PLANE BANK DEGREES Data Plane bank degree deg EVENT_SIM_ PAUSE Event Pause event lorO When using the visual frame rate there is always and exactly one motion command sent per visual frame Moreover the visual frame rate also allows for the simple modification of the motion simulation rate by modifying the visual frame rate directly through the in game FSX settings The information requested from FSX through FSXConnect cpp can be seen in Table 7 Essentially the bulk of the data requested from FSX is the data required by the classical WF algorithm translational
16. lengths m number_of_actuator platformType udpAccelerations udpLengths gt udpVelocities gt velocities Mediator change_speed int void connect string string execute string string isFSXConnected bool isReceivingUDP bool simulation void 156 GalilPrivate A Galil d timeout ms conversion_factor 9 ipAddress number_of_actuator platformType prevent_movement speed v acceleration_command double void change_speed int void connect string string int2string int string is_platform_moving bool length_command double void make3DOFCommand string double string process_user_input string string velocity_command double void Galil const std string amp addresses static std ve arrayDownload const std vector lt double gt amp const std string amp void arrayDownloadFile const std string amp void arrayUpload const std string amp std vector lt double gt arrayUploadFile const std string amp const std string amp void command const std string amp const std string amp const std string amp bool ste commandValue const std string amp double connection std string firmwareDownloadFile
17. mainwindow cpp IDE Visual C 2008 Express Edition author 2Lt Jean Le Bouthillier Latest Revision 07 May 2010 ek Description This class contains the necessary code to implement the graphical user interface GUI that is used by the motion platform user It includes GUI object listeners as well as some convenience methods to facilitate the operation of the GUI When an event occurs on the GUI the appropriate actions are taken and most of time the Mediator is made aware of the user input ek 3 include mainwindow h include ui_mainwindow h pe MainWindow Constructor ek param parent pointer to the QWidget parent MainWindow MainWindow QWidget parent QMainWindow parent ui new Ui MainWindow feedback true Feedback on ui gt setupUi this m new Mediator Initialize the Mediator motionSimulation false Set the motion simulation flag to false MainWindow pe MainWindow Destructor 2 MainWindow MainWindow delete ui m gt execute park Send request to mediator MainWindow ee The Q2S method is used to change a QString to a std string ek param qs the QString to be changed into a std string return the std string obtained from the QString string MainWindow Q2S QString qs return string const char qs toLocal8Bit Q2S The S2Q method is used to change a std string to a QString ek param s the String to
18. E INFORMED CONSENT PROCESS 1 a If participants are to be given a Letter of Information and asked to sign a Written Consent attach copies b If free and informed consent is to be sought in some way other than in writing indicate how and justify DT Information letter with verbal consent attach a copy Annex C C Information cover letter attach a copy Cl Other Specify 184 Justification A cover letter will be included with the questionnaire which is reviewed by the participants prior to the conduct of any flight simulation As the study does not collect any personal information on the participants a signature is not desired Proceeding with the flight testing and answering the questionnaire indicates consent from the participants c If consent will not be obtained from potential participants provide a justification n a 2 a If this study involves persons who are not in a position to give their valid consent to participate attach a copy of the Letter of Information and Permission Form to be used to obtain permission from those with legal authority to give it If voluntary assent at the time of participation is not a precondition for the incompetent individual s participation provide a justification n a b If this study involves persons for whom consent to participate in the research project is not appropriate in the context of their cultural background indicate how their
19. F S 3 f me o L 3 x 0 F ff 4 f L Input KL Filtered 0 5 T i T i i i 0 2 4 6 8 10 12 14 16 18 20 Time sec Figure 14 Tilt Coordination Low Pass Filter x Component Sample Run 30 As can be seen in the gain magnitude plot as the signal s frequency increases a negative gain 1s applied to reduce its amplitude As can be seen from Figure 14 where an aggressive run with low frequency components is conducted the filtered accelerations remain very close to the unfiltered accelerations The parameters of the filter can be adjusted to increase the attenuation of the signal as required During the tuning phase described in Chapter 5 filters are tuned in conjunction with the rate limiter of the tilt coordination channel to achieve the desired effect Although a slightly higher filter break frequency wp 8 rad sec is used for the y component low pass filter it exhibits a behaviour very close to that of the x component filter the signal simply begins to be attenuated at a slightly higher frequency For this reason only the x component low pass filter was discussed in this section 2 5 Summary This chapter provided insight into the reference frames and axis system used throughout this study The classical WF algorithm as reported by Reid and Nahon 8 9 10 has been reviewed with an emphasis on each individual channels The limitations of RMC s 3DOF motion platform were discussed along with the necessa
20. First the connection management functionalities were thoroughly tested and some errors were found under special circumstances when the motion platform controller was energized but not the motion platform These issues were resolved by introducing fail safes in the code Then the manual controls were tested to ensure that they were fully functional and that the motion platform was responding as intended During this test it was found that some commands were inverted and that under specific conditions the software would crash Indeed when trying to press several manual motion command buttons in a rapid succession the software would crash 89 This discovery led to several improvements in the manual command handling code as well as the introduction of a safety feature that prevents two manual commands from being executed at the same time Finally the flight simulation component was thoroughly tested in combination with the manual control features to ensure that the software would comply with all user requests Only minor issues were found during this phase and were easily corrected in the code 4 3 2 User Validation After having thoroughly tested each of the software s functionalities a group of three volunteers were asked to test the software in an attempt to discover undiscovered bugs or problems This additional testing phase the user validation phase is necessary to ensure that no bugs were overlooked by the author Testing is a destruc
21. Surging forward Feedback message else if else if uInput surgeBackward DOF CODE HERE feedback Surging backward Feedback message W else if else if uInput actuator1Up 6DOF CODE HERE feedback Actuator 1 moving up Feedback message else if else if uInput actuator Down 6DOF CODE HERE feedback Actuator 1 moving down Feedback message else else if uInput actuator2Up DOF CODE HERE feedback Actuator 2 moving up Feedback message else if else if uInput actuator2Down 6DOF CODE HERE feedback Actuator 2 moving down Feedback message Jelse if else if uInput actuator3Up 6DOF CODE HERE feedback Actuator 3 moving up Feedback message W else if else if uInput actuator3Down DOF CODE HERE feedback Actuator 3 moving down Feedback message else if else if uInput actuator4Up 6DOF CODE HERE feedback Actuator 4 moving up Feedback message else if else if uInput actuator4Down DOF CODE HERE feedback Actuator 4 moving down Feedback message else if else if uInput actuator5Up DOF CODE HERE feedback Actuator 5 moving up Feedback message else if 216 else if uInput actuatorSDown 6DOF CODE HERE feedback Actuator 5 moving down Feedback message else if else if uInput actuator6Up
22. a 5 a i a a zZ 3 6 A F amr 3 O 1 lt en N Time s Figure 59 Plot of Descent Profile Motions against Time Once the experimenter has optimized the scaling and filtering blocks and when gains on performance are deemed no longer possible the classical WF algorithm is considered to be optimally tuned After having studied all channels and achieved optimal tuning for RMC s 3DOF motion platform the system is evaluated by pilots to collect additional information on performance 5 2 4 Experimental Procedure To protect the experimental validity of results between each test run it is important to follow a rigorous and exact protocol each time In order to determine the best coefficients for the classical WF algorithm the experimenter must be certain that the results observed between each run depends only on the modification applied to the system and are not the result of an 107 experimental error precise experimental procedure is also important to ensure the reproducibility of an experiment by others Table 13 shows a step by step list of the experimental procedure used for each test runs Table 13 Experimental Procedure for Each Test Runs Step Description Configure the Classical WF algorithm with the scaling and filtering coefficients under study Ensure that all equipment motion platform inertial 2 measurement unit is energized and ready for the test r
23. button ui gt FSXDisconnect gt setDisabled true Disable the Stop FSX conenction button ui gt SimulinkConnectLabel gt setText Simulink Connection Disconnected ui gt FSXConnectLabel gt setText FSX Connection Disconnected else if else if m gt isReceivingUDPO amp amp m gt isFSXConnected ui gt begin_button gt setDisabled true Disable the Begin Motion Simulation button ui gt stop_button gt setDisabled true Disable the Stop Motion Simulation button ui gt FSXConnect gt setEnabled true Enable the Start FSX connection button ui gt FSXDisconnect gt setDisabled true Disable the Stop FSX conenction button ui gt SimulinkConnectLabel gt setText Simulink Connection Established ui gt FSXConnectLabel gt setText FSX Connection Disconnected else if else if m gt isReceivingUDP amp amp m gt isFSXConnected ui gt begin_button gt setDisabled true Disable the Begin Motion Simulation button ui gt stop_button gt setDisabled true Disable the Stop Motion Simulation button ui gt FSXConnect gt setDisabled true Enable the Start FSX connection button ui gt FSXDisconnect gt setEnabled true Disable the Stop FSX conenction button ui gt SimulinkConnectLabel gt setText Simulink Connection Disconnected ui gt FSXConnectLabel gt setText FSX Connection Established else if QApplication processEvents Allow the GUI to process events
24. fastsmooth Dataset signals values 1 len 3 10 Dataset signals values 1 10 3 head Dataset signals values len 10 len 3 tail Smooth the remaining Angular rate signals p head mean Dataset signals values 1 10 4 tail mean Dataset signals values len 10 len 4 Dataset signals values 1 len 4 fastsmooth Dataset signals values 1 len 4 10 Dataset signals values 1 10 4 head Dataset signals values len 10 len 4 tail q 159 head mean Dataset signals values 1 10 5 tail mean Dataset signals values len 10 len 5 Dataset signals values 1 len 5 fastsmooth Dataset signals values 1 len 5 10 Dataset signals values 1 10 5 head Dataset signals values len 10 len 5 tail f head mean Dataset signals values 1 10 6 tail mean Dataset signals values len 10 len 6 Dataset signals values 1 len 6 fastsmooth Dataset signals values 1 len 6 10 Dataset signals values 1 10 6 head Dataset signals values len 10 len 6 tail Now the Response variables are calculated for each motion channels RV mean abs FlightProfile signals values 1 len 1 6 Dataset signals values 1 len 1 6 Now calculate the standard deviation for each response variables STD std abs FlightProfile signals values 1 len 1 6 Dataset signals values 1 len 1 6 RV RV 2 STD A plot of reference flight profile and dataset is printed plot Dataset time Dataset signals values 1 l
25. gt mSocket Closes the socket this gt mSocket INVALID_SOCKET HI if SocketCleanup pe This method is used to create a network socket and initialize the UDP networking param nDestlp a pointer to the first character of the string containing the destination ip adress bool UdpSend Initialize const char nDestIp if this gt mSocket INVALID_SOCKET Check for invalid state return false Return a failure Mif if nDestIp NULL Make sure there is an destination ip adress return false Return failure Mif ex Creating a network socket AF INET We are using IP Internet family of protocols SOCK_DGRAM We are using datagram style communication IPPROTO_UDP We are using UDP protocol this gt mSocket socket AF_INET SOCK_DGRAM IPPROTO_UDP if this gt mSocket INVALID_SOCKET Check if socket was successfully created return false Return failure Mif ex Create destination address structure ZeroMemory amp this gt mDestAddr sizeof this gt mDestAddr Initialize mDestAddr stucture this gt mDestAddr sin_family AF_INET IP family of address this gt mDestAddr sin_port 0 Initialize destination port with 0 It will be set later in SendDatagram this gt mDestAddr sin_addr s_addr inet_addr nDestlp Convert string IP address to numeric representation if this gt mDestAddr sin_addr s_addr INADDR_NONE Check if IP string conversion was suc
26. speed 6750 Initialize the speed at 50 for 3DOF prevent_movement false Initialize the prevent movement flag to false Mif else if platformType 6DOF number_of_actuator 6 Setting number of actuators 211 conversion_factor INSERT 6DOF CONVERSION FACTOR HERE WI if return platformType The connection information is returned connect px Used to change the speed of the platform used for pre programmed motions 4 param s the new speed a percentage from to 100 void Model change_speed int s if platformType 3DOF s 135 The maximal speed is 13500 100 speed int2string s Set the motion platform speed string variable Mif else if platformType 6DOF 6DOF CODE HERE speed int2string s else if change_speed Used to transform an integer into a string 4 param i the integer to be transformed into a string return the string version of the integer string Model int2string int i ostringstream ss ss str ss amp lt amp lt i return ss str int2string Takes a prefix i e PA AC SP and an array of 3 doubles to generate a properly formatted 3DOF Galil command ek param prefix a string containing the command prefix param v a double array containing the three values of the command return a string containing the properly formatted 3DOF Galil command string Model make3DOFCommand string
27. the system 25 In this study the response variables suitable to establish the performance of the system are first taken to be the mean absolute differences between FSX s motions F and amp 44 and the simulated motions measured on RMC s 3DOF zs and Wss Two standard deviations for each motion channel are then added to each value to account for variation across the entire test run within a confidence interval of 95 Equation 5 1 and 5 2 show the calculations used to establish the response variables for accelerations and angular rates where n represents the number of motion data points within a simulation run RV TEEN 20 5 1 RV Le 1l aamM sscol 20 5 2 n As such six response variables were calculated to evaluate the system s performance The response variables represent the mean margins of errors for each motion channel with a confidence interval of 95 20 The sample standard deviation is used to measure dispersion for each response variable in accordance with Equation 5 3 where is the sample standard derivation x a dataset value i e x faao fssol and y the arithmetic mean of a dataset 1 n 1 Q ila 1 5 3 In theory as the accuracy of a motion channel increases the value of its response variable decreases Depending on the channel under study not all response variables share the same relevance to performance For example when tuning the translational acceleration channel 99
28. 0 2130 Climb ee 0 4322 2 1833 0 7500 4 425 1 9378 5 0277 Descent 0 7376 0 0608 0 3730 0 2275 3 0498 0 2469 First a definitive improvement can be found in the tilt coordination channels As is expected this improvement on tilt coordination had a slightly detrimental effect on angular channels This is because rotational cues are generated when the platform is conducting tilt coordination to simulate low frequency translational accelerations Next there is minimal variation on the z axis translational motion channel However from the pilot s perspective the improvement is quite significant due to the absence of unsolicited overshooting motions As can be seen in the plot of z axis accelerations for the second test run featured in Figure 70 the signal is substantially dampened compared to that of the first test run It is so because the filter damping ratio of this channel s filter equation was increased while the channel s scaling coefficient remained at 0 5 127 Table 22 Response Variable Variations between Tests Run 1 and 2 Response Variable Variations for Each Motion Channels Relevance Translational Tilt Coordination Angular Flight Profile 2 2 2 1 1 ay mes ay m s a mes p deges q deges r deges Continuous Climb 0 6187 0 0009 0 5051 da o 18618 0 1899 Turn Descent 0 8006 0 0148
29. 1 1905 9 2 9 4 9 6 m 9 8 w wI amp 10 10 2 10 4 10 6 Descent Profile Test Run 2 10 8 0 5 10 15 20 25 30 Time s Figure 70 z Axis Accelerations Climb Profile vs Test Run 2 Despite the seemingly low amplitude of the a signal on this particular test manoeuvre the sensation is quite realistic from the pilot s standpoint It is important to keep in mind that the accelerations simulated in this test manoeuvre only vary by a few tenth of mes The next test 128 run will feature a slightly higher scaling coefficient to investigate if more fidelity can be achieved on the heave motion channel Next the tilt coordination channel is studied for a slightly slow and underpowered tilt coordination on pitch attitudes Turn Profile Test Run 2 4 1 1 1 1 1 0 5 10 15 20 25 30 Time s Figure 71 y Axis Accelerations Turn Profile vs Test Run 2 By observing the variation of response variables in Table 22 it is found that gains on the x axis for tilt coordination are smaller than on the y axis It is so because motions of heave will typically compete with motions of pitch Indeed pitch and heave tend to occur simultaneously as the plane is climbing or descending It is not the case for motions of roll which typically occurs independently of z axis accelerations in level flight Consequently additional tuning is required for the x axis tilt coordination as there are interactions b
30. 140 CHAPTER 6 CLOSING REMARKS 2 55m einen nent name re cine nine 142 6 1 Introductions ner nn ir ste tite ET 142 6 2 Future Work 28 1 ceasesvoces caches mienne Rava trades ety jose lisent nn iles pw id ere en 143 6 2 1 Classical WF Algorithm Improvement ss 143 6 2 2 Flight Simulator Hardware Improvement eesceseceseceseceseceseeeeeceseeeseeesneesseeeseeeneeenaes 144 6 2 3 Future Applications eesriie en ae sonetoessdeseedsbacvettlolageel stiged eel tuewsles 146 ANNEXES sedate see ook caus vast Beet vlog ts 00s SE uate aide tres Sardi de entre Shae net nn Mt PAP den de cubes a 148 Annex A Motion Platform Security Briefing ss 149 Annex B Flight Simulator Components and Cost sus 151 Annex C Flight Simulator Wiring Diagram ss 153 Annex D UME Class Diagram ess usa dire ciate nn et nr con th been aaah et 155 Annex E Response Variable Calculation Code ss 157 Annex F Crossbow IMU 320 Simulink Interface 160 Annex G Recording and Playback y is lt cgalegscdecsstctesaseicaectensclesebesdodeunes EE e E EENET Eeoa 164 Annex H Pilot Rating Questionnaire ss 168 Annex I Pilot Information and Ratings ss 172 Annex J Ethics Board Documentation sise 176 Anex K C Implentation Coder oprie eet iie ee cue a EE e cece e teen E aea 188 REFERENCE S en Re Men e a E te N MR Re E E Ee 225 LIST OF FIGURES Page Figure 1 RMC s Triad 1500 E 3DOF Motion Platform 5 Figure 2 Motion Platf
31. 2 The second test run featured several modifications to the coefficients of the classical WF algorithm In addition some modifications unique to this implementation were introduced to tackle problems identified during the first test run The main objectives of this test run were to note any improvement on heave and tilt coordination motion major improvement to heave motions was expected as a result of implementing an overshoot management segment to the algorithm A study of the tilt coordination algorithm also led to some changes that are expected to yield a better simulation After several investigative test runs it was determined that the cause of underpowered signals in the tilt coordination channel was the tilt rate limiting block of the algorithm Indeed to limit the rate of tilt coordination from an angle of tilt the derivative is taken to obtain the tilt rate The integral of limited tilt rate is then taken to obtain the final tilt coordination angle As it is often the case the process of taking the derivative of a discrete signal is associated with heavy noise and zero crossings In this case signal sampling is discontinuous at a rate of 50 Hz and zero crossings can be found in typical test runs where values plateau momentarily at certain levels A sample plot featuring the derivative tilt rate taken in the tilt limiting block can be seen in Figure 69 It was discovered that the tilt coordination limiting block would scale down the
32. DOF CODE HERE feedback Actuator 6 moving up Feedback message W else if else if uInput actuator6Down DOF CODE HERE feedback Actuator 6 moving down Feedback message else if else if Mif return feedback process_user_input Udp h pee Udp h IDE Visual C 2008 Express Edition author Martin Vidovic author 2Lt Jean Le Bouthillier Latest Revision 13 November 2009 ek ee Description of UDP ek UDP is in nature a datagram protocol A datagram is a block of data with a well known size When communicating using UDP we are sending datagrams and receiving datagrams With each send or recv operation exactly one complete datagram is sent or received ek With contrast to TCP there is no connection in UDP A program reading datagram packets can do so from multiple sources on the same socket simultaneously This is a unique feature that can be taken advantage of in certain applications ek However UDP has a disadvantage over TCP since it does not offer any data reception acknowledgements This means that we may be sending datagrams but the other side of the communications channel will not received it This usually happens on unreliable data links or if the data link is overloaded ek To sum up the essential thing to remember is the datagram a packet of arbitrary data When doing send sendto you are sending a datagram
33. In order to ensure the safety of personnel and equipment involved with the operation of the flight simulator both software and hardware protections are incorporated into the design of the system Additionally some measures are taken to ensure the safe behaviour of personnel in contact with the flight simulator The specific design sub objectives relating to safety can be seen in Figure 21 Relative Position Emergency Shut Down Emergency Shut Down Figure 21 Safety Design Sub Objectives The motion platform hardware consists of commercially available components assembled into the final system The system has been certified to maintain its integrity provided that it is operated below certain limits However the system s raw capabilities far exceed the tolerance limits set by the manufacturer Because of this it is possible to cause damage to the platform if it is Operated without safeties such as limiters Because the platform is driven by a motion controller the most effective way to limit the platform s velocity acceleration and displacement is to implement software limiting To ensure safety software limiting can be implemented in the controller as well as in the motion cueing software Additionally an emergency shut down functionality can be implemented using a simple button on the GUI used to control the simulation If required an emergency button can swiftly overwrite current commands and send a stop co
34. Level View of Flight Recorder Block oo eile eeeeeeeeeeseecnseceseeeseeesaeesnaeeneenses 165 Figure 78 Inside View of Flight Recording Block 165 Figure 79 Top Level View of Playback and Crossbow Blocks 166 Figure 80 Inside View of Flight Playback block RMRR 167 XIV LIST OF TABLES Page Table 1 RMC s 3DOF Motion Platform Limits us 6 Table 2 Filter Parameters with Most Favourable Pilot Response UTIAS 11 22 Table 3 Cost Benefit Analysis of Flight Simulation Software 49 Table 4 Cost Benefit Analysis of the Three Major Types of Display 51 Table 5 Cost Benefit Analysis Between three Commercially Available Pilot Input Systems 57 Table 6 Programming Languages Compatibility for Key Components eee eeeeeeeeeeeeeeeee 64 Table 7 Information Requested from FSX Using FSXConnect ou ee eeeeeeeeeeneeeseeeeneeneeeeenees 75 Table Pert Allocation tor UDP Trabtie 2 2 24 en nin aa aai saiia roas iais 86 Table 9 Observations and Problems Discovered during User Validation 90 Table 10 Specal Mounting Parts csie si e eed ee nega AEE ages 91 Table 11 Crossbow IMU 320 Performance Dat 0 0 0 eee eesceseeeseceeeecnaeeeeesseeeeaeesnaeenseeees 100 Table 12 Experimental Flight Profiles Manoeuvres 104 Table 13 Experimental Procedure for Each Test Runs 107
35. Modifications were also made to the classical implementation of the classical WF algorithm to resolve issues of 2 order filter overshooting and underpowered tilt coordination channel These modifications led to a significant improvement to motion simulation The trials conducted with qualified pilots also provided invaluable information from potential users of such a system The positive ratings collected from pilots are an indicator that low cost motion simulation using commercially available products is a promising approach The information collected from these trials opens the door to many possible improvements on the existing system Overall the work and conclusions drawn in this work will be relevant to RMC as it moves to a more advanced flight simulator using a 6DOF motion platform This chapter will discuss the areas of interest for future work on RMC s 3DOF and the use of motion simulation at RMC Specific areas of possible improvement for the classical WF algorithm and the flight simulator hardware are presented followed by a range of possible applications for RMC S 3DOF motion platform in the light of the findings of this study 6 2 Future Work The study of RMC s 3DOF flight simulator through the experimental tuning and pilot validation processes has lead to several observations that can lead to significant improvements to the motion cueing algorithm and flight simulator hardware The following sections will discuss these observations i
36. Only small angles can be observed even on aggressive cabin tilts Rotational Rotational motion appears to be proportional However both pitch and roll motions appear underpowered Table 18 Response Variables for Each Motion Channels Test Run 1 115 Response Variables for Each Motion Channels Relevance Translational Tilt Coordination Angular Flight Profile 2 2 2 1 1 4 dy mes dy mes a mes p deges q deges r deges Continuous 13336 0 0331 0 2829 0 2437 1 5822 0 1601 Climb Standard Rate 0 4830 4 0451 0 9405 4 4526 2 1189 4 8868 Descent 1 5382 0 0272 0 3582 0 2744 1 8593 0 1625 For this test run the response variables Equations 5 1 and 5 2 serve a limited purpose by establishing a baseline level for comparison They will be of great value after the second test run as they will serve to objectively compare performance between test runs In future test runs motion channel performance improvements should be characterized by lower response variables 116 In light of pilot observations experimental plots are studied to determine the possible causes and solutions of motion cueing problems The descent flight profile featuring the largest translational acceleration error of 0 3582 mes is ideal to investigate heave overshoot problems As can be seen in Figure 63 the heave overshoot problem found dur
37. System Use Case Diagramme 69 Figure 40 Comparison between Coupled and Functional Approach 71 Figure 41 The Many to Many Relationship and the Mediator Design Pattern eee 72 xii Figure 42 SimConnect TCP IP Communications Model seeseseseseesessessresresseseresresserererreeseese 74 Figure 43 Different Axis Systems between Classical WF Algorithm and FSX 77 Figure 44 Simulink Implementation of the Classical WF Algorithm 79 Figure 45 Internal Simulink Block View Sample su 80 Figure 46 Geometrical Layout of the Actuators Top Frame View 81 Figure 47 Superimposition to Calculate Actuator Extensions 82 Figure 48 Networking in the Flight Simulation System 83 Figure 49 Classes Required for the Implementation of UDP uu eee eeeeeeeseeeneeeneeneeeeeaees 84 Figure 50 Typical UDP Communication 22e nine unes dieili sites 85 Fig r sl CHEF Code Sample bn nn nn RE RS EA nt ns 87 Figure 52 Installed Special Mounting Parts seu ste nt Mere tits 92 Figure 53 Head Mounted Display and TrackIR 5 ss 93 Figure 54 Comparison between Concept Draw and Actual Implementation 93 Figure 55 The Position of the IMU 320 on RMC s 3DOF Motion Platform eee 100 Figure 56 FSX s C 172 Key Information and Geometrical References 102 Figure 57 Plot of Climb Profile Motions against Time 0 ee ceeeeceescecseeeseeeseeesseessaeeneeees 104 Fi
38. Unified Modelling Language UML toolbox for the design and visualization of object oriented software 17 The granularity or level of detail of use case diagrams is generally left to the discretion of the author However it is 69 preferable to keep those diagrams at a low level of granularity to ensure that no architecture details are suggested by the diagrams In use case diagrams the actors or users are represented by stick man figure and the actions that the user must be able to accomplish are illustrated by ovals One such use case diagram can be seen in Figure 39 where the three fundamental use cases for Manage Connection Manual Control Flight Simulation Figure 39 System Use Case Diagram the system can be seen Simulator Operator First the simulator operator must be able to use the software to manage a connection with the motion platform More precisely this includes the ability to connect and disconnect from the motion platform Then the software must allow the simulator operator to manually control the motion platform This includes controlling the speed of the motion platform and moving it in any possible position The third functionality that must be provided by the software is to allow the simulator operator to conduct flight simulation This includes the initiation of a flight simulation and its termination from the software GUI The software implementation is conducted while keeping the use cases in mind
39. able to ask questions about the work conducted in the flight simulation lab The result of the study will be made available to satisfy the curiosity of any participants on the matter D POTENTIAL RISKS FROM THE STUDY 1 For each procedure used in this study describe any known or foreseeable risks to the participants including physiological psychological emotional social and economic risks The use of motion platforms for flight simulation presents psychological and physiological risks 183 Psychological risks include Disorientation and motion sickness Physiological risks include Injury could be caused by the unlikely event of a platform breakdown 2 Describe the procedures or safeguards in place to protect the participants in light of the risks identified in D1 Psychological risks Disorientation and motion sickness can occur when motion experienced by the passenger contradicts the visual cues he receives while flying the aircraft This is unlikely to happen as the algorithm used to control the motion platform will produce motions that support the visual cues experienced in the aircraft This has been experimentally verified using an inertial measurement unit used to measure motions experienced on the platform The algorithms used to drive the motion platform are also well known and understood In addition a clearly identified pause button is available to the passenger on the
40. accelerations pitch bank and angular rates The pitch and bank information are required to calculate the effect of gravity on the aircraft for the translational acceleration channel It is necessary to collect this information to address the fact that FSX 76 accelerations returned for each axis of the body do not include the effects of gravity As such the individual acceleration on each axis due to the force of gravity is calculated using Equations 2 12 to 2 14 from pitch and bank information obtained in FSX where a is the pitch angle and is the bank angle This formulation for y axis acceleration is however not proper for simulation as it does not account for the centrifugal force which is not available in FSX and generally brings the y axis acceleration to zero in a coordinated turn If this simplified formulation would be used unrealistic flight simulator tilt would occur during the simulation of coordinated turns Due to this limitation the y axis acceleration due to gravity is set to zero for this study This zeroing of 8y is appropriate as it will level the platform during a turn which is what is supposed to happen as the centrifugal force would neutralize the y axis acceleration gx sin a 9 81 m s7 4 1 Jy sin B 9 81 m s 4 2 gz cos a 9 81 m s7 x cos 4 3 Then the pause simulation event is requested in order to know when the simulation is paused or running This information is used to know when
41. array VGA and audio extensions were required to accomplish this task It is interesting to note that two VGA cables are required for the visual display It is so because the head mounted visual display acts as the secondary screen of the control station with a VGA splitter which sends a signal to the Pilot View Display a display designed to allow spectator to see what the pilot sees The IMU 320 is powered using a power supply which is not illustrated here Power is supplied with the serial cable Headset 2 ft Audio Cable 2x 25 ft VGA Cable 32 ft USB Cable Infrared ee Pilot View Display 25 ft VGA Cable 32 ft USB Cable ei Control Station Leo Visual Display 25 ft Serial cee O Safety Switch LL Power Cable IMU 320 Joystick Throttle 25 ft Ethernet Cable Galil Controller Motion Platform Figure 73 Flight Simulator Wiring Annex D UML Class Diagram 155 MainWindow feedback m motionSimulation platform Type ui UdpSend mDestAddr mSocket Initialize const char bool SendDatagram PortNumber const void size_t bool SendDouble PortNumber double bool SocketCleanup void UdpSend UdpSend Udp success ISOk bool Udp Udp Feedback string void MainWindow QWidget Q2S QString string 20 string QString checkConne
42. be changed into a QString return the Qstring obtained from the std string QString MainWindow S2Q string s return QString fromLocal8Bit s c_str S2Q pe Event handler for a click on the Connect Button This method asks the mediator toattempts a connection to a controller at the IP adress specified by the user 196 void MainWindow on_connectButton_clicked Ilui gt feedback gt clear Clear the display Feedback Attempting to connect Feedback message platformType m gt connect Q2S ui gt ipAddress gt text Ask the mediator to attempt a connection at the specified IP if platformType No Controllers Detected Checking to see if a connection was established ui gt connectedTo gt setText Connected to S2Q platformType Updating the GUI with either 3DOF or 6DOF Feedback m gt execute markPositionParked Ask the mediator to mark the initial platform position as parked Feedback m gt execute center Center the motion platform ui gt connectButton gt setDisabled true Disable the Connect button ui gt FSXConnect gt setEnabled true Enable the Begin Motion Simulation button enableManualControls true Enable the manual controls M if on_connectButton_clicked pe This method is used to check the status of the UDP and FSX connections Once it has established the status of those connections it enables or disables the appropria
43. center or they can ask the model to process length velocity and acceleration commands No platform information is required to communicate with the model Because of this this model can be used to interface RMC s motion platform with any other C application with ease pps eh include Model h pe Used to establish a connection with either RMC s 3DOF or 6DOF platform Some parameters are also initialized depending on which platform the connection is established with k param ip the ip adress of the controller to which a connection is attempted return a string containing either 3DOF or 6DOF string Model connect string ip ipAddress ip platformType 3DOF Default platform type 3DOF try Attempts to connect to the 3DOF g new Galil ipAddress Creates a Galil Object g gt timeout_ms 1000 Allow for a 1 second timeout g gt connection Connect to the controller WI try catch string e No 3DOF controller detected platformType 6DOF Change platform type for DOF try Attempts to connect to the DOF ek a 6DOF CONNECTION CODE HERE ee platformType No Controllers Detected for now until there is a 6DOF HI try catch is a catch all form platformType No Controllers Detected catch catch if platformType 3DOF number_of_actuator 3 Setting number of actuators conversion_factor 30000 Metric to 3DOF conversion factor
44. const std string amp void interrupt int int libraryVersion message int std string programDownload const std programDownloadFile const std programUpload std string programUploadFile const std string amp void read std string record const std string amp std recordsStart double void setSource const std string amp const std string amp const std string amp void source const std string amp const std string amp std string sourceValue const std vector lt char gt amp const std string amp double sources std vector lt std string gt write const std string amp int Galil rector lt char gt Annex E Response Variable Calculation Code 157 158 The calculation of response variables between test runs is a repetitive task As such a MATLAB function CalculateRV was developed to expedite the experimental process This code performs the smoothing of recorded signals and returns the response variable for each motion channel The code found on the MATLAB exchange to perform the triangular smoothing technique used for smoothing the signals is also included function RV CalculateR V Dataset FlightProfile Function CalculateRV Calulates the Response Variables and standard deviation Dataset represents the recorded motions o
45. flight simulator He can press this button at any time to interrupt the simulation and center the motion platform A verbal cue from the passenger can also alert the operator who can stop the simulation and center the motion platform So far no passenger of RMC s 3 DOF flight simulator has ever exhibited disorientation or motion sickness Physiological risks The system in use is controlled by a robust software interface which allows the operator to stop the motion platform at any time This software was debugged and tested by other faculty members Its behaviour is very stable and predictable In addition an emergency power switch hardware located close to the operator station can be used to completely stop the system at any time This redundancy between software and hardware shutdown mechanism guarantees that the operator will always be able to stop the motion platform Although the motions generated by the system are not powerful enough to throw a passenger off the platform the passenger is always secured to his seat with a four point harness bolted to the motion platform base In the unlikely event of a platform breakdown the passenger will remain securely fastened to his seat and the motion platform will be immediately de energized Finally a strict procedure as highlighted in Annex B is always be followed during the boarding and un boarding of passengers Overall the operation of RMC s 3 DOF flight simulator is extremely safe
46. flight simulator hardware assessment The complete results for each pilot can be found in Annex I Table 27 Average Pilot Ratings for Flight Simulation Hardware Pilot Visual as Controls Audio Cockpit Overall Tracking Average 2 17 2 21 2 21 1 38 1 63 1 69 Deviation 0 81 0 96 1 05 0 48 0 71 0 51 Ratings from 1 to 5 where 1 is excellent and where 5 very poor Overall pilots deemed the hardware used to be adequate for the conduct of flight simulation As reflected by the average ratings and standard deviations pilots consistently rated the audio cockpit and overall attributes with very positive ratings Higher average ratings and 137 larger standard deviations between pilot ratings for the visual head tracking and controls attributes was noted These also represent the areas where pilot had the most complaints and suggestions Table 28 Summary of Pilot Complaints for the Flight Simulator Hardware Equipment Complaints Pilots with prescription glasses had a very hard time with the head mounted display Some pilots complained about the Head mounted resolution of the display Some pilots display found it hard to adjust the head mounted display Some pilots reported slight discomfort as the head mounted display rested on their nose Some pilots were dissatisfied by the gain on head motions especially in the yaw channel Pilots felt a learning period was required make the b
47. for the Heave up button As long as this button is pressed the platform will heave up The request is passed to the mediator in this method ie void MainWindow on_heaveUp_pressed Feedback m gt execute heaveUp Send request to mediator on_heaveUp_pressed Jx Event handler for the Heave Up button When the button is released the motion stops The request is passed to the mediator in this method void MainWindow on_heaveUp_released Feedback m gt execute stopMovement Send request to mediator on_heaveUp_ released pe Event handler for the Heave down button As long as this button is pressed the platform will heave down The request is passed to the mediator in this method af void MainWindow on_heaveDown_pressed Feedback m gt execute heaveDown Send request to mediator on_heaveDown_pressed pe Event handler for the Heave Down button When the button is released the motion stops The request is passed to the mediator in this method f void MainWindow on_heaveDown_released Feedback m gt execute stopMovement Send request to mediator on_heaveDown_released pe Event handler for the Surge forward button As long as this button is pressed the platform will surge forward The request is passed to the mediator in this method void MainWindow on_surgeForward_pressed Feedbac
48. involvement will be negotiated and how their consent will be obtained n a 3 The Tri Council Policy document suggests that community approval may be required when the research involves indigenous people the cultural property of indigenous people or where individuals speak on behalf of an indigenous group If this study requires the approval of an indigenous community as a whole describe how this approval will be obtained n a F PRIVACY OF PARTICIPANTS AND CONFIDENTIALITY OF DATA 1 Describe the procedures to be used to ensure privacy of participants and confidentiality of data both during the research and in the release of the findings The data will be obtained using the pilot rating questionnaire The data collected is anonymous and does not identify the participants The findings will only include pilot ratings and observations along with statistical information The participants cannot be identified through this information 2 How will written records video audio tapes completed questionnaires and recordings be kept secure 185 The questionnaires will be kept in a secure location during the research After the research the questionnaires will be scanned and stored on a computer The paper questionnaires will be destroyed after an electronic copy is available 3 What is the final disposition of the data C Confidential shredding of the paper records C Erasing of electronic data a
49. key information and geometrical references for the C 172 model used in this study It is important to note from 103 Figure 56 that the position of the pilot is slightly to the left of the aircraft s center of gravity CG The accelerations and angular rates obtained from FSX are from the CG position In order to compare the motions of the FSX aircraft with those of RMC s 3DOF motion platform both motion measurements must be taken at the same position with respect to the pilot In this case the motion measurements are taken at reference frame F4 for the simulated aircraft and reference frame Fs for the flight simulator The motions recorded for the simulated aircraft in FSX are however reported from the center of gravity of the aircraft which is represented by reference frame Fg Therefore for the C 172 it is necessary to use classical mechanics to determine the accelerations that are generated slightly left 46 cm of the CG position where reference frame F4 can be found under the pilot s seat It is well known that in a rigid body the angular rates are the same for any point of the body 29 However translational accelerations a distance away from the CG are affected by angular accelerations of the body around its CG To obtain the accelerations at an arbitrary point from the CG in a rigid body it is necessary to account for translational accelerations of the body normal accelerations and angular accelerations Equation 5 4 29 is used to ca
50. level of 65 portability as it can be compiled on many platforms and many programmers are likely to be able to modify the code Finally it can also be seen in Table 6 that Visual Studio Professional 2008 VS2008 has been included as a flight simulator component Although it is not a flight simulator component per se it is included here to show its high level of compatibility towards the other components In addition VS2008 supports a full integration of the GUI design toolset QT This integration is done seamlessly through the installation of an add in to VS2008 Because of the quality of this integrated development environment IDE and its integration of GUI designer tools it is the tool of choice to program the control software 3 5 Summary This chapter provided insight into the design phase of RMC s 3DOF flight simulator In accordance with mainstream engineering design process 13 the design objectives and constraints of RMC s 3DOF flight simulator were evaluated This chapter provided a thorough discussion of those objectives and constraints Finally the flight simulator s key components selected for the design of RMC s 3DOF flight simulator were discussed Having completed a thorough design process and identifying the components required for the design the next chapter will discuss the implementation of the flight simulator 66 CHAPTER 4 IMPLEMENTATION 4 1 Introduction After the analysis of the design obje
51. mSocket INVALID_SOCKET Check if the object state is valid return false If something went wrong pee Receive the datagram return this gt ReceiveDatagram nValue sizeof nValue nTimeoutMs ReceiveDouble pe This method is used to receive a double array over the network through a UDP datagram pp param nValues a pointer to a double array to store the incoming datagram param nNumberValues the expected size of the double array coming over UDP param nTimeoutMs the timeout in milliseconds to wait for the double array return amp lt true amp gt if a double value was succesfully received amp lt false amp gt otherwise bool UdpReceive ReceiveDoubleArray double nValues size_t nNumberValues long nTimeoutMs pe Check if Winsock is ready and socket is created if IsWsaReady false this gt mSocket INVALID_SOCKET Check if the object state is valid 221 return false If something went wrong Ja Receive the datagram Ff return this gt ReceiveDatagram n Values nNumberValues sizeof nValues nTimeoutMs ReceiveDoubleArray UdpSend h ee UdpSend h TIDE Visual C 2008 Express Edition author Martin Vidovic Modified by 2Lt Jean Le Bouthillier Latest Revision 13 November 2009 ek pragma once ifndef _SP_UDP_SEND_H_ define _SP_UDP_SEND_H_ include Udp h class UdpSend private pe TP socket used for sending UDP Datagram The so
52. motion cueing algorithm and the Flight simulation environment Tn other words the GUI nor the algorithm or flight simulation environment knows anything about the motion platform 4k 37 include Mediator h pe Mediator Constructor Mediator Mediator udpLengths Initialize 127 0 0 1 10107 Initialize the lengths UDP server udpVelocities Initialize 127 0 0 1 10108 Initialize the velocities UDP server udpAccelerations Initialize 127 0 0 1 10109 Initialize the accelerations UDP server m new Model Initialize the Model f new FSXConnect Creates an FSXConnect object Mediator Jx The connect method is called by the GUI to perform a connection This method simply passes the request to the model which will attempt a connection and return wether the connnection was successful 3DOF or 6DOF or unsuccessful ek param ipAddress a string containing the ip adress of the controller return a string containing the platform type either 3DOF or 6DOF string Mediator connect string ipAddress platformType m gt connect ipAddress Pass the connection request to the model if platformType 3DOF number_of_actuator 3 lengths new double number_of_actuator Setting the size of the double array for lengths velocities new double number_of_actuator Setting the size of the double array for velocities accelerations new double number_of_actuator Setting
53. my work Their patience availability and commitment to education are commendable attributes which played an essential role in the success of my work Within the department I would like to thank Mr Ben Brownlee and Dr Yong Kui Wang for their support in developing solutions to problems which arose during the study Their knowledge and willingness to support me in unfamiliar areas contributed greatly to the quality of this work I also have to express my thanks to Martin Baxter of the Electrical Engineering department for his support in machining and metal working Next I would like to thank the volunteer pilots who gave their time and allowed this study to benefit from their precious experience and opinions The RMC Flying Club and the Ontario Fun Flyers organizations were instrumental in providing volunteers for this effort My iv outreach for volunteer pilots was met with great enthusiasm by Mr Tucker Densmore and Mr Barry Smith who recruited and organized volunteers Finally I would also like to extend special thanks to my house mates Dan Gabriel and Alex from whom I received many encouragements and motivation throughout the year To Sylvia I must thank you for your understanding patience and support even as I cultivated the habit of working very late each night It is clear to me that your support and love allowed me to focus my energies and creativity into this work ABSTRACT Flight simulation is extensively used by the Canadia
54. pe Event handler for a click on the Clear Display button The feedback window is cleared void MainWindow on_clearButton_clicked ui gt feedback gt clear Clear the feedback window on_clearButton_clicked 198 Event handler for a check in the Feedback ON OFF Checkbox In this method the feedback window is enabled or disabled according to the user input ek param s the state of the checkbox 0 is unchecked 2 is checked f void MainWindow on_checkBox_stateChanged int s if s 0 If the checkbox is unchecked ui gt feedback gt setDisabled true Disable the feedback text area feedback false Set the feedback flag to false Mif else if s 2 ui gt feedback gt setEnabled true Enable the feedback text area feedback true Set the feedback flag to true else on_checkBox_stateChanged pe Event handler for the speed slider When the slider is moved the speed of the motion platform must be changed Everytime the value changes it is sent to the mediator af void MainWindow on_speedSlider_sliderMoved int int slidervalue ui gt speedSlider gt value ui gt speedLabel gt setText Speed QString number slidervalue 10 m gt change_speed slidervalue on_speedSlider_sliderMoved pe Event handler for the Erase Parked Position button The request is passed to the mediator in this method wh void Main Window on_e
55. position flat screens onto the platform Although positioning flat screens or a projection system off the platform is a possible configuration the design of an enclosure is beyond the scope of this study and without it the immersion of the pilot would be limited Figure 28 Z800 3D Visor The head mounted display is a convenient visual display with its light weight availability in the department and immersive effect When combined with a head motion tracker it becomes an acceptable display for this study The Z800 3D visor used for RMC s 3DOF flight simulator is shown in Figure 28 The use of an infrared head tracking device allows the passenger to move his head in 6DOF to obtain an unlimited field of regard as if he is sitting inside the actual cockpit of the aircraft This feature combined with a head mounted display is the next best thing to a cockpit replica It is important to note that other tracking devices utilize motion sensors to 53 capture head movement However such tracking devices are not compatible with the motion platform as motion cues generated by the platform could create undesirable effects by creating false positive gyroscopic cues Other technologies utilizing magnetic field head tracking could be viable for motion platform head tracking but were not commercially available The only commercially available head mounted infrared tracking system the TrackIR 5 by NaturalPoint is shown in Figure 29 This motion track
56. prefix double v ostringstream ss ss str ss setf ios fixed ios floatfield ss precision 2 ss amp lt amp lt prefix amp lt amp lt amp lt amp lt v 0 amp lt amp lt amp lt amp lt vp 1 amp lt amp lt amp lt amp lt v 2 return ss str makeCommand px Used to send an actuator length command to the motion platform ek param command a double array containing the three actuator lenghts void Model length_command double command for int i 0 i amp lt number_of_actuator i command i conversion_ factor Converting metric to 3DOF units ee Here the 3DOF motion platform is limited in its lower reach to 5000 due to a design flaw of the platform More information about this flaw and the limitation it forces upon the operation can be found in thesis if command i amp lt 5000 command i 5000 if command i amp gt 6000 command i 6000 for g gt command FL 6050 6050 6050 Define the forward limit slightly higher for flight simulation g gt command BL 5050 5050 5050 Define backward limit for all actuators only 5050 because of platform design g gt command PT1 1 1 Register all actuator for on the fly absolute position move g gt command make3DOFCommand PA command Send the motion command to the controller 212 213 length_command px Used to send an actuator velocity command to the motion platform 4
57. simulator acceleration is affected during this process However if the tilt coordination angles are sufficiently small this reduction has a negligible effect on the perceived overall motion 8 Simulator Cab Imes 9 75 mes g Zs Figure 5 Example of Tilt Coordination In Box five of the tilt coordination channel the low frequency aircraft accelerations are extracted using a low pass filter and held in the intermediate variable f Tilt coordination takes places in Box six where Bi hf Ofr YB is a vector containing the low frequency tilt coordination Euler angles The low frequency tilt coordination Euler angles of low performance aircrafts with low speed and manoeuvrability such as airliners are described in Equations 2 3 to 2 5 8 fly g op 2 3 Sy OB 2 4 wp 0 2 5 20 In Box seven the low frequency tilt coordination Euler angles are rate limited to be below the vestibular motion detection threshold values As described by Equation 2 6 the output of the tilt coordination channel is combined with the output of the rotational channel to form the simulator Euler angles where Bor gt 0 y PsL Bsu Bs 2 6 2 3 4 Scaling and Limiting The scaling and limiting of aircraft accelerations and angular rates is essential to generate motion platform commands that are proportional to those of the simulated aircraft while at the same time suited for the physical limitations of th
58. so because the expected 111 signal is of a smooth nature Indeed the noise observed is not expected when observing a flight simulation system such as the one under study and is not caused by the motion platform motions as it exists even when the IMU is perfectly still For this reason a linear moving window average smoothing function described by applying Equation 5 4 twice on each data point of the Crossbow IMU outputs is used to smooth the signal 30 For this study acceptable noise attenuation without a significant loss of precision is achieved with a moving window average width of nine data points Yj 4tYj 3 Yj 2 Yj 1 Y tY j 1tYj 2tYj 3 tY j 4 5 4 Figure 61 shows a typical angular rate Crossbow IMU signal and the resulting smoothed signal As can be seen this smoothing technique is very effective at neutralizing the noise present in the signals However due to the moving average nature of this smoothing technique it cannot be applied to the extremities of a signal because four values are required before and after a data point to apply the smoothing Equation As such the first and last four data points of a signal are always zeroed by this technique To work around this problem it is convenient to simply average the unsmoothed data points and replace each of them with the mean value obtained In this case such extremity errors are acceptable because only 0 08 seconds of data at each extremity of a signal remains unsmoothed
59. solution which is felt to be very realistic During this process positive improvements were also tracked on relevant motion channels between each test runs by calculating variations on response variables The experimental study reached a conclusion when gains of performance were minimal for modifications to the configuration and where improvements on select motion channels were no longer possible without disproportional lost of performance on related channels To provide an impartial measure of the flight simulation a pilot validation exercise was then conducted using the configuration derived from the experimental study of the motion response 5 4 Pilot Validation Exercise After having optimally tuned the classical WF algorithm through an experimental observation of the system a pilot validation exercise was conducted to assess the perceived pilot performance of the flight simulator In addition this exercise was used to collect pilot ratings on the unique design of RMC s 3DOF flight simulator This validation exercise was seen as an opportunity to collect pilot ratings and subjective performance data Comments were also collected during the pilot validation phase Such comments will prove to be valuable for future projects involving aircraft simulation and RMC s 3DOF motion platform In the section to follow the activities conducted with volunteer pilots will be discussed Finally the pilot ratings obtained during this exercise will be given
60. the IMU is motionless Those constants were established by recording the running average of each rotational velocity signal for a long period Using these constants all three rotational velocity channel exhibit a signal equal to zero when the IMU is motionless Annex G Recording and Playback 164 165 The implementation of the classical WF algorithm was modified twice to perform recording and playback operations For recording operations a flight recorder block was simply added to the classical WF algorithm implementation Using this block the motions were recorded to the MATLAB workspace as a Structure With Time featuring a double decimation and recording at the same frequency as the system Figure 77 shows the top level view of the Fligth Recorder block and Figure 78 shows the inside view of the same block Acc lerations ex ay az UDP Server Flight Recorder Figure 77 Top Level View of Flight Recorder Block Sink Block Parameters To Workspace To Workspace Write input to specified array or structure in MATLAB s main workspace Data is not available until the simulation is stopped or paused Parameters Variable name Standard _Turn Limit data points to last z py To Workspace Sample time 1 for inherited 0 02 a Save format Structure With Time E Log fixed point data as an fi object Co cancel Figure 78 Inside View of Fli
61. the mediator in this method sI void MainWindow on_m2down_released Feedback m gt execute stopMovement Send request to mediator on_m2down_released pe Event handler for the Motor 3 Up button As long as this button is pressed actuator 3 will move up The request is passed to the mediator in this method ial void MainWindow on_m3up_pressed Feedback m gt execute actuator3Up Send request to mediator on_m3up_pressed px Event handler for the Motor 3 Up button When the button is released the actuator stops The request is passed to the mediator in this method EI void MainWindow on_m3up_released Feedback m gt execute stopMovement Send request to mediator on_m3up_released ee Event handler for the Motor 3 Down button As long as this button is pressed actuator 3 will move down The request is passed to the mediator in this method void MainWindow on_m3down_pressed Feedback m gt execute actuator3Down Send request to mediator on_m3down_pressed px Event handler for the Motor 3 Down button When the button is released the actuator stops The request is passed to the mediator in this method we void Main Window on_m3down_released Feedback m gt execute stopMovement Send request to mediator on_m3down_released pe Event handler for the Motor 4 Up button As long as
62. to informed consent or place participants at a higher risk level than anticipated Such unforeseen changes to the protocol during the conduct of the research must be communicated within four working days to the REB Chair as well as the actions taken to protect the dignity of participants Any undesirable experience or response adverse event from participants during their involvement in the study must also be reported within four working days to the REB Chair as well as actions taken by the research team to protect the participants Such adverse event may be emotional psychological physiological or physical in nature Signature Dr J Daniel Lagac Roy Chair RMC REB Phone 613 541 6000 x6572 E mail daniel lagace roy rmc ca Annex K C Implentation Code 188 189 The integrality of the C code used for the implementation of the flight simulation system is available in the pages to follow The code has been printed with reduced spacing and without any empty lines for a more compact printing Indeed due to the length of some code lines it was not possible to print the code in the document without a substantial lost of space The Galil h and SimConnect h code is not included as it is in the public domain and can be easily obtained from either Galil or Microsoft FSXConnect h pee FS XConnect h IDE Visual C 2008 Express Edition author 2Lt Jean Le Bouthillier Latest Revision 20 Sept 2010 ek pra
63. tuning of the classical WF algorithm In addition this experiment was essential in order to optimize the classical WF algorithm and to allow RMC s 3DOF motion platform to achieve its full potential as a flight simulator In the sections to follow the experimental design for this study will be discussed 5 2 1 Statement of the Problem and Objectives An experiment is usually conducted in order to gather the necessary information to address a specific problem 25 In the case of this study the problem of interest is to determine the optimal configuration of the classical WF algorithm in order to generate the most realistic flight motion simulation using RMC s 3DOF motion platform Several coefficients such as scaling and filtering coefficients can be altered to affect the response of the motion platform As reported by 8 9 10 and 12 the fine tuning of the classical WF algorithm is best accomplished through an iterative process where the different coefficients of the algorithm are modified in order to generate the best motion simulation with the available hardware As discussed in Chapter 2 the filters and scaling coefficients of each channel interact to generate the final motion signal With this in mind the experiment is conducted with one overall objective to discover the best parameters for each motion channel and how these motion channels can be combined to produce the most realistic motion simulation As the three channels are combin
64. 0 0 If all the actuators have a speed of 0 answer false The platform is not in motion MI if Mif else if platformType 6DOF If connected to the 6DOF 11 6DOF CODE HERE else if return answer Return the velocity for all actuators tell_platform_velocity pe This method processes manual user inputs into motion platform commands This method only processes events that typically come from user input such as a GUL Surprisingly enough it is not possible to do a switch on strings in C For this reason if and else if has to be used to triage the commands ek param ulnput a string containing the command request return a string containing feedback in the execution of the requested command ze string Model process_user_input string uInput string feedback Cannot execute command while Parking if prevent_movement true amp amp is_platform_moving If the platform stopped moving movement can resume prevent_movement false Set the movement prevention flag to false g gt command BL 5000 5000 5000 Define backward limit for all actuators only 5000 because of platform design WI if if prevent_movement If there is no movement prevention if platformType 3DOF If we are connected to the 3DOF g gt command AB1 Abort any current movement of the motion platform if uInput erasePosition g gt command BL 2147483648 2147483648 2147483648 R
65. 14 deges 0 3710 deges 110 At first glance the loss of precision on translational channels is much lower than the lost of precision on angular rate channels This can be explained by the fact that the Crossbow IMU used to measure the motions of the platform exhibits an especially noisy signal for angular rate measurements As can be seen in Figure 60 where an acceleration signal in mes is compared to an angular rate signal in deges the translational acceleration signal exhibits sporadic and low amplitude noise On the other hand the angular rate signal exhibits a prevalent noise of an amplitude 0 5 deges Signal noise is reflected in the margin of error calculated in Table 14 as the slightest phase shift between the flight profile motions and the Crossbow IMU outputs will generate an error similar to the noise amplitude A phase shift between the signals could be caused for example by the system s latency 5 T T T T ia 4 al S3 amp 2 4 gt 1 L I 4 2 EUR ii i du i l l i ul Rh Ly 4 wa P 0 au lan n 2 Da i 2 Acceleration x 7 3 Angular Rate p 4 0 5 10 15 Figure 60 Noisy Crossbow Crossbow Angular Rate Signal Thus the weakest link of the system is the Crossbow IMU which exhibits noisy signals for angular rates This lost of precision can be reduced by smoothing the Crossbow IMU signals to remove the high frequency noise In this case it is appropriate to do
66. 4 Crossbow IMU 320 Packet Format For the S1 packet the Packet Type identifier is 0x5331 S1 and the payload lenght is 18 bytes The accelerations and angular rates are received first in the payload as signed short with 2 s complement and take up a total of 12 bytes of storage As can be seen in Figure 75 the Serial receive block of the Instrument Control Toolbox by Mathworks is used to listen on the serial port for the incoming IMU packets The configurations as seen in Figure 76 are set in accordance with Crossbow s specifications for serial communication with the IMU 320 The Serial Receive block is configured to receive packet with a header of 0x55555331 UUS1 Then only the following 13 bytes of data the payload lenght and the relevant 12 bytes of data are kept A simple process is applied to transform the bytes into relevant acceleration and angular rates First the most significant byte of each variable is multiplied by 256 for it to gain its true value In this case the values are transmitted using the Big Endian standard which means the most significant byte is first Then both bytes are added together to form a double value the output type is set in 162 the add block parameters Finally a scaling factor is applied to the values as per Crossbow s re i 7 x 20 specifications In this case the accelerations are scaled by a factor of FT and the angular rates are scaled by a factor of Terminator COM1 3
67. 4 is comprised of a translational motion channel a tilt coordination channel and a rotational motion channel Due to the limited motion envelope of any motion platform it is not possible to reproduce constant low frequency translational or rotational motions as they would constantly drive the motion platform to its physical limits As such the classical WF algorithm filters out lower frequency motions in both 15 its translational and rotational motion channels An approximation of low frequency translational motions is performed in the tilt coordination channel where the motion platform is oriented slightly off the gravity vector in order to create a translational motion sensation The inputs to the classical WF algorithm are the aircraft accelerations Taa and the aircraft angular rates aq where fa ay ay az g and waa p q r In this formulation ay ay and a are the accelerations along the x y and z axes respectively and g is the acceleration due to gravity 9 8 mes The variables p q and r represent the Euler angle rates at which the aircraft is moving around each of its axis from x to z in order Lis Translational Motion Channel l i 1 1 a 1 m i f fh fo A HP Sy 1 La Ss faa SCALE FILTER SE i i g l eee E EE i ________JiltCoordinationChannel ______ I 1 LP f TET B RATE Bs FILTER COORD LIMIT S E i Rotational Motion Channel jimi ee NN 1 i I E T
68. 8400 A 8 none 1 Serial Configuration Crossbow Accelerations Data com1 Status 5 a Serial Receive Terminator1 Crossbow Rates Crossbow Rates Subtract1 uint8 sl Conversion Acceleration Scaling Constant4 c 4 ConstantS Divide Constant2 Math Function Constant3 Rate Scaling Divide1 Figure 75 Crossbow IMU 320 Simulink Interface M Block Parameters Serial Configuration M Source Block Parameters Serial Receive Serial Configuration Configure the parameters for the serial port Parameters Communication port COM1 Baud rate 38400 Data bits 8 Parity none Stop bits 1 Serial Receive Receive binary data over serial port Parameters Communication port COM1 Header UUS1 Terminator lt none gt Data size 13 1 Data type ints Byte order BigEndian none Flow control Timeout 10 ox cance Help Enable blocking mode Action when data is unavailable Output last received value x Custom value 0 Block sample time 0 02 Figure 76 Simulink Serial Configurations IMU 320 163 As can be seen in Figure 75 two constants are substracted from the x and y angular velocity channels It is so because the particular unit used for this study exhibits a slight signal deviation from the origin even when
69. C Code Sample As can be seen in Figure 51 which shows a code sample the highest commenting standard has been applied to the entire C implementation Finally a UML class diagram of the system complements the code documentation and comments The UML class diagram of the system can be found in Annex D 88 4 3 Software Validation Following the software implementation phase software validation activities were conducted to ensure that the software implementation allows the operator to conduct all of the selected use cases scenarios flawlessly In this case the operator needed to be able to manage a connection to the motion platform manually control the motion platform and perform flight simulation Two distinct software validation activities were conducted during this phase technical verification and user validation 4 3 1 Technical Verification Technical verification consists of verifying each of the intended use cases thoroughly to ensure that any combination of actions performed by the users are supported by the software and that there are no errors in the operation of the system Several tests were conducted with a single purpose to crash or uncover errors in the system The author of this study was able to crash the initial system several times and uncover a myriad of errors both in the code and the logic Therefore the technical verification phase has proven to be very effective in improving the software before its release
70. CKNOWLEDGEMEN T Seet T AEN aes delves ioe OA A dies T AORO AAEE A Ill ABSTRACT E RS lie URN rentes de LENO VO CARS ilu ete bes V R SUM sn Mii net AR te teal Ne et Ret a ti Cae Sd en cn VI LABLE ORGON TENTS ER nn a ete A nisn hiss VII LIST QF FIGURES cn hide tnt listen tte rE EEEE EEEE nent est EENET ca ten latente ide X LIST OR TABLES sentant urine nt son risen ced Bose a eeee ah Gases eed Sea eee Ba en Cannes XIV LISPOFS MBOES ERA RR ne Re eT XVI CHAPTER 1 INTRODUCTION net ren femme den ner trees rires nes REAREA mette 2 1 1 Intr dUCUO N 3 fuss Saas AR es REA RE ie SAA AR tas 2 12 Fight Simulationen ae nn a E a T Ea A a ind 2 1 2 1 Flieht Simulation HIStory eeaeee ee ee Re ogc ees iis isa Et E hae ee Le TA 2 1 2 2 Advantages of Flight Simulation ss 3 1 2 3 Modern Flight Simulation 3 1 3 RMC 3DOF Motion Platform sisi 4 1 4 MOTIVATION PMR te eee EE Re oe EE EEN RS A 7 1 5 NeT CRE PP PRE PDP ED RES D PR E PER EE PU LR TE ET 8 1 6 CONADUHONS A345 eles nus Tr Ana cie tte et gata ns Ti a EON EAEE T E tien ie 8 1 7 Th sis Stru ture s she nl Sas nn ne SEs mette ONE teens Co 10 CHAPTER 2 MOTION DRIVE ALGORITHM AND 3DOF ADAPTATION cccccccccsssceceesseeeeeens 12 2 1 IN trOQUCTIONT 533 2 5522 IE E se nn me ne TT TE ae ee aa 12 2 2 Reference Frames and Axis System sise 13 2 3 Classical Washout Filter Algorithm 14 2 3 1 Translational Motion Channel ccccccccssssssscecececeesen
71. Double double nValue long nTimeoutMs bool ReceiveDoubleArray double nValues size_t nNumberValues long nTimeoutMs E endif UdpReceive cpp ee UdpReceive cpp IDE Visual C 2008 Express Edition author Martin Vidovic Modified by 2Lt Jean Le Bouthillier Latest Revision 13 November 2009 1k ek Description A helper class to make web service adressing WSA Windows socket Winsock initialization and finalization easier and more reliable or include UdpReceive h UdpReceive Constructor di UdpReceive UdpReceive Initialize the mSocket to an initial value this gt mSocket INVALID_SOCKET UdpReceive px UdpReceive Destructor UdpReceive UdpReceive Release networking resources this gt SocketCleanup UdpReceive px This method is used to close the current network socket void UdpReceive SocketCleanup Close the socket if this gt mSocket INVALID_SOCKET closesocket this gt mSocket Closes the socket this gt mSocket INVALID SOCKET HI if 219 220 SocketCleanup px This method is used to create a network socket and initialize the UDP networking ek param nSourcelp a pointer to the IP address of the Datagram source param nSourcePort the port number of the Datagram source return amp lt true amp gt if the initialization was succesful amp It false amp gt otherwise eh bool UdpReceive Initialize
72. Edition author 2Lt Jean Le Bouthillier Latest Revision 20 Sept 2010 t a pragma once ifndef MAINWINDOW_H define MAINWINDOW_H include Galil h include Mediator h include amp lt QtGui QMainWindow amp gt include amp lt Windows h amp st include amp lt cstddef amp gt include amp lt string amp gt include amp lt cassert amp gt include amp lt iostream amp gt using namespace std namespace Ui class Main Window Ui class MainWindow public QMainWindow Q OBJECT public MainWindow QWidget parent 0 MainWindow private per Flag for the feedback option bool feedback pee Flag for the motion simulation bool motionSimulation pee String to hold connection information 194 string platformT ype pee Pointer to a mediator object F Mediator m px Pointer to a ui object Ui MainWindow ui string Q2S QString qs QString S2Q string s void checkConnections void Feedback string s void enableManualControls bool private slots void on_FSXDisconnect_clicked void on_FSXConnect_clicked void on_extend_clicked void on_yawRight_released void on_yawRight_pressed void on_yawLeft_released void on_yawLeft_pressed void on_swayRight_released void on_swayRight_pressed void on_swayLeft_released void on_swayLeft_pressed void on_surgeBackward_released void on_surgeBackward_pressed void on_surgeForward_re
73. Good Average Poor Copies of FSX available in the department 50 The Flight Gear v2 0 flight simulator was taken into consideration because of its compatibility with Mathworks Simulink a simulation environment that allows for rapid prototyping However after a preliminary evaluation it was determined that this flight simulation environment was not suited to drive a full motion flight simulator Indeed the quality of its physics is questionable There are no development tools available and very little documentation which makes it a very unreliable product On the other hand both X Plane 9 and FSX represent acceptable flight simulation software for motion simulation Indeed they both deliver realistic physics of flight combined with good graphics quality Both those products come with development tools that allow for the writing of plug ins to communicate with the simulation environment Microsoft s product features very good documentation and easy access to the relevant flight information Indeed FSX offers a complete software development kit SDK to easily interface with the simulation environment This SDK SimConnect allows for the abstraction of many of the communications and timing issues associated with the flight simulation interfacing This is also the case for X Plane 9 where users can use similar libraries to easily interface with the flight simulation environment However X Plane features inferior libraries with
74. Operation Design Sub Objectives The GUI and underlying program should never enter a state where they become stuck and cannot recover Furthermore the GUI should be intuitive and easily understood without any explanation This includes the use of significant terms which must be used to describe the functions of each GUI controls As a user controls the flight simulation he must also remain aware of the system s state For this reason the GUI should provide feedback on the operation of program and the simulator Finally the GUI is an abstraction of the complicated function calls and code Essentially the GUI should allow for the straightforward operation of the simulator without a need to understand any of the underlying systems 3 2 5 Maintainability Maintainability represents the ease with which a software system or component can be modified to correct faults improve performance or other attributes or adapt to a changed environment 16 Closely related to portability this design objective is aimed at creating a software system that will be easy to maintain As it is intended that the software produced in this study will be used by others in the future it is important that it is properly commented and documented Proper software documentation includes sections on the characteristics of the 40 system its architecture and design the code algorithms and APIs used The specific design sub objectives relating to maintai
75. Run 2 Parameters Scaling Limiting TUA Misc Parameters ax 1 wn 5 1 ay I n 8 0 1 _ _ Overshooting oa warea Management p 0 6 On 1 q 0 6 On 1 r 1 Tilt af i oo 5 4 5 deges Tilt Coordination Gain Table 20 Pilot Observations Test Run 2 Motion Channel Observations bon Almost no overshooting present The magnitude of accelerations is sufficient Tilt coordination is present and proportional Tilt Coordination Tilt coordination sometimes lags slightly behind on pitch motions Rotational motion is proportional for both Rotational pitch and roll motion However pitch motions seem to be underpowered The response variables for this test run can be found in Table 21 The variations of relevant response variables between test run 1 and 2 which is the information of interest can 126 be found in Table 22 The response variables of this test run were subtracted from those of the first test run As such a positive variation indicates that there has been an improvement on the motion response for a specific channel Table 21 Response Variables for Each Motion Channels Test Run 2 Response Variables for Each Motion Channels Relevance Translational Tilt Coordination Angular Flight Profile a mes ay mes a mes p deges q deges r deges Continuous 0 7149 0 0589 0 2820 0 3775 2 0873
76. SIMCONNECT_RECV_SIMOBJECT_DATA pObjData SIMCONNECT_RECV_SIMOBJECT_DATA pData Cast pData into o bject data switch pObjData gt dwRequestID Obtain the id of the simObject case REQUEST_PDR int count 0 StructDatum pS StructDatum amp pObjData gt dwData pe There can be a minimum of and a maximum of maxReturnedItems in the StructDatum structure The actual number returned will be held in the dwDefineCount parameter if paused while count amp lt int pObjData gt dwDefineCount For each element we will extract the data and send over UDP ex Switch statement on the data received and the appropriate action In this case an UDP broadcast on select port switch pS gt datum count id case DATA_X_ACCELERATION Cast into double and send value over UDP thisFS XConnect gt udpSend SendDouble thisFS XConnect gt udpDataPorts 0 double pS gt datum count value break case DATA_Y_ACCELERATION Cast into double and send value over UDP Polarity is inverted for the y axis since FSX uses different convention than the SAE right hand rule thisFS XConnect gt udpSend SendDouble thisFS XConnect gt udpDataPorts 1 double pS gt datum count value break case DATA_Z_ACCELERATION Cast into double and send value over UDP thisFS XConnect gt udpSend SendDouble thisFS XConnect gt udpDataPorts 2 double pS gt datum count value break case DATA_P_ANGULAR_VELOCITY Cast into d
77. THE DESIGN IMPLEMENTATION AND ANALYSIS OF A LOW COST FLIGHT SIMULATOR USING RMC S 3DOF MOTION PLATFORM LE DESIGN L IMPLEMENTATION ET L ANALYSE D UN SIMULATEUR DE VOL A FAIBLE COUT UTILISANT LA PLATFORME DE MOUVEMENT A TROIS DEGRES DE LIBERTE DU CMR A Thesis Submitted to the Faculty of the Royal Military College of Canada by Jean Le Bouthillier RMC Second Lieutenant In Partial Fulfillment of the Requirements for the Degree of Master of Science in Computer Science October 2010 This thesis may be used within the Department of National Defence but copyright for open publication remains the property of the author ii ROYAL MILITARY COLLEGE OF CANADA COLLEGE MILITAIRE ROYAL DU CANADA DIVISION OF GRADUATE STUDIES AND RESEARCH DIVISION DES ETUDES SUPERIEURES ET DE LA RECHERCHE This is to certify that the thesis prepared by Ceci certifie que la th se r dig e par Jean Le Bouthillier entitled intitul e The design implementation and analysis of a low cost flight simulator using RMC s 3DOF motion platform Le d sign l impl mentation et l analyse d un simulateur de vol faible cout utilisant la platforme de mouvement trois degr s de libert du CMR complies with the Royal Military College of Canada regulations and that it meets the accepted standards of the Graduate School with respect to quality satisfait aux r glements du Coll ge militaire royal du Canada et qu elle respecte les normes
78. Table 14 Margin of Error for Three Identical Runs Unsmoothed Signals eee 109 Table 15 Margin of Error for Three Identical Runs Smoothed Signals 112 Table 16 Classical WF algorithm Configuration Test Run 1 114 Table 17 Pilot Observations Test Run 1 115 Table 18 Response Variables for Each Motion Channels Test Run 1 115 XV Table 19 Classical WF algorithm Configuration Test Run 2 125 Table 20 Pilot Observations Test Run 2 22 nn homard tetas bettas 125 Table 21 Response Variables for Each Motion Channels Test Run 2 126 Table 22 Response Variable Variations between Tests Run 1 and 2 127 Table 23 Classical WF algorithm Configuration Test Run 3 130 Table 24 Pilot Observations Test Run 3 sun mana dre dns 131 Table 25 Response Variable Variations between Tests Run 2 and 3 132 Table 26 Pilot Validation Exercise Procedure 134 Table 27 Average Pilot Ratings for Flight Simulation Hardware 0 ee ceeeeeeeeeseeceeeeeeeeees 136 Table 28 Summary of Pilot Complaints for the Flight Simulator Hardware 137 Table 29 Average Pilot Ratings for Motion Simulation cee eeeescecsseceseeeseeesneeesaeeneeees 138 Table 30 Average Pilot Ratings for Motion Impact and Flight Training Potential 139 Table 31 Pilot Suggestions for Potential Training Use oe eee eeeeescecnec
79. The study of an acceleration feedback mechanism as in 32 shows that the classical WF algorithm can also be studied and included into hardware in the loop formulations In light of the time consuming and repetitive task of algorithm tuning it seems that the formulation of an auto tuning method would be a major breakthrough Such a method if sufficiently reliable could also reduce the need for pilot evaluation exercises in establishing a flight stmulator s performance 6 2 2 Flight Simulator Hardware Improvement Improvement of system performance through the acquisition of better hardware components can become a never ending process In the case of flight simulation this process ends when a complete recreation of motions audiovisuals and tactile aspects of flight has been achieved If economical constraints are an important consideration it will not be possible to design or acquire the advanced and costly hardware necessary to fully achieve this goal The 145 comments formulated in this section reflect the observations of pilots and the author in order to improve the simulator while maintaining the low cost philosophy of this study As identified during the pilot testing phase the head tracking system is perhaps one of the most important component in the system However due to the nature of the TrackIR 5 pilots could only move their head within a 120 degree cone in front of the sensor This led to a non proportional head tracking w
80. UEST_ID ex An enumeration of the simulation variables we are interested in static enum DATA_NAMES DATA_X_ACCELERATION DATA_Y_ACCELERATION DATA_Z_ACCELERATION DATA_P_ANGULAR_VELOCITY DATA_Q_ANGULAR_VELOCITY DATA_R_ANGULAR_VELOCITY DATA_P_AIRCRAFT_PITCH DATA_B_AIRCRAFT_BANK DATA_NAMES This method must be static since it is used as a callback Will receive this pointer in the pContext parameter static void CALLBACK MyDispatchProcPDR SIMCONNECT_RECV pData DWORD cbData void pContext public FSXConnect FSXConnect constructor void FSXConnectRequest Method requesting a connection to FSX void FSXDisconnect Method disconnecting from FSX void simulation Method performing one simulation step bool isConnected Method returning the connection status FSXConnect endif FSX_CONNECT_H FSXConnect cpp pe FS XConnect cpp IDE Visual C 2008 Express Edition author 2Lt Jean Le Bouthillier Latest Revision 20 Sept 2010 ek C SimConnect Data Request for FAA and wAA inputs to the classical washout algorithm ek fA A ax ay az wAA p q r kk Description After a flight has loaded the program will request the accelerations of the aircraft referenced to its center of mass and its angular velocities Anytime the values change in the flight simulator the program will return the accelerations in meters s and the angular ve
81. accelerations into Be the simulator low frequency Euler angles The low frequency Euler angles Be represent the angles at which the simulator cab must be positioned to effectively simulate continuous translational accelerations They are obtained in the tilt coordination channel from the low frequency constant accelerations of the aircraft Due to the limited translational motion envelope of motion platforms it is hard to simulate continuous translational motion in the x and y axis as sensed by an aircraft pilot However by slightly tilting the simulator cab with appropriate visual cues it is possible to orient to the local gravity vector in such a way that the simulator generates a translational motion feeling for the pilot This process called tilt coordination is commonly used in flight simulators 8 When using tilt coordination to simulate translational accelerations it is important to limit the angular velocity at which the simulator cab tilts Indeed as reported by 3 vestibular motion detection threshold values range between 2 6 to 3 6 deges Limiting the simulator cabin tilt coordination rates below these threshold values is paramount to achieve 19 realistic simulation In the example shown in Figure 5 the simulator cab is slowly tilted about its y axis until it reaches approximately six degrees Then the pilot experiences an approximate translational acceleration 9 81 m s7 sin 6 1m s The a component of the motion
82. accept es par la Facult des tudes sup rieures quant la qualit for the degree of pour le dipl me de MASTERS OF SCIENCE IN COMPUTER SCIENCE MAITRISE ES SCIENCES EN INFORMATIQUE Signed by the final examining committee Sign par les membres du comit examinateur de la soutenance de th se Chair Pr sident External Examiner Examinateur externe Int External Examiner Examinateur interne externe Internal Examiner Examinateur interne Supervisors Directeurs de th se Approved by the Head of Department Approuv par le Directeur du D partement Date To the Librarian This thesis is not to be regarded as classified Au Biblioth caire Cette th se n est pas consid r e comme publication restreinte Supervisors Directeurs de th se iii ACKNOWLEDGEMENTS First I would like to thank my parents for the great qualities they haved instilled in me Their unwavering support guidance and encouragements throughout the years in all my endeavours have been invaluable They instilled in me the values of hardwork and discipline which have always served me well My success is but an extension of their remarkable parenthood and dedication I would like to express my deepest thanks to Dr Yawei Liang and Dr Paul Allard for giving me the opportunity to study the challenging problem of flight simulation Their knowledge guidance and critical outlook allowed me to structure and improve
83. ailable as part of the G940 flight system used in this study 146 Finally the third area of complaints from pilots targeted the joystick controller which was deemed to have a very limited range when compared to an actual aircraft stick As such the use of a joystick with a wider motion envelope could improve the flight simulator control Such a joystick could potentially be found commercially at customized flight simulation businesses It can be pointed out that many of these hardware limitations experienced will be removed as hardware components continue to be improved by the industry 6 2 3 Future Applications Although an acceptable level of flight simulation has been achieved with RMC s 3DOF it remains an experimental flight simulator designed to study the problem of flight simulation and the possibilities associated with RMC s 3DOF motion platform One of the next steps consists in determining a potential realistic application for this kind of flight simulator Then a careful design could be made to be specifically oriented to this task As suggested by many pilots this platform could potentially be used to train new pilots to accomplish manoeuvres before doing them in the real aircraft Most pilots with training experience also saw great potential for such a flight simulator as an inexpensive IFR trainer Most pilots felt that the motion experienced would contribute greatly to IFR scenarios where the visibility is inexistent In such ca
84. analysis is performed for these components because they bring the greatest contribution to the flight simulation system The table used for comparison and for the cost benefit analysis for each of these components can be found in Section 3 4 of this chapter where specific flight simulator components are discussed 3 2 3 Software Portability From the very beginning of this study issues of software portability were a primary concem in the design of RMC s 3DOF flight simulator Software portability one of the key concepts of high level programming is essential to create a system that can be easily adapted to other simulation hardware Specifically software portability refers to the features allowing the 37 successful reuse of code when moving software from one environment to another 15 Indeed it is intended that the software used for this study be easily transferable to other motion platforms in order to conduct other research work Specifically the software produced in this study is destined to be adapted and used for the operation of RMC s 6DOF motion platform Software portability is closely related to cost effectiveness because the development of portable software targeting many platforms with the same application saves time and resources Several factors are considered to contribute to the portability of software and are used in this study to determine how to best select and implement the software components driving the flight si
85. anced debriefing capabilities Flight simulators also reduce the risk of aircraft training accidents as a large portion of training can be conducted in the simulated environment 1 2 3 Modern Flight Simulation Typical modern full motion simulators used in pilot training such as the ones sold by CAE Inc use electrically or hydraulically driven six degree of freedom 6DOF motion platforms and retail for several million dollars 2 Although it has been demonstrated that in most cases a three degree of freedom 3DOF motion platform is capable of producing a simulation quality comparable to that of a 6DOF motion platform 3 there are very few commercially available 3DOF flight simulators Most flight simulators include several subsystems that approximate sensations felt in a real aircraft during flight The computer generated images the visual display the sound system the motion simulation system and other enhancements such as realistic controls and cockpit instrumentation all contribute to the quality and realism of the simulation In this study the subsystem of most interest is a 3DOF motion platform its associated motion cueing algorithm and the control system that lead to a more realistic sensation of flying Modern simulators usually employ electrically or hydraulically driven motion platform to generate motion All of these motion platforms have limited motion envelopes compared to the motion envelope of a real aircraft The phys
86. and is complete when the simulator operator is able to conduct the use cases with ease 70 4 2 2 Functional Approach to Implementation As stated in the design objectives of Chapter 3 it is important to design software components that are simple to use maintain and that are portable to other applications such as the control of RMC s 6DOF motion platform As such a functional approach has been taken for the implementation of the software components in this study A functional approach is the term utilized here to describe the implementation of the required software components in a way that ensures that each software components of the system serves a unique and independent purpose from the others The functional approach described here is in line with object oriented best practices 18 The idea is to separate the software into several simpler parts components to accomplish each of the main functionality of the system In addition the goal is to implement the system in a way that separates software components from each other as much as possible As opposed to implementing all the main functionality of the system together they are separated into different software components that only perform specific and unique actions To accomplish this the C object oriented programming language is very useful as it is already designed to facilitate this kind of approach in the implementation This approach provides advantages later in the software life c
87. ant is debriefed and additional feedback is collected trough a discussion 135 More specifically the rating criteria were explained to the pilots prior to performing the simulation This information is covered to ensure that pilots have an idea of the important criteria for the evaluation during their flights Table 26 summarizes the procedure which was followed for each test pilot during the pilot validation phase The evaluation featured the conduct of an airport circuit flight manoeuvre with and without motion simulation The pilot validation exercise focused on collecting pilot ratings on three topics First the pilots were asked to rate unique aspects of hardware used on RMC s 3DOF flight simulator i e the controls used the head tracking etc A scale from 1 to 5 1 representing the most positive rating and 5 representing the worst rating was used to rate each element of the simulator s hardware Then the pilots are asked to rate the quality of the motion based on seven criteria developed at the UTIAS by 4 during a pilot validation exercise These criteria were selected as they are used in the literature to characterize motion quality and because they are in line with the information of interest for this study Then the pilots were asked to rate the impact of motion simulation by rating their perceived value of motion simulation by comparing the motionless run to the full motion one Finally the pilots were asked to commen
88. ar PortNumber bool SendDatagram PortNumber const void size_t bool ReceiveDatagram void size_t long bool SendDouble PortNumber double bool ReceiveDouble double long bool SocketCleanup void ReceiveDoubleArray double size_t long bool UdpSend SocketCleanup void UdpSend UdpReceive UdpReceive Figure 49 Classes Required for the Implementation of UDP The UdpSend class is used to send single double values over UDP datagram while the UdpReceive class is used to receive single double values or double arrays of customizable size from UDP datagram The Udp class is a convenience class used by both UdpSend and UdpReceive for the initialization and status checking of Winsock Since the relevant simulation data is only obtained from FSX when it has changed in the flight simulation the accelerations and velocities are not considered to be outputted from FSX as arrays but rather as single double values Because of this it is convenient to be able to send simulation data one double at a time using the SendDouble method in the UdpSend class Once the data has been processed by the 85 classical WF algorithm inside Simulink the lengths velocities and accelerations of the motion platform actuators are returned The lengths velocities and accelerations are broadcasted from the Simulink environment as arrays over UDP using Simulink s UDP Send block As such
89. ar signals values DataValues var signals dimensions DimValues Parameters eC 2 Data gyz Turn Sample time 0 02 N Interpolate data WAA E Enable zero crossing detection Form output after final data value by Setting to zero Loc cane Figure 80 Inside View of Flight Playback block Annex H Pilot Rating Questionnaire 168 169 RMC s 3DOF Flight Simulator Evaluation Pilot Information Please answer the following questions about your flying experience 1 What is your total flight time hrs 2 Please list the main aircraft s on which you have acquired this flight time 3 Please list your pilot qualifications i e private commercial instructor etc 4 Have you ever flown the Cessna 172 Yes No If so how many hours do you have on the C 172 hrs Flight Simulation Hardware Please indicate your assessment of the simulation hardware based on the following rating scales Attribute Rating Response Unrealistic feeling Excellent creates a 8 Visual de 5 detrimental to the very realistic feeling experience Unrealistic feeling Excellent creates a 8 Head tracking ee 5 detrimental to the very realistic feeling experience Excellent response to Very poor control Controls inputs provides l 5 detrimental to the excellent con
90. ary airplane functions such as flaps and trim The seat and controls must be installed onto the platform in a way that allows passengers of different dimensions to fit comfortably Ideally the seat or controls positions are adjusted to suit the preference of the pilot Figure 34 shows a tri dimensional concept drawing for the positioning of the pilot seat and controls Figure 34 Concept Drawing of the Pilot Seat and Controls In this design it is envisioned that the pilot seat remains in a central position to respect the center of gravity constraints The seat can be reclined back and forth but cannot move up or down Consequently it is the controls position that is adjustable to best suit the pilot s preferences The control column with the joystick on top can be lowered down to allow the pilot to board and fasten his seat belt Once the pilot is in position the control column is raised 60 between the legs of the pilot The height of the joystick can be changed by collapsing the main control column up or down and the distance of the joystick can be modified by changing its position on the upper joystick receiving frame The height of the throttle is fixed but it can be moved forward and backward along a series of pre drilled holes at the top of the support box Figure 35 shows a close up of the adjustments mechanisms Figure 35 Close up on Control Adjustment Mechanisms Concept This design could also include an extension t
91. ary to ensure feasibility of design and to allow for the selection of appropriate components An objective tree 13 is a useful and clear method for visualizing an objective statement It is a way to display all objectives and sub objectives of a given project into their hierarchical levels A diagram of the top level objective tree used in the design of RMC s 3DOF motion simulator can be seen in Figure 15 Note that each of the six objectives in this tree branch into sub objectives that will be discussed in the sections to follow RMC s 3 DOF Flight Simulator Realistic Simulation Cost Effectiveness Portability Maintainability Figure 15 Objective Tree for RMC s 3DOF Flight Simulator 33 The specific design objectives of RMC s 3DOF flight simulator were derived from discussions with thesis surpervisors about the desirable characteristics of a flight simulator for this study Driven by those characteristics the design of RMC s 3DOF flight simulator was made while keeping the six main design objectives of Figure 15 in mind One of the primary design objectives is to create a flight simulator that is capable of approximating a real flight environment with the highest possible quality given the limited resources available Additionally the simulator must meet other design objectives that pertained to cost effectiveness software portability maintainability ease of use and safety In the sections to follow each of the design obje
92. ating the flight simulator The exercise will take place in the Girouard 333 laboratory where the flight simulator is located This is where flight simulation will take place and where the participants will be asked to fill the rating questionnaire To prevent the risk of influencing the ratings and opinion of participants during the evaluation staff will not discuss or give their opinion on the fidelity of the flight simulator Participants will be asked to complete the rating questionnaire before discussion the experience with staff To prevent a consensus when multiple participants are present for one session the participants will be asked to wait until all participants have completed the questionnaire to discuss the rating 4 Participants Involved in the Study a Describe any special characteristics of potential participants in this study e g race gender mental and physical disabilities age range Are there participant exclusion and or inclusion criteria If so describe and provide justification Participants in this study will be individual with flight experience of various levels There will be no restriction on age or gender The study will be conducted in English b How many participants are expected to be involved in this study Around 10 participants are expected to be involved c From what source will the potential participants be recruited RMC undergraduate or graduate classes _ Other RMC sou
93. ational channel In this case it is observed that overshooting has a dampening effect on the correct motion cues sent to the motion platform Indeed if two opposing commands are sent to the motion platform within a short time interval the commands cancel each other and the motion platform remains still In this case the overshoot cannot be attributed to the filter equation as it is of the 1 order These overshoots are caused by the fact that turn entry and turn exit rotational velocities typically exhibit higher frequency on their falling edge This is a normal phenomenon that is usually dampened by the tilt coordination channel In the absence of sufficient tilt coordination this phenomenon becomes a motion cueing problem The amplitude of simulated angular rates seems to be slightly low as it visually hovers around 20 40 of turn profile angular rates for a scaling of 0 5 This suggests that scaling coefficients should be increased on this channel As such the rotational motion channel scaling coefficients are increased for the next test run Having evaluated the optimal configuration of the classical WF algorithm as per the literature an improved configuration is formulated for the second test run This formulation is 119 based on pilot observations combined with an analysis of response variables and relevant motion plots Details of this new formulation and its performance are available in Section 5 3 4 p deg s 6 N
94. atum ID d pS gt datum count id To catch errors break switch count while MI if else if paused for int i 0 i amp lt 8 i thisFSXConnect gt udpSend SendDouble thisFS XConnect gt udpDataPorts i zero 193 for else if break case default break switch break case pe Tf the simulator sends the event quit we stop this program a case SIMCONNECT_RECV_ID_QUIT thisFSXConnect gt quit true break case default This happens when the program has established a connection with the simulator but there is no flight underway printf nNo simulation running or Unknown Event ID d pData gt dwID break switch MyDispatchProcPDR pe Tn the FSXConnectRequest method a connection to SimConnect is opened and the variables we will be looking at are subscribed to A simulation start event is also requested as it will be used to know when to start listening for the variables we are interested in void FSXConnect FSXConnectRequest HRESULT hr Handle return object ex Tf there was a successful connection to FSX if SUCCEEDED SimConnect_Open amp hSimConnect Tagged Data NULL 0 0 0 connected true Set the connection status flag to true pe Set up the data definition ensuring that all the elements are in Float32 units to match the StructDatum structure The number of entries in the DEFINITION_PDR definition should be
95. ault those ports are 10101 to 10106 4 param pData SimConnect receive structure object param cbData callback data allocator param pContext pointer to the context structure void CALLBACK FSXConnect MyDispatchProcPDR SIMCONNECT_RECV pData DWORD cbData void pContext double zero 0 a value used to send zero data static bool paused flag used to keep the state of the pause unpaused event from FSX pee Cast pContext to FSXConnect This is the instance of FSXConnect that caused this callback to be called thisFS XConnect can be used like this in any non static method FSXConnect thisFSXConnect FSXConnect pContext switch pData gt dwID Switch on SimConnect DWord ID pe Tf we receive some event from the simulator on the events we subscribed to perform actions case SIMCONNECT_RECV_ID_EVENT 191 SIMCONNECT_RECV_EVENT evt SIMCONNECT_RECV_EVENT pData Cast the data into SimConnect object structure switch evt gt uEventID Check the event ID case EVENT_SIM_PAUSE If it is the pause event if evt gt dwData 1 Paused paused true HI if else if evt gt dwData 0 Unpaused paused false else if break default break switch break case 192 ee Tf we receive some data from the simulator on the data we tagged perform actions case SIMCONNECT_RECV_ID_SIMOBJECT_DATA In case we receive a Simobject Data simulation variable
96. button is also the only enabled button during flight simulation in order to draw the user s attention to it should there be the need to hastily stop the simulation This simple interface is highly abstracted from the complex operations required to perform the flight simulation 3 4 6 Programming Language The selection of a programming language is important to ensure the full integration of all of the key flight simulator s components In addition the use of a mainstream programming language is conducive to a high level of portability When looking carefully at the programming languages inherent to the key components used in the system it can be seen that C is the language of choice for the system Table 6 Programming Languages Compatibility for Key Components Programming Flight Simulator Component Languages Galil Motion Controller C C C VB Microsoft Flight Simulator X C C Graphical User Interface QT C C Visual Studio Professional C C C VB 2008 As key systems such as the Galil motion controller and FSX use C it is imperative that other systems be integrated using the same language The native language of key components of the flight simulation system can be seen in Table 6 In addition to providing a single language for integrating all of the key components of the flight simulator C is a modern widely used and object oriented programming language Those features are conducive to high a
97. calculation of actuator s extensions l using Equation 4 4 where the Euclidian norm for the difference in position of the actuator attachments between the upper and base platform frames is taken The polarity of the z axis is kept for the z axis coordinate with the sign operation sgn It is necessary to use the sign operation since the sign is lost when taking the Euclidian norm to get the length l sgnz Lis Ais S1 Bis Lis Ais Sr Bis 4 4 The washout position of the motion platform is reached when the actuators are extended to half their maximum elongation at what is called the central position In other words the motion platform will always seek to return to this position For RMC s 3DOF this leaves an available distance of 0 20 meters for each actuator to move either up or down It is from this 83 position that the lengths calculated from 4 4 are applied As can be seen from Equation 4 5 and 4 6 the velocities v and accelerations a are derived from the lengths using its first and second derivative respectively Finally the lengths velocities and accelerations are limited according to SSSI s specification Table 1 using a limiter as described in Equation 2 8 d v t F 4 5 a t 2 4 6 4 2 6 Networking The implementation of the flight simulation system relies on the use of networking for communications with FSX and Simulink In addition network communications are used to communicate from t
98. ce of a cabin or not 35 Finally the simulation can be further enhanced by providing a realistic cabin environment and controls to the passenger Various systems are available and typically range from simple consumer level gaming joysticks throttles and rudders to fully replicated and functional aircraft cabins with simulated instrumentation panels Some of the most realistic controls can provide tactile sensation in the form of force feedback from the simulation environment depending on the situation For this study it is felt that at a minimum the controls necessary to operate an aircraft should be present along with a seat for the pilot to be properly seated A seat providing adjustments is desirable as it will allow individuals of different heights to be seated comfortably 3 2 2 Cost Effectiveness As resources allocated to this study are not unlimited cost effectiveness is an important objective driving the development of RMC s 3DOF flight simulator Furthermore as highlighted in Section 1 4 an objective of this research is to design a flight simulator at a relatively low cost while maintaining a high level of realism The conduct of a cost benefit analysis is an important step in achieving cost effectiveness Indeed the process of cost benefit analysis consists of comparing the amount of money spent to the benefits that will be achieved through the spending of this money to acquire a specific equipment or software This can onl
99. cessful this gt SocketCleanup If it wasnt close the current socket return false Return failure Wif return true Everything went well return a success Initialize pe This method is used to send a UDP Datagram over the network 4 param nDestPort the destination port for this datagram param nData a pointer to the memory block containing the data to be sent param nDataSize number of bytes of data to be sent return amp lt true amp gt if the the number of bytes sent matches the number of bytes intended to send amp lt false amp gt otherwise bool UdpSend SendDatagram PortNumber nDestPort const void nData size_t nDataSize ex Prepare the destination address sockaddr_in destAddr this gt mDestAddr Destination IP was set in Initialize destAddr sin_port htons nDestPort Change the destination port htons converts from host to network format ex Send the UDP datagram to the network The Datagram will be assigned a source TP address from the egress network interface The source port is chosen 223 224 randomly by Windows int sendToResult sendto this gt mSocket Send using this socket source IP and port char nData Send bytes from this memory block nDataSize Number of bytes to send in this datagram 0 No flags are required sockaddr amp destAddr Send to this IP address and port destination IP address and port sizeof destAddr Th
100. channel of the classical WF algorithm allows for an approximation of the desired translational effects in the x and y axis by tilting the simulator s cabin towards these axes as described in Section 2 2 3 In addition the absence of a yaw motion in the simulation of fixed wing aircrafts is not a significant problem as there is usually little yaw motion in normal flight As such for the simulation of low performance aircraft the performance of 3DOF motion platforms is comparable to that of 6DOF motion platforms 3 2 4 1 Translational Motion Channel As suggested by 11 the optimal scaling factor for most commercial 6DOF simulators is of 0 5 However depending on the behaviour of the translational channel high pass filter utilized as well as RMC s 3DOF specific motion envelope this coefficient may need to be modified to suit this study An appropriate scaling factor will be chosen during the experimental phase of this study in which fine tuning of the algorithm will occur According to the platform manufacturer 24 SSSI 6 the z axis accelerations of RMC s 3DOF are limited to a maximum of 7 85 mes due to the physical limitations of the platform RMC s 3DOF motion platform is only capable of translational displacements in the z axis For this reason the x and y translational accelerations are not required in the calculation of the simulator translational displacement Si However the x and y translational accelerations are
101. chnique inspired from 8 is used in this study for RMC s 3DOF Using the simulator displacement S from the translational motion channel in this case only in the z axis and the Euler rotation matrix from the tilt coordination and rotational motion channel it is possible to calculate the position of the upper motion platform frame Locations Fs Fr 0 0 0 As Bs 0 49 0 0 As Bsz 0 39 0 45 0 As3 Bs3 0 39 0 45 0 Les Figure 46 Geometrical Layout of the Actuators Top Frame View Figure 46 shows the geometrical layout of the motion platform base where the positions of the actuator attachments are denoted in by As and represented in space using the Cartesian system in meters In the case of RMC s 3DOF the positions of the actuator attachments on the upper motion platform frame denoted by Bs are the same as for the motion platform base Knowing the geometrical layout of the motion platform allows for the superimposition of the 82 upper and base platform frames in order to derive the actuator s lengths In Figure 47 an example of such a superimposition can be seen where the upper motion platform frame is undergoing a pitch backwards combined with a heave up Pitch Backwards 30 7 Heave up 10 cm 27 Actuator Length Actuator Length 7 B Sl Upper Frame Base Frame Figure 47 Superimposition to Calculate Actuator Extensions This superimposition technique allows for the
102. cket is an IP and port pair and must be unique on the system SOCKET mSocket Ja A structure containing the destination address template Tt is a template since the Port field is not set only the IP address is set The port value is set in SendDouble SendFloat for each invocation sockaddr_in mDestAddr void SocketCleanup bool SendDatagram PortNumber nDestPort const void nData size_t nDataSize public UdpSend UdpSend bool Initialize const char nDestlp bool SendDouble PortNumber nPort double nValue endif UdpSend cpp ee UdpSend cpp IDE Visual C 2008 Express Edition author Martin Vidovic Modified by 2Lt Jean Le Bouthillier Latest Revision 13 November 2009 ek Description This class is used to send UDP datagram over the network Tt is currently only capable of sending double values over UDP datagram but it could be modified to send any kind of data over the network k include UdpSend h ee UdpSend Constructor UdpSend UdpSend Initialize the mSocket to an initial value this gt mSocket INVALID_SOCKET UdpSend 222 UdpSend Destructor UdpSend UdpSend Release networking resources this gt SocketCleanup UdpSend pe This method is used to close the current network socket void UdpSend SocketCleanup Close the socket if this gt mSocket INVALID_SOCKET closesocket this
103. classical WF algorithm is studied and optimized as a whole by playing back a series of pre recorded flight profiles through the system Then by comparing recorded 102 motions to those of the platform the fidelity of motion simulation can be studied and optimized in order to take full advantage of RMC s 3DOF motion platform capabilities The Cessna 172 SP Skyhawk C 172 is selected as the aircraft for which we will attempt to simulate motion The C 172 is selected because it is a small and very popular aircraft in the civilian aviation world In addition it is widely used as a pilot trainer in most civilian flight schools 27 and some military establishments 28 For this reason it is interesting to study the degree of realism that can be achieved using this airplane and RMC s 3DOF motion platform x Fuel and Payload Information Pilot and copilot weight 170 lbs each Y Fuel 53 U S Gallons 318 Ibs a Airplane gross weight at takeoff 2308 Ibs Key Geometrical References m Reference Datum Position 0 0 0 CG Position 0 91 0 0 A Pilot 0 91 0 46 0 and Copilot 0 91 0 46 0 Figure 56 FSX s C 172 Key Information and Geometrical References Positive results would mean that the system under study could be of use for the simulation of such airplanes Finally because of its popularity it is also very easy to find qualified pilots for a validation exercise of the flight simulator Figure 56 shows
104. compared to those of the second test run in order to objectively corroborate the improvement noted by pilot testing It was found that the variance of response variables was very small for most motion channel It was also observed that where gains of performance were present a decrease in performance was also found on related channels These observations indicate that the system is reaching its limits were improvements on channels are no longer possible without an associated loss on a related channel It is felt that the current solution is well balanced between each motion channel and that the final result is quite realistic Table 25 shows the response variable variations between the second and third test run As can be seen from Table 25 the only major gain of performance can be found on the x axis tilt coordination channel a channel for the continuous climb manoeuvre This gain of performance is associated with a loss of performance on pitch motion q channel for this 132 manoeuvre Other response variables of this test run remain relatively stable with slight variations Table 25 Response Variable Variations between Tests Run 2 and 3 Response Variable Variations for Each Motion Channels Relevance Translational Tilt Coordination Angular Flight Profile 2 2 2 1 gt 1 ay mes ay mes az mes p deges q deges r deges A 0 1970 0 0006 0 1576 Climb SERA EE 0 0777
105. const char nSourcelp PortNumber nSourcePort Check for invalid state if this gt mSocket INVALID_SOCKET return false Mif if nSourcelp NULL return false Mif ex Creating a network socket AF INET We are using IP Internet family of protocols SOCK_DGRAM We are using datagram style communication IPPROTO_UDP We are using UDP protocol this gt mSocket socket AF_INET SOCK_DGRAM IPPROTO_UDP if this gt mSocket INVALID_SOCKET Check if socket was successfully created return false Mif ex Set socket SO_REUSEADDR option This makes reuse of the socket possible int solEnable 1 int setSockOptResult setsockopt this gt mSocket Socket to change options on SOL_SOCKET Level of options We are changing options on the socket level SO_REUSEADDR The option ID char amp solEnable New option value sizeof solEnable Size of the value data type if setSockOptResult 0 If setting of the socket options failed return false Mif ex Creating source address structure Bind our socket to the source port number and address This is needed so that we receive only packets sent to this address ZeroMemory amp this gt mSourceAddr sizeof this gt mSourceAddr Initialize sourceAddr stucture this gt mSourceAddr sin_family AF_INET IP family of address ex Set port to nPort since we are sending _from_ this port Use htons to convert f
106. ct to F7 Smoothed point of a dataset at index j Transformation from angular velocity to Euler angle rates Individual actuator velocity Aircraft angular rates p q r Simulator angular rates p q r Standard deviation Arithmetic mean Filter damping ratio Filter break frequency CHAPTER 1 INTRODUCTION 1 1 Introduction In this chapter the topic of flight simulation from an historical and practical point of view is introduced to the reader Then the motion simulation hardware that makes this study possible is discussed Finally the motivation objectives and structure of this study are discussed to provide the reader with an understanding of the context leading to the conduct of this work 12 Flight Simulation This section provides a short introduction to the topic of flight simulation from an historical and practical perspective Events leading to the development of flight simulation are briefly discussed Then the advantages of flight simulation which drove the development of flight simulation to its modern state are summarized Finally the elements of modern day flight simulation and the need for motion cueing algorithm are introduced 12 1 Flight Simulation History In the early days of flying teaching new pilots was an activity that was both costly and dangerous This was in part due to the absence of training techniques other than progressive flights that would take place in actual aircraft and eventually bring a s
107. ctions void enableManualControls bool void metaObject const QMetaObject on_FSXConnect_clicked void on_FSXDisconnect_clicked void on_begin_button_clicked void on_center_clicked void on_checkBox_stateChanged int void on_clearButton_clicked void on_connectButton_clicked void on_erasePosition_clicked void on_extend_clicked void on_heaveDown_pressed void on_heaveDown_released void on_heaveUp_pressed void on_heaveUp_released void on_midown_pressed void on_midown_released void on_mtup_pressed void on_miup_released void on_m2down_pressed void on_m2down_released void on_m2up_pressed void on_m2up_released void on_m3down_pressed void on_m3down_released void on_mSup_pressed void on_m3up_released void on_m4down_pressed void on_m down_released void on_m4up_pressed void on_m4up_released void on_mSdown_pressed void on_mSdown_released void on_mSup_pressed void on_mSup_released void on_m6down_pressed void on_m6down_released void on_m6up_pressed void on_m6up_released void on_markPositionParked_clicked void on_park_clicked void on_pitchBackward_pressed void on_pitchBackward_released void on_pitchForward_pres
108. ctives and constraints and having selected the system s components the implementation of the system can begin In the implementation phase the software components are created in accordance with the conclusions drawn in the design phase The most significant part of the implementation in this study is the software implementation The software implementation follows a methodology that ensures the design is respected The installation of the hardware components onto the motion platform is also undertaken during the implementation phase In this chapter the implementation of the software components required to enable flight simulation will be discussed More specifically the implementation process and decisions taken during this process will be explained The intent is to provide the reader with an understanding of the work conducted during the software implementation phase An explanation of the implementation philosophy will be provided as well as a discussion on the implementation of specific components of the system Next the networking elements used in the flight simulation system will be explained to further expand the reader s understanding of the system References to the annexes where in some cases more specific implementation details are shown will also be given throughout this chapter Then the installation of physical components onto the motion 67 platform will be detailed Finally the validation of both the software and hardwa
109. ctives are discussed 3 2 1 Achieving a Realistic Simulation One of the primary objectives driving the design of RMC s 3DOF flight simulator is the achievement of a realistic simulation It is of importance to generate motion visual and audio signals that are conducive to a high degree of realism Indeed the generation of visual stimuli is considered to have the greatest effect on the overall realism of the simulation along with motion stimuli 14 In designing a flight simulator it is possible to spend thousands of dollars on systems that will generate the aforementioned stimuli to various degrees of quality However regardless of the quality of the systems used in the flight simulator the most fundamental element remains the virtual environment Great care must be taken in selecting flight simulator software that creates a virtual environment offering high quality audio visuals generation as well as a realistic physics of flight Once a high quality virtual environment has been selected it is important to ensure that high quality motion simulation and audio visual systems bear fidelity to the virtual environment and effectively immerse the pilot inside the simulation Of particular importance to the motion simulation aspect of flight simulation is the proper motion cueing 34 Indeed it has been demonstrated that poorly cued motion is detrimental to flight simulation 4 The classical WF algorithm the most widely used algorithm in the fli
110. ctuators g gt command DC 235500 235500 235500 Set deceleration to 235500 7 85 m s s for all actuators g gt command PT 1 1 1 Register all actuator for on the fly absolute position move g gt command PA 0 0 0 Center the motion platform feedback Moving platform to central position Feedback message else if else if uInput park g gt command SP 2000 2000 2000 Set jog speed to 2000 for all actuators g gt command AC 235500 235500 235500 Set acceleration g gt command DC 235500 235500 235500 Set deceleration g gt command PT 1 1 1 Register all actuator for on the fly absolute position move g gt command BL 6000 6000 6000 Define backward limit for all actuators g gt command PA 6000 6000 6000 Park the simulator prevent_movement true When the platform is parking prevent other movements feedback Parking the platform Feedback message W else if else if uInput stopMovement g gt command STX YZ Stops any current movement of the motion platform feedback Movement Stopped Feedback message W else if else if uInput pitchForward g gt command JG speed speed speed Actuator 1 down actuators 2 and 3 up g gt command BGX YZ Begin motion on all actuators feedback Pitching forward Feedback message else if else if uInput pitchBackward g gt command JG speed
111. d of the details of implementing mathematical operations Simulink also leverages all the power of Mathwork s MATLAB a high level programming language and environment and is able to perform almost any mathematical operations faster than traditional programming languages 23 It also allows developers to quickly implement and test algorithms in a drag and drop fashion For these reasons and because of its availability to students at RMC it was selected for the implementation of the classical WF algorithm in this study Figure 44 shows the implementation of the classical WF algorithm in the Simulink environment In Figure 44 the top level view of the classical WF algorithm can be seen where the channels and other key components are denoted using different colors It includes the main components of the algorithm as well as some UDP networking components that will be discussed in the next section Simulink allows for a clean layout of the top level algorithm blocks while hiding many layers of underlying complexity Each Simulink block can be opened with a double click to reveal its internal components This is a useful feature which allows the designer 79 to quickly navigate from the top level of the algorithm all the way down to the details of the implementation Speed Displacement fAA Scaling and Limiting fAA high pass filter Actuator Extension Actustor Limiting Degrees to Radians AA low pass filter T
112. d true if platformType 6DOF If connected to the 6DOF Additional options ui gt m4up gt setEnabled true ui gt m4down gt setEnabled true ui gt mSup gt setEnabled true ui gt mSdown gt setEnabled true ui gt m6up gt setEnabled true ui gt m6down gt setEnabled true ui gt swayLeft gt setEnabled true ui gt swayRight gt setEnabled true ui gt surgeForward gt setEnabled true ui gt surgeBackward gt setEnabled true ui gt rollLeft gt setEnabled true ui gt rollRight gt setEnabled true MI if Mif else if b false if asked to disable manual controls ui gt speedSlider gt setDisabled true ui gt pitchForward gt setDisabled true ui gt pitchBackward gt setDisabled true ui gt rollLeft gt setDisabled true ui gt rollRight gt setDisabled true ui gt heaveUp gt setDisabled true ui gt heaveDown gt setDisabled true ui gt mlup gt setDisabled true ui gt m1down gt setDisabled true ui gt m2up gt setDisabled true ui gt m2down gt setDisabled true ui gt m3up gt setDisabled true ui gt m3down gt setDisabled true ui gt m4up gt setDisabled true ui gt m4down gt setDisabled true ui gt mSup gt setDisabled true ui gt m5down gt setDisabled true ui gt m6up gt setDisabled true ui gt m6down gt setDisabled true ui gt swayLeft gt setDisabled true 200 ui gt swayRight gt setDisabled true ui gt surgeForward gt setDisabled true ui gt sur
113. d with respect to the aircraft The direction of a positive rotation around each axis of the coordinate system is determined by using the right hand rule A simulator reference frame Fs is geometrically centered on the upper frame of the motion platform The location of Fs with regards to F is represented by the vector S which connects the origin of frame Fy to the origin of frame Fs Frame Fps the simulator pilot reference frame is located at the pilot s head between his left and right vestibular systems Furthermore Fps is parallel to Fs and both frame s origins are connected by Re Vector Rs represents the position of the pilot s head with regards to the simulator reference frame Fs Fpa the aircraft pilot reference frame is located and oriented exactly like Fp With respect to the cockpit In addition aircraft reference frame F shares the same orientation and relative cockpit location as the simulator reference frame Fs Frame F1 is required in order to compare the simulated aircraft motions to that of the motion platform in frame Fs In other words Ra and Rs are equal in magnitude and parrallel Finally aircraft reference frame Fg is fixed at the aircraft s center of gravity with its x axis pointing forward towards the nose and its z axis point downward with respect to the aircraft Vector Py connects the origins of frames Fg and F4 in space 2 3 Classical Washout Filter Algorithm The classical WF algorithm shown in Figure
114. ded Yes X No J If Yes provide Funding Source RMC DRDC Period of Funding Until December 2010 If No is funding being sought Yes No 179 Funding Source s Period of Funding 6 Has this application been submitted to another institution s Research Ethics Board Yes _ No X If Yes provide the name of the institution date of ethics review and the decision Attach a copy of the ethics clearance certificate if applicable B SUMMARY OF PROPOSED RESEARCH 1 Purpose and Rationale for Proposed Research Briefly describe the objectives of the proposed project and the hypothesis es research questions to be investigated Include the potential benefits and the rationale for the study or studies Include a copy of a research proposal 2 pages maximum where applicable The objective of the study is to determine the level of flight simulation fidelity which can be achieved by using RMC s 3 degree of freedom DOF motion platform as a flight simulator a low cost system Traditionally expensive 6 DOF motion platforms are utilized for flight simulation However as evidenced by 1 3 DOF motion platforms exhibit potential in producing acceptable levels of motion simulation It is theorized that the use of a 3 DOF motion platform can greatly reduce the cost of flight simulation while providing fidelity levels that are conducive to its use as a flight training aid within the Canadian Forces Flight simu
115. ded into several software components it is necessary to implement an effective way for each software components to interact with each other As will be seen in the next section the use of design patterns is an effective way to solve the communication problem caused by the separation of software functionalities into different software components 4 2 3 Mediator Design Pattern When attempting to design software components that are reusable and easily maintained it is important to ensure that there are no dependencies among them This can be easily 12 accomplished by separating the software components in accordance with their respective functionality In order to control this level of abstraction between software components it is necessary to use a software design pattern that is tailored for this kind of application where a many to many relationship naturally occurs between the system s software components The Mediator design pattern is well suited for this application because it abstracts the behaviour of the system s software components from each other 19 In other words when software components need to work together they must all go through the mediator that handles the components interactions The mediator is the only software component to know and understand the behaviour of all of the system s software components The main advantage of this design pattern is that the software components or objects in object oriented progra
116. dern and effective force For this reason the motivation for this study is to provide an evaluation of the flight simulation capabilities of RMC s 3DOF motion platform as a tool to create a flight simulation environment As there is currently no flight simulator at RMC this research serves as a reference work in the design of such systems at RMC The work and conclusions drawn in this study are relevant to the design and implementation of simulators using 6DOF motion platform technologies that RMC has recently acquired Finally as the research budgets allotted to motion simulation researches are limited it was also of great interest to investigate whether a realistic flight simulator can be designed and implemented using low cost and commercially available components and software 15 Aim The first aim of this research is to create a low cost flight simulator using RMC s 3DOF motion platform The design and implementation of this flight simulator is made in order to obtain a realistic experience from RMC s 3DOF motion platform Furthermore the flight simulator s capabilities and degree of realism when driven by the classical WE algorithm are investigated The classical WF algorithm is the most widely used algorithm for 6DOF motion platforms in the flight simulation industry 4 It is used to filter the accelerations of a simulated aircraft so that the platform is not moved beyond its limits and so that it will washout or re
117. directly into the Simulink environment using a serial connection through the RS 232 connector of the unit Specific details on the implementation of an interface to receive the IMU 320 signal into the Simulink environment are discussed in Annex F Table 11 shows the relevant performance data of the IMU 320 100 Figure 55 The Position of the IMU 320 on RMC s 3DOF Motion Platform Table 11 Crossbow IMU 320 Performance Data Performance Acceleration Angular Rate Range 4g 150 deges Resolution lt 0 5 mg lt 0 02 deges Bandwidth 20 Hz 20 Hz Scale Factor Error lt 1 lt 1 As can be seen from the performance data of the IMU 320 it is well suited for this experiment as it is capable of measuring acceleration and angular rate at ranges and resolutions well beyond the values that will be encountered during the tests 5 2 3 Choice of Factors Factors represent the independent variables that are varied throughout the experiment to affect its outcome Typically these variables are varied over a pre selected range established by 101 the experimenter based on his knowledge of the process and theorical considerations 25 In this study the factors of interest are the parameters that define the behaviour of each motion channel Scaling coefficients and filtering equations for each motion channel represent these parameters In addition the rate limiting coefficient of the tilt coordination channe
118. dure usd in this study The pilots were asked to rate the flight simulator on four topics of interest The pilot validation exercise provided critical information that helped answer some very important questions about RMC s 3DOF flight simulator Although the pilot evaluations remain largely subjective and will vary greatly depending of pilot experience and perspective they remain 141 important when trying to evaluate the fidelity of a flight simulator Overall pilots were satisfied with the flight simulator and there was a conscensus that motion was greatly beneficial to flight simulation Areas of complaints emerged almost exclusively from limitation of the hardware used in this study It is fair to say that almost all pilot complaints could be eliminated or alleviated using better equipment which is available in most cases From these results it is fair to say that RMC s 3DOF motion platform is suitable for use as a flight simulator This represents the most important conclusion emerging from this study With that in mind the next step is to define a more ambitious project where a training application is selected and where the system is designed specifically to suit that application 142 CHAPTER 6 CLOSING REMARKS 6 1 Introduction The work conducted in this study was aimed at designing implementing and analysing a flight simulator that takes full advantage of RMC s 3DOF motion platform More specifically the challenge was t
119. e 90 4 4 1 Special Mounting Parts and Installation eeeseeeeeeeeeeseseeesessessersrsstesesressesressesrresessrsrese 91 4 5 SUMUMALY a eae a ea eia A e A a ere tr dr en rte ere sie 94 CHAPTER 5 EXPERIMENTAL STUDY AND PILOT VALIDATION ss 95 1X 5 1 Introduction release e A Eee etl ees ee 95 5 2 Experimental Study of the System 95 5 2 1 Statement of the Problem and Objectives ss 96 522 2 Measures of Performance and Response Variables cesccessseceeseceeeeeceeeeeenaeceeaeeeenees 97 5 2 3 Choice of Factors vata een eee ee ee eS 100 5 2 4 Exp rimental D signe atelier Rene fn nt en eur nt 101 5 2 4 Experimental Procedure erraminene dirt he einen 106 5 3 Conduct oF the Experiment seen tres ent 108 5 3 1 Reproducibility and Repeatability Considerations 108 5 3 2 Test Rum Fl sis seras sien nt AET 113 5 3 3 Solving the Problem of 2 Order Filter Overshooting ccsscssssssessessssssessesssscseesessseeees 119 5 3 4 Test Rutt 2 norna eine inner ene nant Mn andes ne E met nn tee 123 5 3 5 Test RUN Fr ER aie M MN RE ee a ee ee eS 130 5 3 5 Conelusions ss sn is nt tent ttes nn serie one une tie nt nero ste 132 54 Pilot Validation EXercis eines amine eme ere et ed ne er entente ie ie td o ins sites 133 54 1 Procedure 2er cic divans eal eos E AE es 134 5 4 2 Pilot Ratings and Comments ten nent ein e 136 5 5 SUMMA Vs ent Aee TE E ete tn ati le M tt nn etat M Pr ts tin dits ce Pet
120. e flight simulator exhibits a very weak tilt coordination 0 5 a m s Turn Profile Test Run 1 0 5 10 15 20 25 30 Time s Figure 64 y Axis Accelerations Turn Profile vs Test Run 1 The time interval in which tilt coordination should occur is identified by red markers on the plot of motions The two y axis acceleration peaks on the turn signal are caused by motion generated by the rotational motion channel as the aircraft enters and leaves the turn After further investigation into the tilt coordination channel it was determined that filtering equations exhibit the appropriate behaviour In order to allow for greater tilt coordination there are then two 118 possible options First scaling of the x and y axis acceleration channel can be reduced to allow for a greater amplitude in these signals Second the rate limiting coefficients of the tilt coordination channel currently at 3 deges can be increased Thus the next test run features a combined increase of scaling and tilt limiting coefficients to promote tilt coordination Finally the underpowered rotational motion channel cues are investigated The standard rate turn profile is selected in light of its high angular rate error of 4 4526 deges The plot of motions found in Figure 65 is revealing on the conditions that create motion cueing errors on rotational motions As was the case for the heave channel overshooting can also be found in the rot
121. e motion platform Figure 6 Scaling and Limiting 6 The process used in the classical WF algorithm is illustrated by Figure 6 where L is the limiting value and the slope K is the scaling value 6 The scaling process can also be interpreted as a gain function as described in Equation 2 7 Furthermore the limiting process can be interpreted as a conditional statement as described in Equation 2 8 In the case of scaling for the aircraft accelerations ore the input variable in Equations 2 7 and 2 8 is represented by the specific accelerations the resulting accelerations when the effect of gravity is removed a only This is required because the simulator cab is under the constant effect of the local g thus only 21 the aircraft accelerations that depart from g are simulated 8 Scaling faa g would lead to a false motion cue when the local gravity vector is added after Block 2 of the classical WF algorithm The aircraft angular rates w4 are taken as is and scaled according to Equations 2 7 and 2 8 Input K Scaled Input 2 7 Input If Input lt Li Limited Input Input 2 8 i rer Else 2 3 5 Filtering Filtering plays an essential role in the classical WF algorithm Indeed due to the physical limitations of motion platforms it is necessary to attenuate the low frequency accelerations that tend to drive motion platforms to their limits 12 High frequency accelerations are usually of short duration and do n
122. e ratings for smoothness complained that the initial input to manoeuvers was slightly overpowered Unfavorable ratings for phase lag were attributed by most pilots to an initial phase lag when entering a manoeuvre Unfavorable ratings for the amplitude rating were most of time targeted at decelerations that were sometimes felt to be slightly overpowered especially in turns Overall the motion simulation was felt to be adequate and devoid of major problems After having completed both a motionless and full motion run the pilots were asked to rate the impact of motion simulation on the overall flight simulation In addition pilots were asked to give their opinion on the flight simulator potential as a training aid These ratings provided interesting data on two subjective topics of great importance in defining the future of RMC s 3DOF flight simulator The average ratings collected for both these attributes are shown in Table 30 Table 30 Average Pilot Ratings for Motion Impact and Flight Training Potential Motion Simulation Flight Training Pilot Impact Potential Average 1 52 1 42 Deviation 0 43 0 47 Ratings from 1 to 5 where 1 is excellent and where 5 is very poor The results show that there is a strong conscensus between pilots on the positive impact of motion simulation In addition all pilots feel that there is a potential for training applications using a flight simulator such as RMC s 3DOF f
123. ect to certain constraints that are inherent to the resources available as well as the equipment used for this study Indeed the flight simulator is designed around the existing motion platform RMC s 3DOF and associated hardware This motion platform is not specifically designed to be used as part of a flight simulator and is intented to serve only general purposes As such it comes with certain limitations and deficiencies that must be taken into consideration during the design phase of RMC s 3DOF flight simulator As was done for the design objectives the design constraints are 43 arranged into tree diagrams The top level tree diagram for the design constraints can be seen in Figure 22 Figure 22 Design Constraints The constraints associated with the design of the flight simulator can be divided into hardware and software constraints In the sections to follow the constraints influencing the design of the flight simulator are discussed in details 3 3 1 Hardware Constraints The hardware constraints associated with RMC s 3DOF motion platform come from limitations of each of the different components utilized in its design Additionally the manufacturer SSSI lists constraints for the safe operation of the motion platform Figure 23 shows the sub level tree diagram representing the hardware design constraints Motion Platform Size Center of Gravity Communication Payload Weight Figure 23 Desi
124. ed in the design phase the hardware implementation began This phase required the fabrication of special parts to secure the 91 components to the platform Some of these parts were manufactured locally while others were machined and fabricated off site by a commercial machining shop This section will provide information on the origin and installation of special parts required to install the simulator components in accordance with the design Details of the installation of the flight simulator components will also be discussed 44 1 Special Mounting Parts and Installation Special mounting parts were required to safely mount the key components selected for RMC s flight simulator onto the motion platform Most parts requirements were driven by the concept drawings developed during the design phase as detailed in Chapter 3 Table 10 lists the special mounting parts that had to be fabricated as well as their source Table 10 Special Mounting Parts Mounting Part Source RMC s Electrical Engineering Seat Mount Machining Shop Adjustable Joystick Mount Local Machining Shop Adjustable Throttle Mount Author Adjustable Infra red Sensor Local Machining Shop Mount RMC s Electrical Engineering Seat Harness Installation Machining Shop An aluminum seat mount was designed and fabricated by the local Electrical Engineering machining staff to safely anchor the passenger seat to the motion platform base This
125. ed to produce the final lengths velocities and accelerations of the platform actuators the individual effects of a channel are augmented or dampened by the other channels The flight simulator designer must determine the scaling coefficients so as to create a balanced motion simulation where each channel contributes proportionally and realistically to the final output This is best accomplished when the simulator designer collects the observations of pilots between each run 4 10 The author of this study has more than 100 flight hours in low 97 performance fixed wing aircrafts As such the motion responses can be studied between each run by the author To meet the objective of this experiment the author relied on the self evaluation of the motion stimuli coupled with the use of motion sensors installed on RMC s 3DOF It is felt that the combination of subjective and objective means of measuring motion simulation fidelity will provide the author with ample data for the effective study and optimization of the system Indeed the use of a motion sensor provides an additional method of measuring motion fidelity instead of following the traditional method where WF algorithm tuning relies exclusively on pilot inputs 4 Finally to validate the work conducted during the experiment experienced pilots also flew the simulator during the pilot validation phase and produce additional feedback on the quality of motion simulation This is also neces
126. eedback m gt execute actuator Down Send request to mediator on_mldown_pressed pe Event handler for the Motor 1 Down button When the button is released the actuator stops The request is passed to the mediator in this method eh void MainWindow on_m1down_released Feedback m gt execute stopMovement Send request to mediator on_mldown_released pe Event handler for the Motor 2 Up button As long as this button is pressed actuator 2 will move up The request is passed to the mediator in this method fal void MainWindow on_m2up_pressed Feedback m gt execute actuator2Up Send request to mediator on_m2up_pressed pe Event handler for the Motor 2 Up button When the button is released the actuator stops The request is passed to the mediator in this method gi void MainWindow on_m2up_released Feedback m gt execute stopMovement Send request to mediator on_m2up_released pe Event handler for the Motor 2 Down button As long as this button is pressed actuator 2 will move down The request is passed to the mediator in this method void MainWindow on_m2down_pressed Feedback m gt execute actuator2Down Send request to mediator on_m2down_pressed pe Event handler for the Motor 2 Down button When the button is released the actuator stops 204 The request is passed to
127. eeeeeeeeeeneesnaeenseenees 140 Table 32 Pilot Informati n siie i n eaten eran Lae ele as 173 Table 33 Complete Hardware Rates 2e nette ttes dt and ne 174 Table 34 Complete Motion Simulation Ratings 0 eee eseessecsseceseeeseeceaeceeeeseeeeneecsaeenseesees 174 Table 35 Complete Impact and Training Potential Ratings eee eee eeseeeeeeeseeeeeeenaeeneeees 175 XVI LIST OF SYMBOLS Acronyms 3DOF Three Degree of Freedom 6DOF Six Degree of Freedom C 172 Cessna 172 SP Skyhawk Airplane CF Canadian Forces DLSE Directorate of Land Synthetic Environment DRDC Defence Research and Development Canada FAA Federal Aviation Administration FSX Flight Simulator X GUI Graphical User Interface IDE Integrated Development Environment IFR Instrument Flight Rules IMU Inertial Measurement Unit RMC Royal Military College of Canada SAE Socitety of Automotive Engineers SDK Software Development Kit SGN Sign operation taking only the sign of a variable SSSI Sarnicola Simulation Systems Inc UML Unified Modelling Language USB Universal Serial Bus UTIAS University of Toronto Institute for Aerospace Studies VGA Video graphics array VS2008 Microsoft Visual Studio Professional 2008 WE Washout Filter Winsock Windows Sockets Variables faa xvii Aircraft pitch angle Individual actuator acceleration High pass filtered accelerations Actuator attachments position on the motion platform base Actuator attachments position on the upper motio
128. elevant information such as the rotational velocities and accelerations of the aircraft in real time by writing a plug in 47 Flight Simulator Plug in Compatibility Figure 26 Design Software Constraints Motion Controller Programming Language Compatibility Furthermore it is important to ensure that this information is available from the simulation environment at a sample rate that is sufficiently high to conduct realistic simulation It is also important that a plug in can be written in a language that can be interfaced with all the systems used in motion simulator It is best to select a single programming language in order to write software that can interface all the systems of the flight simulator without unnecessary complications With that in mind the selection of a programming language is restricted to those that are compatible with both the motion controller and the flight simulator software selected to drive the simulation When selecting a programming language issues of portability as described in Section 3 2 3 are also taken into consideration 3 4 Flight Simulator Components By carefully analyzing the design objectives and constraints decisions can be made in order to select the appropriate commercially available components for the flight simulation system This section contains relevant information on the final design components that are selected based upon those criteria More specifically this sect
129. emory In order to simplify the UDP communications it was decided to use a distinct port for each variable in the system The ports used for the transfer of all data over UDP can be seen in Table 8 Table 8 Port Allocation for UDP Traffic Data Data Type Port X axis acceleration Double 10101 Y axis acceleration Double 10102 Z axis acceleration Double 10103 P Angular velocity Double 10104 Q Angular velocity Double 10105 R Angular velocity Double 10106 Actuator lengths Double Array 10107 Actuator velocities Double Array 10108 Actuator accelerations Double Array 10109 a plane pitch angle Double 10110 B plant bank angle Double 10111 In total six different double values are sent from FSXConnect to the Simulink environment Then there are three double arrays of size three the number of actuators for RMC s 3DOF that are sent from Simulink to the Mediator More information about the UDP implementation can be found in Annex K where the C code is available 4 2 7 Documentation The proper documentation of code in software implementations is essential to ensure the portability ease of use and maintainability of any system In addition proper documentation can contribute to cost effectiveness as no time is wasted when a newcomer to the project tries to 87 understand the system As such proper project documentation containing class fields constructors and methods informa
130. emove backward limits on actuators g gt command FL 2147483647 2147483647 2147483647 Remove forward limits on actuators feedback Parked position erased Feedback message 214 Wif else if uInput markPositionParked g gt command DP 6000 6000 6000 Define initial position g gt command FL 6000 6000 6000 Define forward limit for all actuators g gt command BL 5000 5000 5000 Define backward limit for all actuators only 5000 because of platform design feedback Position marked as parked Feedback message W else if else if uInput extend g gt command SP 2000 2000 2000 Set jog speed to 2000 0 066 m s for all actuators g gt command AC 235500 235500 235500 Set acceleration to 235500 7 85 m s s for all actuators g gt command DC 235500 235500 235500 Set deceleration to 235500 7 85 m s s for all actuators g gt command PT 1 1 1 Register all actuator for on the fly absolute position move g gt command PA 6000 6000 6000 Extend the motion platform feedback Extending the platform Feedback message else if else if uInput center g gt command BL 5000 5000 5000 Define backward limit for all actuators only 5000 because of platform design g gt command SP 2000 2000 2000 Set jog speed to 2000 0 066 m s for all actuators g gt command AC 235500 235500 235500 Set acceleration to 235500 7 85 m s s for all a
131. en 1 6 FlightProfile time FlightProfile signals values 1 len 1 6 end function Smooth Y fastsmooth Y smoothwidth fastsmooth Y w smooths vector Y by triangular smooth of width smoothwidth Works well with signals up to 100 000 points in length and smooth widths up to 1000 points Faster than tsmooth for smooth widths above 600 points T C O Haver 2006 w round smoothwidth SumPoints sum Y 1 w s zeros size Y halfw round w 2 for k 1 length Y w s k halfw SumPoints SumPoints SumPoints Y k SumPoints SumPoints Y k w end s s w SumPoints sum s 1 w Smooth Y zeros size s for k 1 length s w Smooth Y k halfw SumPoints SumPoints SumPoints s k SumPoints SumPoints s k w end Smooth Y SmoothY w Annex F Crossbow IMU 320 Simulink Interface 160 161 The Crossbow IMU 320 was interfaced with the Simulink environment in order to collect accelerations and angular rates from RMC s 3DOF flight simulator The IMU comes with a user s manual 26 which contains a programming guide including all the necessary information for interfacing the unit Many data packets can be sent and received from the IMU For this study the packet of interest the S1 packet contains the accelerations angular rates and temperature information As can be seen from Figure 74 all Crossbow IMU 320 packets conform to a specific packet format and are always headed by the bit pattern 0x5555 Figure 7
132. en available can be made available upon request Proceeding with the flight testing and filling the rating questionnaire will be interpreted as consent to participate in this study If you wish to have a copy of this sheet you may simply detach it from the questionnaire and take it away with you Thank you for your time and consideration 187 ETHICS APPROVAL LETTER File number 2010 22 MCS Project title The Design Implementation and Study of a Low Cost 3 Degree of Freedom Flight Simulator using RMC S 3 DOF Motion Platform Principal investigator 2Lt Jean Le Bouthillier Co Investigator Supervisors Dr Yawei Liang Date of submission 15 October 2010 Anticipated commencement date 25 October 2010 Anticipated completion date 5 November 2010 Date of approval 21 October 2010 Period of approval 21 October 2010 to 20 October 2011 This is to inform you that RMC Research Ethics Board REB has granted approval to the above mentioned project and it can now proceed The approval is based only on the documents submitted and only in the language s presented This approval is valid for 12 months If the project goes beyond this date you must inform REB and obtain approval for an extension Any intentional changes to the protocol prior to the start of data collection must be submitted to and approved by the Chair Researchers should not proceed with a project if unforeseen changes to the protocol threaten participants right
133. ent cena eae ee REE EE e 49 3 4 2 Visual Display and Sound sise 51 3 4 3 Pilot S at and COnttols 2 semis es dirt one sn tea Aa image nee 54 3 4 4 Simulator Operation and Safety ss 61 3 4 5 Graphical User Interface han ana ent ht denis 62 3 4 6 Programming Language sine diamant ini a iia i 64 3 5 Summar VE 55528025 ne sun eau ig len E Ni A de ne Seat ne Mr E ANNER 65 CHAPTER 4 IMPLEMENTATION inner T E L TEA E e 66 4 1 Introduction EE E EE tiene EE E O E E E E E net 66 4 2 Software Implementati t sn siie erriei eiia iaaiaee Ee Se Es Eeri r ee R Saia 67 4 2 1 Purpose and Use Cases iti nde Mt een ee aE EAEE a ETE TEA 68 4 2 2 Functional Approach to Implementation ss 70 4 2 3 Mediator Design Pattern 1 sise aep nie dites RAR nt tn dti A sets 71 4 2 4 Flisht Simulator X Pl g ini5 sssse ane Mise de nt tiens listen inerte an lee 73 4 2 5 Classical WF Algorithm Implementation ss 77 2 4 4 Calculating the Actuator Extensions Velocities and Accelerations 81 4 2 6 Networking lens amine a nain eee nr ee tite 83 4 2 7 Doctimentation serora aii ann nn Wine ea art heel he NN eee 86 4 3 Software Validation state te ees ee Rene ee ee ee 88 4 3 1 Technical Ventilation unissent ee Al da nal nines 88 4 3 2 User Validation sarees cst gi te coerce nant en RAR UN ts stat anse 89 4 4 Hardware Implementation oenen e ee occu ea cases E ae ae Earra cove teestseen
134. equal to the maxReturnedItems defined There is an inversion in the axis to convert to the proper axis for the washout algorithm hr SimConnect_AddToDataDefinition hSimConnect DEFINITION_PDR ACCELERATION BODY Z meters per second squared SIMCONNECT_DATATYPE_FLOAT32 0 DATA_X_ACCELERATION hr SimConnect_AddToDataDefinition hSimConnect DEFINITION_PDR ACCELERATION BODY X meters per second squared SIMCONNECT_DATATYPE_FLOAT32 0 DATA_Y_ACCELERATION hr SimConnect_AddToDataDefinition hSimConnect DEFINITION_PDR ACCELERATION BODY Y meters per second squared SIMCONNECT_DATATYPE_FLOAT32 0 DATA_Z_ACCELERATION hr SimConnect_AddToDataDefinition hSimConnect DEFINITION_PDR ROTATION VELOCITY BODY Z degrees per second SIMCONNECT_DATATYPE_FLOAT32 0 DATA_P_ANGULAR_VELOCITY hr SimConnect_AddToDataDefinition hSimConnect DEFINITION_PDR ROTATION VELOCITY BODY X degrees per second SIMCONNECT_DATATYPE_FLOAT32 0 DATA_Q_ ANGULAR_VELOCITY hr SimConnect_AddToDataDefinition hSimConnect DEFINITION_PDR ROTATION VELOCITY BODY Y degrees per second SIMCONNECT_DATATYPE_FLOAT32 0 DATA_R_ANGULAR_VELOCITY hr SimConnect_AddToDataDefinition hSimConnect DEFINITION_PDR PLANE PITCH DEGREES degrees SIMCONNECT_DATATYPE_FLOAT32 0 DATA_P_AIRCRAFT_PITCH hr SimConnect_AddToDataDefinition hSimConnect DEFINITION_PDR PLANE BANK DEGREES degrees SIMCONNECT_DATATYPE_FLOAT32 0 DATA_B_AIRCRAFT_BANK
135. er Private C 150 C 152 C 172 Katana DV 20 11 85 55 Bayete Eia yik C 172 Grob 120A training Air Force 12 260 70 Private Er vary Ehei C 150 C 152 C172 Grob 120A training Air Force Pilot flight time is in most cases approximate as reported by the pilots on the evaluation questionnaire Table 33 Complete Hardware Ratings 174 Pilot Visual eae Controls Audio Cockpit Overall Tracking 1 1 5 1 5 2 1 1 1 25 2 3 4 5 3 5 2 2 3 3 2 3 2 5 2 3 2 4 4 2 1 5 1 1 1 5 5 2 2 2 1 2 2 6 2 1 2 2 2 1 7 2 2 1 5 1 1 1 5 8 3 3 5 2 2 2 9 1 2 1 5 1 5 2 5 1 5 10 2 2 5 1 5 1 1 5 11 1 5 1 2 1 1 5 12 2 2 1 5 1 1 5 Average 2 17 2 21 2 21 1 38 1 63 1 69 Deviation 0 81 0 96 1 05 0 48 0 71 0 51 Ratings from 1 to 5 where 1 is excellent and where 5 is very poor Table 34 Complete Motion Simulation Ratings Pilot Smoothness Sense Amplitude Phase Lag Discomfort Disorientation Overall 1 2 1 5 2 1 75 1 5 1 1 75 2 25 2 2 3 1 1 5 2 3 2 9 2 2 2 5 1 1 2 4 3 5 1 5 229 1 5 1 1 1 5 5 2 5 2 2 2 9 1 1 2 6 2 3 2 2 2 2 2 7 2 5 2 3 1 5 1 1 5 1 8 4 2 4 1 2 3 3 9 2 1 2 2 1 1 1 5 10 2 1 5 1 5 1 5 1 1 1 11 4 3 25 1 5 1 3 2 5 12 a 1 5 1 7 2 1 1 1 5 Average 2 67 1 92 227 1 90 1 21 1 50 1 81 Deviation 0 75 0 60 0 67 0 56 0 40 0 77 0 58 Ratings from 1 to 5 where 1 is excellent and where 5 is very poor
136. er functions of the aircraft cockpit A comparison between three commercially available pilot input systems can be seen in Table 5 These pilot input systems were considered for use in the flight simulator Table 5 Cost Benefit Analysis Between three Commercially Available Pilot Input Systems acy X52 Flight Pro Flight Yoke Flight System System System G940 Manufacturer Saitek Saitek Logitech Price Around 100 Around 200 Around 300 Force Feedback No No Yes Rudder Included Yes No Yes Number of Levers on Throttle 3 i Control Column Joystick Yoke Joystick Number of Buttons 23 14 32 Customizable Buttons Yes Yes Yes Hard points for anchoring None Desk Clamps Mounting Holes Two sets of X52 Flight Systems are Available in the Department The X52 Flight System available in the department was the first option Despite its immediate availability the lack of anchoring points the absence of force feedback and the fragility of components were negative points that warranted the search for a better pilot input system The Pro Flight Yoke System was also evaluated for potential suitability It features a realistic yoke as well as a 3 lever throttle Despite its many interesting features this system comes without rudders and is provided with a clamp only designed to grip the side of a desk 58 Although it does not feature force feedback the main control column of the Pro Flight Yoke System i
137. ergized and ready 2 Enter the controller IP address 3 Press the Connect button to establish a connection to the controller After a succesful connection the platform should automatically center itself At this point the manual controls are enabled 4 Open Microsoft FSX 5 Press the Start FSX Connect button to establish a connection After a successful connection the FSX Connection status changes to Established 6 Run the Classical Washout algorithm Simulink model Time step must be inf infinite 7 The simulink connection should be automatically detected and the Begin Motion Simulation button will become enabled V Feedback ON OFF l Clear Display Start FSX Connection Stop FSX Connection FSX Connection Simulink Connection Begin Motion Simulation Stop Motion Simulation Pitch Forward Backward Surge Forward Backward V RIT DEVOIR VAILLANCE Up Down Position Extend Center Park Motor 6 Up Down Erase Parked Position Mark Position as Parked Figure 37 Final GUI Design A feedback window prominently displays relevant information to the user as he controls the motion hardware Detailed instructions on how to begin a flight simulation are also included in this feedback window The entire flight simulation setup can be controlled by four buttons 64 below the feedback window The Stop Motion Simulation
138. ersonnel or damage equipment The motion platform operator is responsible for the safety of platform passengers and spectators during the operation of the motion platform To ensure safety during platform operation the following rules must be followed at all times 1 Before applying power to the motion platform ensure that all personnel and equipment is at a safe distance of the motion platform 2 When passengers are boarding and un boarding the motion platform it should always be in the parked position and powered off using the safety switch 3 Before the motion platform is energized the motion platform operator must assist the passenger in safely fastening his seat harness The passenger harness must remain fastened until the platform is completely shut down safety switch is off 4 The motion simulation can be halted at any time by the passenger by pressing the pause button This button must be clearly identified to the passenger by the motion platform operator 5 The motion platform operator must provide a short familiarization briefing covering the layout of controls and other pertinent information to new passengers 6 In the event of an emergency the following procedure should be followed a Attempt to stop the motion simulation using the software interface b If the software fails power off the motion platform using the safety switch c When the platform is powered off manage the emergency situa
139. es 0 0608 deges 0 1971 deges Angular g 0 1040 deges 0 0996 deges 0 3031 deges Angularr 0 0451 deges 0 0249 deges 0 0948 deges 113 Smoothed Crossbow Signal Raw Crossbow Signal Margin of Error 26 y a p q r Figure 62 Margin of Errors Smoothed vs Raw Crossbow Signals As such it was decided that all Crossbow IMU channels would be smoothed during the experiment Having studied the precision of the system and the measurement technique used during the experiment it is deemed that sufficient precision is exhibited to proceed with the experiment This precision guarantees a high degree of repeatability between each test run In light of these facts it is decided that a recording for each test run is sufficient to establish a significant measure of performance Performing several recordings and averaging over each test run would increase the time required to conduct the experiment and would not provide any substantial gain in the quality of the data 5 3 2 Test Run 1 The first test run featured the configuration with the most favourable pilot responses from 4 and 12 This configuration can be seen in Table 16 In essence this test run was an attempt to reproduce the conditions that received good pilot feedback for the Classical WF algorithm 4 and to evaluate how well they work with RMC s 3DOF 114 Table 16 Classical WF algorithm Configuration Test Run 1
140. est runs it was discovered that this technique is only partially effective as most overshoot begins in the middle of an acceleration peak Because of this they benefit from a steep slope that lets the overshoot signal permeate through to the filtered signal Another shortfall of the derivative switch technique is that unfiltered signals are not perfect and exhibit numerous areas where the signal s slope is exactly zero This can be seen in Figure 67 where an interval of the signal exhibiting a slope of zero is magnified These areas of flat slope exist when a new data packet is not received within the time step interval It is so because the algorithm is configured to output the last received value when no new value is available a m s a m s 121 Unfiltered Filtered 3 10 15 20 25 30 Time s Figure 67 Overshoot Management Using a Derivative Switch Unfiltered Filtered 5 10 15 20 25 30 Time s Figure 68 Overshoot Management Using a Polarity Switch 122 This phenomenon causes a very noisy and underpowered signal that hops between zero and peak values when using derivative based switching An example of this technique where filtered signals are switched based on the derivative of unfiltered signals is shown in Figure 67 In light of the problems encountered when using the derivative another method was designed and tested to remove overshoots The method in question is based on the parity of the unfil
141. est use of the head Head tracking tracking Some weird visual orientations were reported when pilots reached the limits of head tracking Finally some pilots noted a slight delay between head movement and the image displayed Some pilots were dissatisfied by the non linear gain exhibited by the joystick Some pilots were also dissatisfied by the limited amplitude of the joystick Resistance was deemed to be slightly weak when Controls compared to the control column of an actual aircraft Pilots reported oversensitive for progressive flaps retraction and trim control Most pilots complained about the absence of rudder pedals 138 Pilot complaints for the flight simulator hardware where targeted at limitations exhibited by the visual head tracking and control equipment It was found that the bulk of pilot feedback was aimed at the hardware utilized on the motion platform rather than the sensation of motion As such Table 28 summarizes the complaints collected from pilots during the validation exercise Certain complaints were more significant than others Indeed most pilots found the use of head tracking to be a major improvement over the use of a static display However most pilots were also critical of the limitations of the head tracking system in use but agreed that this technique showed great potential if complaints were addressed Next the absence of rudder pedals was noted by pilots as the most detrimental to the
142. etween motions of heave and pitch Figure 71 shows the y axis accelerations for this test run where it can be seen that the simulated translational accelerations are similar to that of the flight manoeuvre The next test run will 129 feature a slightly higher tilt rate limiting gain to ensure that pitch attitudes are felt by the pilot at a level proportional to those of roll attitudes 6 T T T T T Turn Profile Test Run 2 p deg s 0 5 10 15 20 25 30 Time s Figure 72 p Angular Rates Turn Profile vs Test Run 1 Finally the underpowered feeling experienced on the angular motion channel is studied This problem is closely related to the pitch tilt coordination problem where interactions with the heave motion channel create a dampening effect on pitch motion For the pilot a feeling of underpowered pitch is created by comparing the motion response of the flight simulator to roll manoeuvres with pitch manoeuvres As such it is critical to ensure that both roll and pitch motions create a sensation of the same magnitude for equivalent rotations about the x and y axis In light of these observations the next test run will feature an increased scaling coefficient on the y axis angular motion channel 130 The second run featured many improvements to the motions generated during flight Having identified the remaining causes of pilot complaints the next test run will feature a configuration intended to c
143. execute stopMovement Send request to mediator on_m6up_released pe Event handler for the Motor 6 Down button As long as this button is pressed actuator 6 will move down The request is passed to the mediator in this method af void MainWindow on_m6down_pressed Feedback m gt execute actuator6Down Send request to mediator on_m6down_pressed pe Event handler for the Motor 6 Down button When the button is released the actuator stops The request is passed to the mediator in this method void MainWindow on_m down_released 206 Feedback m gt execute stopMovement Send request to mediator on_m6down_ released Mediator h pee Mediator h IDE Visual C 2008 Express Edition author 2Lt Jean Le Bouthillier Latest Revision 11 May 2010 ek pragma once ifndef MEDIATOR_H define MEDIATOR_H include FSXConnect h include UdpReceive h include Model h include amp lt string amp gt include amp lt math h amp gt include amp lt iostream amp gt include amp lt sstream amp gt using namespace std class Mediator public Mediator string execute string uInput void simulation string connect string ipAddress bool isReceivingUDP bool isFSXConnected void change_speed int s private rere Pointer to a Model object Model m pee Pointer to an FSXConnect object FSXConnect f pee Double d
144. f a test run FlightProfile represents the reference flight profile Author 2Lt Jean Le Bouthillier Date September 5 2010 First the effect of gravity will be added to the Flight profile data as itis seperate from specific accelerations len length FlightProfile signals values FlightProfile signals values 1 len 1 3 FlightProfile signals values 1 len 1 3 FlightProfile signals values 1 len 4 6 FlightProfile signals values 1 len 4 6 FlightProfile signals values 1 len 7 9 Weare saving the first and last 10 values from the z axis force channel This is necessary to prevent the smoothing algorithm from zeroing those values Their average is calculated and added to the smoothed dataset Smooth the Crossbow Acceleration signals Ax head mean Dataset signals values 1 10 1 tail mean Dataset signals values len 10 len 1 Dataset signals values 1 len 1 fastsmooth Dataset signals values 1 len 1 10 Dataset signals values 1 10 1 head Dataset signals values len 10 len 1 tail Ay head mean Dataset signals values 1 10 2 tail mean Dataset signals values len 10 len 2 Dataset signals values 1 len 2 fastsmooth Dataset signals values 1 len 2 10 Dataset signals values 1 10 2 head Dataset signals values len 10 len 2 tail Az head mean Dataset signals values 1 10 3 tail mean Dataset signals values len 10 len 3 Dataset signals values 1 len 3
145. fects associated with this frame are less than that of filtering in Fs Bode Diagram a r n Magnitude dB D D Oo Oo oO 1 D oO N 1 80 amp r ror rrrrrE r fe cp papopupore r p or rerprE r eoe OY Co wf 0 pa ae ne Re 1354 LL a 90 N Phase deg fl NC 4 0E ror pe eee ror rrrrrrF SN E naea aua E 1 0 1 2 10 10 10 10 10 Frequency rad sec Figure 8 Translational z Component High Pass Filter w 4 rad sec Acceleration m s s 3H Input Filtered T 44 r r r r r r r r 0 2 4 6 8 10 12 14 16 18 20 Time sec Figure 9 Translational Channel High Pass Filter Sample Run w 4 rad sec 26 Figure 8 shows the Bode plot of the high pass filter used to filter the z translational acceleration component where the frequency is in rades the standard unit for bode plots As can be seen from the gain magnitude plot the magnitude of a signal is progressively reduced by a negative gain as its frequency lowers This has a scaling effect on the signal For this reason fine tuning in conjunction with the scaling values of the translational motion channel is indicated to obtain the overall desired scaling effect The Phase plot is commonly used to illustrate the frequency response phase shift The general behaviour of the z component high pass filter can be seen in Figure 9 where the results of a sample run are shown As can be seen fro
146. flying experience Pilots were dissatisfied by the inability to center the aircraft properly during landing and takeoff manoeuvres Finally pilots with prescription glasses had a hard time with the head mounted display This aspect was not discussed during the design phase and highlights the fact that such a system is not suitable for all Despite the seemingly long list of complaints and comments it is felt that improvements can be easily achieved by purchasing new equipment or changing some simple configuration settings The average motion simulation pilot ratings collected during the validation exercise can be found in Table 29 As expected the areas of least performance are found to be the smoothness and amplitude attributes of motion All pilots experienced little discomfort or disorientation during the flight simulation Table 29 Average Pilot Ratings for Motion Simulation Pilot Smoothness Sense Amplitude 2 Discomfort Disorientation Overall Average 2 67 1 92 2 27 1 90 1 21 1 50 1 81 Deviation 0 75 0 60 0 67 0 56 0 40 0 77 0 58 Ratings from 1 to 5 where 1 is excellent and where 5 is very poor 139 Overall it was felt by pilots that the sense of motions was appropriate and without any noticeable phase lags By comparison to the hardware evaluation there were very little complaints with regards to the motion simulation aspect of flight simulation Most pilots who reported unfavourabl
147. for motion cueing calculations Fortunately FSX comes with a powerful client server communications protocol SimConnect that allows accessing the internal data of FSX in an asynchronous manner The protocol is considered asynchronous because a publish subscribe method is used to establish the communication between the client and server 20 As a result once a client subscribes to a data or event the server internally processes and returns the information at a frequency that can vary and is specified in the initial client request Since the requests are internally processed by the server and never handed for processing to the client this communication is considered to be asynchronous 20 Request Data gt SimConnect SimConnect FSXConnect CPP Client Server FSX Receive Data Figure 42 SimConnect TCP IP Communications Model As can be seen in Figure 42 the TCP IP Winsock networking protocol is used to establish network communication between a SimConnect client on top of an external out of process executable and the SimConnect server in process component of FSX In this study a C object FSXConnect cpp was implemented to establish a SimConnect communication with FSX and to obtain the relevant data and events It is possible to retrieve data from FSX in three 75 different frequency modes every visual frame every simulation step or every second 1 Hz
148. ft s ailerons and elevator Figure 32 Corbeau Four Points Harness Belt Typically two types of control column are used to control an aircraft s pitch and roll a yoke or a joystick The yoke is traditionally used in commercial aircraft whereas the joystick is mostly used in military or aerobatic airplanes The most readily available type of input device to control computer based consumer flight simulator is the joystick High end commercially available joysticks offer force feedback to simulate the response of the aircraft s control surfaces to the aerodynamic effects on the airplane Next the throttle is also an important control for the aircraft pilot It allows modifying the engine s rpm and consequently impacts the aircraft s velocity Commercially available consumer models often feature as many as three levers that allow for the additional control of the airplane propeller pitch and engine gas mixture With both the throttle and joystick the pilot is using both his hands to control the aircraft In order to control the yaw of the aircraft the pilot must use his feet to action rudder pedals In addition to controlling the aircraft s yaw rudder pedals are used for ground steering as well as braking A review of the commercially available input devices shows that most high end consumer systems feature a combination of joystick throttle and rudder pedals assorted with numerous buttons that 57 can be configured to implement oth
149. ft Vol 25 No 7 July 1988 In Motion Simulation Website 3DOF Accessed on 23 October 2009 http www inmotionsimulation com 3DOF html Sarnicola Simulation Systems Inc Triad 1500 E Motion Platform Owner s Manual Land Force Command Order 28 01 Land Force Simulation Policy in Support of Land Force Training Reid L Nahon M Flight Simulator Motion Base Drive Algorithms Part 1 Developing and Testing the Equations University of Toronto Ontario Canada UTIAS Rept 296 Dec 1985 Reid L Nahon M Flight Simulator Motion Base Drive Algorithms Part 2 Selecting the System Parameters University of Toronto Ontario Canada UTIAS Rept 307 May 1986 Reid L Nahon M Flight Simulator Motion Base Drive Algorithms Part 3 Pilot Evaluations University of Toronto Ontario Canada UTIAS Rept 319 Dec 1986 Reid L Nahon M Simulator Motion Drive Algorithms A Designer s 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 226 Perspective Journal of Guidance Control and Dynamics Vol 13 No 2 1990 pp 356 362 Grant P Reid L Motion Washout Filter Tuning Rules and Requirements Journal of Aircraft Vol 34 No 2 March April 1997 pp 145 151 Haik Y Engineering Design Process Brooks Cole Thomson Learning 2003 270 p Burki Cohen J Soja N Longridge T Simulator Fidel
150. g a software limiter in the motion platform control software Thus the rear motion platform actuators were limited at a height of three centimetres above the drive motors After thoroughly testing the modified software limiting it was established that such an incident would not re occur This incident served as a reminder that hardware constraints must be fully understood and integrated in the design of a flight simulator A thorough evaluation of the manufacturer s claims with regards to performance and limitations is always a good idea 3 3 2 Software Constraints Some of the software constraints associated with the design of RMC s 3DOF flight simulator come from the Galil motion controller provided with the motion platform Software constraints also emerge from the requirement to select flight simulation software for the virtual environment Additionally software constraints emerge when a single programming language is selected in order to interface all the components of the system Figure 26 shows the sub level tree diagram representing the software design constraints The Galil DMC 40x30 motion controller is built upon native C programming language libraries and also extends compatibility to Windows programming languages such as VB and C As such either of these languages it supports could be utilized to communicate with it Next when selecting a flight simulation software it is important to ensure that there is a possibility to extract r
151. geBackward gt setDisabled true ui gt rollLeft gt setDisabled true ui gt rollRight gt setDisabled true ui gt extend gt setDisabled true ui gt center gt setDisabled true ui gt park gt setDisabled true ui gt erasePosition gt setDisabled true ui gt center gt setDisabled true ui gt markPositionParked gt setDisabled true else if enableManualControls Event handler for the Pitch Forward button As long as this button is pressed the platform will pitch forward The request is passed to the mediator in this method ck void MainWindow on_pitchForward_pressed Feedback m gt execute pitchForward Send request to mediator on_pitchForward_pressed pe Event handler for the Pitch Forward button When the button is released the motion stops The request is passed to the mediator in this method void MainWindow on_pitchForward_released Feedback m gt execute stopMovement Send request to mediator on_pitchForward_released Event handler for the Pitch Backward button As long as this button is pressed the platform will pitch backward The request is passed to the mediator in this method iwi void MainWindow on_pitchBackward_pressed Feedback m gt execute pitchBackward Send request to mediator on_pitchBackward_pressed Jx Event handler for the Pitch Backward button When the button is released the motion stops
152. ght Recording Block 166 The playback functionality was implemented by replacing the UDP Server block by the Flight Playback block which is configured to read Structure With Time from the MATLAB environment This block feeds into the unchanged components of the classical WF algorithm implementation to drive the motion platform Finally a Crossbow IMU recording block was added to the model in order to record the platform motions into the MATLAB environment as a Structure With Time Inside the Crossbow Recorder block information is recorded in a similar fashion than for the Flight Recorder block of Figure 78 Flight Playback Crossbow Crossbow Recorder wAA Scaling and Limiting Figure 79 Top Level View of Playback and Crossbow Blocks An inside view of the Flight Playback block where a From Workspace block is used to fetch data at the frequency of the system can be seen in Figure 80 The data interpolation option is selected to ensure that a value is always outputted from the Flight Playback block The block is also configured to output zero after the final data value is read from Structure With Time M Source Block Parameters From Workspace From Workspace Read data values specified in array or structure format from MATLAB s workspace Array or matrix format 1 D signal var TimeValues DataValues For multidimensional signal use structure format Structure format var time TimeValues v
153. ght simulation industry has been reported to exhibit excellent motion cueing capabilities 11 However the filters used in the algorithm must be properly fine tuned in accordance with the motion hardware in order to provide the best possible motion cues 12 Figure 16 shows the specific design sub objectives relating to the design objective of achieving a realistic simulation Realistic Simulation Audio Visual Fidelity Cabin Controls Pilot Seat Basic Airplane Controls Graphics Sounds WF Algorithm Sensory Isolation Figure 16 Realistic Simulation Design Sub Objectives Various systems such as screens projectors and head mounted displays are commercially available to generate visual cues Factors such as image quality screen size and resolution have an impact on the cost of such devices It is also important to eliminate any fixed reference points when trying to immerse a passenger in a virtual environment Without proper visual isolation isolating the passenger s vision from the motionless surroundings the pilot could receive misleading cues from his her ambient peripheral vision from the surroundings of the simulator It is possible to deliver audio stimuli to the passenger by either an ambient or headset system To provide a realistic simulation the occupant of the flight simulator should be isolated from outside noises Both ambient of headset systems are suitable depending on the presen
154. gma once ifndef FSX_CONNECT_H define FSX_CONNECT_H include amp lt winsock2 h amp gt include amp lt string amp gt include SimConnect h Microsoft SimConnect C Library include UdpSend h using namespace std class FSXConnect private pee Flag used to notify the end of a connection bool quit ex Flag used to hold the connection status bool connected per SimConnect handle to FSX HANDLE hSimConnect pee UDP Client to send data to Matlab UdpSend udpSend pee Number of UDP ports required PortNumber udpDataPorts 8 pee Buffer for Aircraft Pitch and Bank values double pitchBank 2 ex A structure for the inputs of FSX struct Inputs int id An integer for the id of the simulation variable float value A float variable to hold the simulation variable 190 pee 6 maxReturnedltems as we request 3 accelerations and 3 angular velocities define maxReturnedltems 8 ex A structure that is used to receive Tagged data struct StructDatum Inputs datum maxReturnedItems StructDatum ex An enumeration of the events we are interested in static enum EVENT_PDR EVENT_SIM_PAUSE EVENT_PDR ex An enumeration of the definitions we are interested in static enum DATA_DEFINE_ID DEFINITION_PDR DATA_DEFINE_ID ex An enumeration of the requests we are interested in static enum DATA_REQUEST_ID REQUEST_PDR DATA_REQ
155. gn Hardware Constraints RMC s 3DOF motion platform is controlled by a Galil DMC 40x30 motion controller This motion controller is widely used in various control applications and features two modes of 44 communication Indeed the motion controller can be interfaced through Ethernet or RS232 communication protocols 6 As such either of those two standards could be utilized to communicate with the platform Each of these communication protocols features different speeds which will create different latencies in the system In this case latency should be kept at a minimum to ensure proper synchronization and realtime operation of the motion platform with the simulation visual environment In general the Ethernet protocol offers much more speed and performance than the RS232 protocol Therefore it is better suited for the design of a flight simulator Another hardware constraint of importance is the motion platform size The upper platform frame measures 121 9 by 121 9 cm Any equipment used for the flight simulator cabin and controls must remain within this area Additionally the center of gravity of any payload positioned on the platform must remain geometrically centered as per the manufacturer s specification According to those specifications the platform is designed to hold a maximum payload of 1500 lbs with its center of gravity situated 45 7 cm above the geometrical center of the platform 6 Therefore the passenger and the bulk of
156. gt success true If there is an initialized Winsock WSACleanup Stop networking functions this gt success false Set the flag to false HI if WsaHelper pe This method returns the state of the winsock initialization ex return amp lt true amp gt if there is an initialized Winsock amp lt false amp gt if there is no initialized Winsock bool Udp IsOKk return this gt success IsOk pe Convenience Udp instance to check for Winsock status amp Udp udp pe This method is used globally to check the status of the Winsock Pa return amp lt true amp gt if the Winsock is ready amp lt false amp gt otherwise fe bool IsWsaReady return udp IsOkQ IsWsaReady UdpReceive h pe UdpReceive h IDE Visual C 2008 Express Edition author Martin Vidovic 218 Modified by 2Lt Jean Le Bouthillier Latest Revision 13 November 2009 ek pragma once ifndef _SP_UDP_RECEIVE_H define _SP_UDP_RECEIVE_H include Udp h class UdpReceive private pee Socket used for datagram reception SOCKET mSocket ex Address IP and port of the datagram origin E sockaddr_in mSourceAddr void SocketCleanup bool ReceiveDatagram void nData size_t nDataSize long nTimeoutMs public Class constructor UdpReceive Class destructor UdpReceive bool Initialize const char nSourcelp PortNumber nSourcePort bool Receive
157. gure 58 Plot of Standard Rate Turn Profile Motions against Time 105 Figure 59 Plot of Descent Profile Motions against Time uu cee eeeeeseceseceseeeseeeeneeesaeenseenees 106 Figure 60 Noisy Crossbow Crossbow Angular Rate Signal 110 Figure 61 Smoothing of a Typical Crossbow Angular Rate Signal 112 Figure 62 Margin of Errors Smoothed vs Raw Crossbow Signals cesceseceeseesseeeeeeeees 113 Figure 63 z Axis Accelerations Climb Profile vs Test Run 1 116 Figure 64 y Axis Accelerations Turn Profile vs Test Run 1 117 Figure 65 p Angular Rates Turn Profile vs Test Run 1 119 Figure 66 2 Order High Pass Filter Overshooting s 120 Figure 67 Overshoot Management Using a Derivative Switch 121 Figure 68 Overshoot Management Using a Polarity Switch 2 0 0 ceeeeeeneceeeeeceeeeeeeteeeeeseees 121 Figure 69 The effect of taking the Derivative in the Tilt Limiting Block 124 Figure 70 z Axis Accelerations Climb Profile vs Test Run 2 127 Figure 71 y Axis Accelerations Turn Profile vs Test Run 2 128 Figure 72 p Angular Rates Turn Profile vs Test Run 1 129 Figure 73 Flight Simulator Wiring 226 lettre terne 154 Figure 74 Crossbow IMU 320 Packet Format cedsesidecdis cna neds aeetiiondned eave eis este 161 Figure 75 Crossbow IMU 320 Simulink Interface 162 Figure 76 Simulink Serial Configurations IMU 320 163 Figure 77 Top
158. h the mediator is ask the begin the motion simulation ai void MainWindow on_begin_button_clicked motionSimulation true Set the motion simulation flag enableManualControls false Disable manual controls of the platform Feedback Motion simulation in progress Feedback message checkConnections Check connections while motionSimulation true m gt simulation QApplication processEvents Allow the GUI to process events while on_begin_button_clicked px Event handler for the Stop Motion Simulation button When this button is pressed by the user the platform is parked and the program closes void MainWindow on_stop_button_clicked motionSimulation false Stop the motion simulation Feedback Motion simulation terminating Feedback message Feedback Please wait a moment while the last motion simulation commands are executed Feedback message while m gt execute stopMovement Cannot execute command Center the motion platform QApplication processEvents while Feedback m gt execute center ui gt stop_button gt setDisabled true Disable the Stop Motion Simulation button enableManualControls true Enable the manual controls of the platform checkConnections Check connections MainWindow MainWindow Close the GUI terminate the execution exit 1 Exit the program on_stop_button_clicked
159. hat would displace the actuators to their physical limits The intermediate variable as represents the high pass filtered accelerations 2 3 2 Rotational Motion Channel The rotational motion channel consisting of Boxes eight to eleven transforms the aircraft angular rates amp 44 into simulator high frequency Euler angles Bax fy Oby WBy In Box eight the aircraft angular rates are scaled down and limited in function of the motion platform physical limits additional information can be found in Section 2 3 4 The scaled down and limited angular rates are represented by the intermediate variable w1 In Box nine the angular rates are transformed into Euler angle rates fy using the transformation matrix Ts given by 18 Equation 2 2 8 As can be seen in Box thirteen the transformation matrix Ts is formed using the current simulator orientation given by Euler angles Bs 1 sinptan0 cos dtan Ts 0 cosh sinp 2 2 0 sin secO cos sec In Boxes ten and eleven the Euler angle rates Bix are high pass filtered and integrated once to obtain the simulator high frequency Euler angles Bor Filtering out low frequency rates is r required as they would drive the motion platform to its limit The intermediate variable Psy represents the high pass filtered Euler angle rates before they are integrated in Block 11 2 3 3 Tilt Coordination Channel The tilt coordination channel transforms the scaled aircraft
160. he C implementation to the Galil motion controller In Figure 48 we can see that both UDP and TCP IP Networking protocols are present in the flight simulation system Flight Simulator X TCP IP C Implementation K TCP IP Motion Platform UDP Simulink Figure 48 Networking in the Flight Simulation System The TCP IP networking segment between the C implementation FSX and the motion platform controller are easily handled as they rely on third party libraries that take care of the networking communications Indeed when using SimConnect to connect to FSX the TCP IP networking details are abstracted and operate without the knowledge of the user Similarly when 84 communicating With the motion platform the Galil C library that is provided with the motion controller handles the networking communications On the other hand the UDP communication protocol used to communicate to and from Simulink had to be implemented in C using the Windows Sockets API Winsock The C UDP implementation allows sending and receiving double values over UDP As can be seen in Figure 49 there are three C classes that had to be used to allow for the transfer of those double values Udp SUCCESS ISOk bool Udp Udbp LZ UdpSend UdpReceive mDestAddr mSocket mSocket mSourceAddr Initialize const char bool Initialize const ch
161. here the pilot s head would move more in the virtual world than it did in reality According to most pilots who tested the flight simulator improvements on head tracking would significantly increase their feeling of immersion in a real aicraft As such it would be fitting to look into products which would allow more freedom in head movement Perhaps a combination of several TrackIRs with customized software would allow for more freedom Another option would be to develop an experimental tracking system capable of all around motions at RMC Such a system would certainly find wide usage in various virtual reality applications Future work in this area could also include a study of the techniques used in the cinematographic industy for the tracking of actors in movies where computer generated imagery is used The second area of unanimous complaint from pilots arose from the lack of rudder pedals on the flight simulator Indeed without rudder input it became hard for pilots to achieve centerline on takeoff and landing manoeuvres This became evident during the pilot validation phase where pilots were asked to perform circuits around the airport Although there were no rudders on RMC s 3DOF rudders could easily be implemented on RMC s 6DOF features a large platform area and yaw motion This would allow for the use of the flight simulator as a training aid for feet inputs which are critical during circuit manoeuvres A rudder system is already av
162. hich are very thorough The Simulink environment allows developers to run several iterations of a model as fast as the host computer will allow it However in the case of a real time system such as a motion platform it is necessary to slow down the simulation in order to achieve real time speed Although this is not a feature that is available using the standard version of Simulink it is possible to slow down the simulation by comparing the host computer s clock to the simulation s clock and by putting the simulation to sleep as required in order to achieve real time speed Such tweaking of the Simulink environment is more advanced but is considered a mainstream practice amongst the MATLAB community As such several blocks and code snippets are available on Mathwork s website to achieve real time speed One such block called RTsync was downloaded and is used to control the simulation s speed 81 2 4 4 Calculating the Actuator Extensions Velocities and Accelerations A technique for calculating the actuator extensions is discussed in 8 However this technique was found to be inadequate for this study and for use with RMC s 3DOF motion platform Actuator extension calculation techniques depend on the type of hardware used and vary depending on the type of commands required by the motion platform The work conducted by 8 to solve this problem does not apply well to RMC s 3DOF motion platform Hence a different te
163. ical limits of a motion platform are referred to as the motion envelope Thus a motion cueing algorithm commonly referred to as the Washout Filter WF algorithm is necessary to transform the aircraft s angular rates and accelerations so that the motion platform remains within its motion envelope It has been substantiated through pilot validation testing that a properly driven motion platform for a flight simulator can greatly enhance the simulation while an improperly cued motion platform can adversely affect the simulation 4 13 RMC 3DOF Motion Platform The Department of Mathematics and Computer Science of the Royal Military College of Canada RMC purchased a Triad 1500 E 3DOF motion platform from Sarnicola Simulation Systems Inc SSSI This motion platform was acquired to provide students and researchers with a versatile tool capable of motion simulation in three degree of freedom RMC s 3DOF motion platform shown in Figure 1 5 is capable of executing movements of roll pitch and heave with a payload of up to 1500 lbs Figure 1 RMC s Triad 1500 E 3DOF Motion Platform The motion is generated by three electric motors in a triangular arrangement connected to the upper platform frame by a double shafted gear head actuator Figure 2 Motion Platform Actuator Assembly The actuator assembly shown in Figure 2 and adapted from CAD drawings 6 transforms rotational motion from each of the motors into linear moti
164. ilt Coordination Rate Limit Form L_IS wi wt B_SH Euler Angle Form T_S WAA Scaling and Limiting UDP Client WAA high pass filter Figure 44 Simulink Implementation of the Classical WF Algorithm Figure 45 shows the internal components of a low pass filter block which would be revealed by a double click of the block in Simulink It is important to specify that the implementation of the classical WF algorithm in Simulink required special MATLAB and Simulink add in packages that are called toolboxes and blocksets It was necessary to acquire the Instrument Control Toolbox Control System Toolbox and the Signal Processing Toolbox and Blockset in order to implement the algorithm These additional packages were required to implement low and high pass filters and to establish network connectivity to and from the Simulink environment 80 Second order translational low pass filters for fAA components ax sy sz The constant passed in each transfer function is an acceleration in meter per second squared y Translational LPF az is left untouched az is left untouched to allow for a greater rate limiting in the next phase Figure 45 Internal Simulink Block View Sample Some minor configuration problems relating mostly to the complexity of the environment were encountered during the implementation but these problems were eventually resolved after reading the Simulink help files w
165. implemented on the GUI of the control software to allow instant termination by the operator 3 4 5 Graphical User Interface The GUI is one of the most important software components in the design of RMC s 3DOF flight simulator Using the GUI the flight simulator operator is able to connect to the motion controller manually control the motion platform and begin or stop the flight simulation In addition the operator is provided with relevant information such as the position of the motion platform actuators To allow for the creation of graphical user interface a GUI builder is used Nokia s QT is a cross platform application and user interface framework By using this C compatible tool it is possible to quickly prototype and implement graphical user interfaces The use of a GUI builder rather than programming allows the designer to focus on the layout and functionality of the GUI rather than the details of its implementation QT also features a VS2008 add in that allows for the seamless integration of both development tools The final design of the GUI can be seen in Figure 37 As can be seen from Figure 37 the GUI design allows for extensive manual control of both the 3DOF and 6DOF motion platforms Z Motion Simulation 2Lt Jean Le Bouthillier v10 me Controller IP Address Manual Controls Connected to Speed 50 Feedback Flight Simulation Checklist 1 Ensure that the controller and motion platform hardware are en
166. ing a transition from one value to another Table 2 shows the filter parameters used with the classical WF algorithm that received the most favourable response from pilots during a study at the UTIAS 11 Table 2 Filter Parameters with Most Favourable Pilot Response UTIAS 11 Filter Order On x translational high pass 2 2 5 1 0 y and z translational high pass 2 4 0 1 0 All rotational high pass 1 1 0 x translational low pass 2 5 0 1 0 y translational low pass 2 8 0 1 0 These values are a good starting point for a study but would need to be fine tuned after evaluating their behaviour and their effects on each channel 23 2 4 Three Degree of Freedom Adaptation In order to use the classical WF algorithm with RMC s 3DOF motion platform some modifications are required First RMC s 3DOF motion platform is not capable of generating surging and swaying motions characterized by translational displacement in the x and y axis respectively The platform can only generate heave motions characterized by translational displacements in the z axis Furthermore the platform is limited in its rotational displacement capabilities as it is not capable of generating yaw motions characterized by a rotation about its z axis The platform is only capable of generating roll and pitch motions characterized by rotations about its x and y axis respectively Despite those limitations the tilt coordination
167. ing pilot evaluation can also be found in the plot of motions An example of such an occurrence is identified on the plot of motions with a blue arrow In the example selected it can be seen that the overshoot also has a cumulative effect as it lowers the starting point of the next acceleration peak Overshoot Descent Profile Test Run 1 0 3 10 15 20 25 30 Time s Figure 63 z Axis Accelerations Climb Profile vs Test Run 1 According to a study conducted by 12 this problem is inherent to the use of second order high pass filters It is possible to slightly reduce overshoot effects by increasing the filter damping ratio of the Equation In addition the effects of overshoots can be dampened by increasing scaling on the affected channel However it was found during testing that these solutions reported by 12 only provide a minimal improvement of overshooting As such the 117 problem of filter overshoot was mitigated through a more radical method As the process used in this study for neutralizing overshooting is very specific Section 5 3 3 is dedicated to this subject Next the absence of tilt coordination designed as a cue substitution for translational accelerations on x and y axis translational motion channel is investigated The profile of choice to investigate this error is the standard rate turn profile as it features the highest tilt coordination error of 4 045 mes As can be seen from Figure 64 th
168. ing system features a highly customizable motion tracking software that is compatible with FSX and other simulation software The motion tracking software is installed on the control station where the TrackIR 5 system is connected Figure 29 TrackIR 5 with TrackIR PRO Clip To complete the immersion inside the virtual flight environment a sound system was incorporated Sounds can be generated by a headset or by an ambient speaker system The generation of sounds through a headset offers many advantages for the design of the flight simulator Indeed the headset provides an immersive experience and most importantly isolation from unrelated environmental noises Isolation from unrelated noises is a very useful feature as the motion platform produces a distracting hissing noise when in operation The Creative Fatallty headset as can be seen in Figure 30 a high end commercially available headset with a microphone was selected to generate sounds for the flight simulation The microphone feature 54 can be used to allow two way communications between the simulator s operator and the passenger Figure 30 Fatallty Sound Headset 3 4 3 Pilot Seat and Controls The installation of a pilot seat as well as flight controls is an essential requirement for the flight simulator Without providing active control of the simulated aircraft to the passenger the system remains a flight simulator but it does not simulate the operation of an ai
169. ion explains how design objectives and constraints materialized into the selection of flight simulator components With clearly defined design objectives and knowledge of the relevant constraints potential equipment was 48 evaluated compared and selected The design phase facilitated the selection of the appropriate commercially available components that interact together to meet the design objectives This phase of the design is important as an understanding of the components and their interactions is the key to a successful implementation discussed in Chapter 4 Safety Switch Flight Simulator X Motion Tracking Software A Infrared Tracker All electronic equipment Identified with a is connected p a to the control station Plug In Within Reach 4 a a f T Energi Gay nergize s oat gt v J ontrol isual Display Software Le f Graphical User Motion Cueing Interface Control Station Ethernet Connection 3 ja Headset Classical WF ee Motion Platform Algorithm r Software Components Throttle j _ Joystick Safety Hamess Passenger Seat Hardware Components Figure 27 Interaction of Components in the Final Flight Simulator Design A complete and detailed listing of all the components used in the Design of RMC s 3DOF flight simulator is available in Annex B Various hardware and software components are added to the motion platform in order to create the flight simulation
170. is chapter the objectives associated with the design of RMC s 3DOF flight simulator are discussed Similarly the constraints which must be considered during the design are discussed In light of the design objectives and constraints the design choices and components selected for the flight simulator are discussed in detail 32 3 2 Design Objectives An important step in design is to clarify the design objectives When creating a flight simulator one of the most basic design objectives is to create conditions where a flight environment is approximated to a degree that is deemed satisfactory for the conduct of specific activities 8 In this case the activities that are of interest are the regular flight operations of aircraft that exhibit low translational accelerations and angular rates Indeed due to the platform motion envelope and limitations it would be unreasonable to attempt the simulation of very agile and aggressive aircraft such as fighter jets Other design factors that may pertain to the available equipment for simulator subsystems such as safety maintenance and user interactions must also be taken into consideration when defining design objectives By defining design objectives it is necessary to begin a research phase in which the literature and the industry are investigated in order to meet the design objectives As dictated by mainstream design processes 13 a thorough investigation of proposed design objectives is necess
171. is is just the size of destAddr structure it s needed by Windows return sendToResult nDataSize Datagram sent successfully if this condition is met SendDatagram px This method is used to send a double value over UDP datagram to a specific port number ek param nPort the port number to which the datagram will be sent param nValue the double value to be sent over UDP return amp lt true amp gt if the datagram was succesfully sent amp lt false amp gt otherwise 7 bool UdpSend SendDouble PortNumber nPort double nValue ex Check if Winsock is ready and socket is created if IsWsaReady false this gt mSocket INVALID_SOCKET Check if the object state is valid return false If something went wrong Mif return this gt SendDatagram nPort amp nValue sizeof nValue Send the datagram SendDouble 5 6 7 10 11 225 REFERENCES Page R L Brief History of Flight Simulation Proceedings of SimTecT 2000 Conference Sydney 2000 CAE Website Motion Simulators Accessed on 23 October 2009 http www cae com en sim products sim products asp Pouliot N A Gosselin C M Nahon M A Motion Simulation Capabilities of Three Degree of Freedom Flight Simulators Journal of Aircraft Vol 35 No 1 January 1998 pp 9 17 Reid L D Nahon M A Response of Airline Pilots to Variations in Flight Simulator Motion Algorithms Journal of Aircra
172. it is important to be able to receive double arrays that contain these values The ReceiveDouble method in UdpReceive is used to receive those double arrays In Figure 50 we can see the position of the UdpSend and UdpReceive classes in the C implementation The sequence of a typical UDP communication between the C implementation and Simulink can also be seen FSXConnect Mediator UdpSend UdpReceive 1 4 C Simulink Receive UDP Classical WF Send UDP double double 2 Algorithm 3 array Figure 50 Typical UDP Communication The UdpSend class is used by FSXConnect to send the double values containing the translation acceleration and angular rates as they are received from FSX The UdpReceive class is used by the Mediator to receive the double arrays containing the motion platform actuator lengths velocities and accelerations from the Simulink environment The UDP communications inside Simulink are easily implemented using the UDP receive and send blocks from the Instrument Control Toolbox All UDP communications are conducted on the local host IP 127 0 0 1 which means that they are conducted on the same machine loopback network connection The order of data used for all UDP communications is the Little endian standard This simply means that the least significant byte of the data is always stored at the lowest address 86 in m
173. ition John Wiley amp Sons Arizona State University 2005 Crossbow Technology Inc 320 Series User s Manual Document 7430 0292 01 Revision B September 2009 27 28 29 30 31 32 33 227 Cessna Website C 172 Skyhawk Accessed on 30 July 2010 http www cessna com single engine skyhawk html GlobalSecurity org T 41 A C Mescalero Article Accessed on 30 July 2010 http www globalsecurity org military systems aircraft t 41 htm Lamoureux L Cin matique et dynamique des solides Trait des Nouvelles Technologies s rie M canique Editions Herm s Paris 1992 Dr T O Haver Smoothing Algorithms Website Accessed on September 21 2010 http terpconnect umd edu toh spectrum Smoothing html algorithms Thong Shing Hwang Ming Sung Kuo Optimal Genetic and Adaptive Fuzzy Washout Filter Design in the Motion Cueing Simulator Asian Journal of Control Vol 10 January 2008 pp 88 95 Gutridge C J Three Degree of Freedom Simulator Motion Cueing Using Classical Washout Filters and Acceleration Feedback Master of Science Thesis Virginia Polytechnic Institute amp State University 2004 Bohemia Interactive Website Virtual Battlespace 2 Webpage Accessed on October 01 2010 http www bisimulations com
174. ity Requirements The Case of Platform Motion 9 ITEC International Training amp Education Conference Lausanne Switzerland pp 216 231 Mooney J D Bringing Portability to the Software Process Dept of Statistics and Computer Science West Virginia University Morgantown WV IEEE Std 610 12 1990 Glossary of Software Engineering Terminology in Software Engineering Standards Collection IEEE CS Press Los Alamitos California 1993 Fowler M UML Distilled Third Edition A Brief Guide to the Standard Object Modelling Language Addison Wesley Aug 2009 Weisfeld M The Object Oriented Thought Process Third Edition Addison Wesley Oct 2009 Freeman E Freeman E Head First Design Patterns O Reilly Oct 2004 About SimConnect Microsoft ESP Developer Center Accessed on 10 May 2010 http msdn microsoft com en us esp cc835021 aspx SimConnect SDK Reference MSDN Library Accessed on 10 May 2010 http msdn microsoft com en us library cc526983 aspx Simulink Simulation and Model Based Design Accessed on 11 May 2010 http www mathworks com products simulink MATLAB The Language of Technical Computing Accessed on 11 May 2010 http www mathworks com products matlab Meyers G Software Reliability Principles and Practice Chapter 10 Testing Principles A Wiley Interscience Publication 1976 pp 169 195 Montgomery D C Design and Analysis of Experiments 6 Ed
175. k m gt execute surgeForward Send request to mediator on_surgeForward_pressed Event handler for the Surge forward button When the button is released the motion stops The request is passed to the mediator in this method void MainWindow on_surgeForward_released Feedback m gt execute stopMovement Send request to mediator on_surgeForward_released ee Event handler for the Surge backward button As long as this button is pressed the platform will surge backward The request is passed to the mediator in this method void MainWindow on_surgeBackward_pressed Feedback m gt execute surgeBackward Send request to mediator 202 on_surgeBackward_pressed pe Event handler for the Surge backward button When the button is released the motion stops The request is passed to the mediator in this method af void MainWindow on_surgeBackward_released Feedback m gt execute stopMovement Send request to mediator on_surgeBackward_released pe Event handler for the Sway left button As long as this button is pressed the platform will sway left The request is passed to the mediator in this method void MainWindow on_swayLeft_pressed Feedback m gt execute swayLeft Send request to mediator on_swayLeft_pressed Event handler for the Sway left button When the button is
176. ket in readSet timeval tval Setup timeout tval tv_sec nTimeoutMs 1000 In seconds tval tv_usec nTimeoutMs 1000 1000 In useconds int selRet 1 Integer to hold return of select if nTimeoutMs amp gt 0 fx Wait for a datagram up to nTimeoutMs milli seconds Select returns 1 if datagram is waiting to be received and 0 if there is no datagram to be received selRet select this gt mSocket 1 amp readSet NULL NULL amp tval Mif if selRet amp gt 0 Some data is ready or user wants to wait for the datagram int bytesRecv recv this gt mSocket Socket to receive from char nData User defined buffer to store received data nDataSize Number of bytes to receive 0 No flags are needed if bytesRecv nDataSize If the expected number of bytes were read return true Return success MI if Mif return false Return failure ReceiveDatagram px This method is used to receive a double value over the network through a UDP datagram pp param nValue a pointer to a double variable to store the incoming datagram param nTimeoutMs the timeout in milliseconds to wait for the double value return amp lt true amp gt if a double value was succesfully received amp lt false amp gt otherwise ei bool UdpReceive ReceiveDouble double n Value long nTimeoutMs ex Check if Winsock is ready and socket is created if IsWsaReady false Il this gt
177. l is also of interest This coefficient limits the rate at which tilt coordination can occur Typically the scaling factors of a motion channel will range non inclusively from 0 to 1 Tilt coordination rate limiting is typically between 2 6 and 3 6 deges 3 to remain below the threshold of human sensory detection for rotational velocity It is harder to define a range for filtering equation coefficients since their transfer functions have many attributes such as form degree natural break frequency and damping ratio A complete study on the effect of each filter coefficient for the tuning of the classical WF algorithm can be found in 12 Based upon the findings of 12 it would be unproductive to create an experiment where a list of these filtering variables were selected and methodically tested in order to discover the best results As such this study is conducted with a sequential approach where observations are made between each test runs and where the parameters deemed appropriate for the next iterations are formulated from the results of the previous iterations In fact as reported by 12 an iterative process of filter tuning is the most appropriate form of optimization for the classical WF algorithm By observing the behaviour of a specific channel after each test run the filtering equations can be modified in a deliberate fashion to improve the quality of motions generated by the platform 5 2 4 Experimental Design In this study the
178. l testing of the flight simulator throughout its motion envelope at various speeds was also conducted to validate the installation As can be seen in Figure 54 the final result after the installation of most components remained very close to the concept drawings that were envisioned during the design phase 4 5 Summary This chapter looked at the implementation phase of this study More precisely it covered the implementation of the software by first explaining the goals and approach of the implementation phase Then the mediator design pattern an essential part of the software implementation was discussed In addition this chapter provided the reader with an understanding of the interactions between software components including the networking elements required to bring the system together The leveraging of the powerful Simulink environment and its role in the software implementation was also discussed Important references to more specific implementation material in the annexes were given to allow for a more exhaustive study of the software implementation Finally the implementation of the required hardware from the process of its fabrication to its installation onto RMC s 3DOF motion platform was detailed The final result was also shown to be very close to the system envisioned during the design phase 95 CHAPTER 5 EXPERIMENTAL STUDY AND PILOT VALIDATION 5 1 Introduction After having designed and implemented the desired flight
179. lator fidelity is achieved through a combination of software and hardware Some aspects of flight simulation fidelity can be studied through empirical evaluation of the system using sensors and other measurements 2 However these evaluation techniques have their limitations and can never replace the input of qualified pilots in evaluating subjective aspects of flight simulation i e control response motion sense visual quality etc As reported by Reid and Nahon 3 from the University of Toronto Institute for Aerospace UTIAS pilot evaluation is essential in obtaining the full picture in evaluating the fidelity of flight simulation More specifically this study aims to collect pilot feedback on four topics the hardware utilized on RMC s 3 DOF flight simulator the motion experienced the impact of motion simulation and the potential applications for the flight simulator The motivation of this study is to acquire and publish relevant data on the use of low cost 3 DOF flight simulation systems for use in the Canadian Forces Agencies such as Defense Research Development Canada DRDC Toronto and the Directorate of Land Synthetic Environment DLSE have already shown interest for the results of this study Summary information on volunteer pilots i e flight time flight qualification aircraft flown etc would be collected in order to describe the sample obtained and to study any possible impact of pilot qualifications on the data Pilot
180. lculate the accelerations at the pilot s position for each time step where Pi is a vector between the CG and the pilot s location amp g is the acceleration of the aircraft at the CG Ecc is the angular acceleration of the aircraft at the CG and where Wcg is the angular velocity of the aircraft at the CG By doing this the pilot s location with respect to the point where motions are measured is the same on the aircraft and on RMC s 3DOF F43 and Fs It is also the point Fs where the Crossbow IMU is installed on the flight simulator for this experiment Apitot Ace Ece X Pa Wee X cc X Pa 5 4 Using the C 172 three flight profiles representing the most prevalent manoeuvres in flight are recorded from the FSX environment The flight profiles are recorded from the 104 following manoeuvres a continuous climb a standard rate turn and a descent Each flight profile is of duration of 30 seconds The flight profiles were recorded using Simulink and manually flown by the author in the FSX environment A short description for each of the experimental flight profiles can be found in Table 12 Table 12 Experimental Flight Profiles Manoeuvres Manoeuvre Description Climb Entry into a continuous 1000 feet min climb from level flight From level flight into a 20 bank turn Standard Rate Tum and back into level flight Desni From level flight into a 1000 feet min descent and back into level flight
181. leased void on_surgeForward_pressed void on_m6down_released void on_m6down_pressed void on_m6up_released void on_m6up_pressed void on_m5down_released void on_mSdown_pressed void on_mSup_released void on_mSup_pressed void on_m4down_pressed void on_m4up_pressed void on_m4down_released void on_m4up_released void on_m3down_released void on_m3down_pressed void on_m3up_released void on_m3up_pressed void on_m2down_released void on_m2down_pressed void on_m2up_released void on_m2up_pressed void on_mldown_released void on_midown_pressed void on_mlup_released void on_mlup_pressed void on_heaveDown_released void on_heaveDown_pressed void on_heaveUp_released void on_heaveUp_pressed void on_rollRight_released void on_rollRight_pressed void on_rollLeft_released void on_rollLeft_pressed void on_pitchBackward_released void on_pitchBackward_pressed void on_pitchForward_released void on_pitchForward_pressed void on_park_clicked void on_center_clicked void on_markPositionParked_clicked void on_erasePosition_clicked 195 void on_speedSlider_ slider Moved int void on_stop_button_clicked void on_begin_button_clicked void on_checkBox_stateChanged int void on_clearButton_clicked void on_connectButton_clicked Main Window endif MAINWINDOW_H Mainwindow cpp pe
182. light simulator As part of the rating 140 questionnaire pilots were asked to specifically provide comments on potential applications for RMC s 3DOF flight simulator in flight training Table 31 summarizes the numerous suggestions and ideas which were identified by pilots for RMC s 3DOF flight simulator Overall there is a general consensus amongst pilots that the flight simulator could be used to perform numerous tasks that do not require a full size realistic cockpit It is important to mention that most pilots also cautioned that the system s flaws would need to be ironed out before it could be used commercially for any of those tasks In light of these suggestions a role could be attributed to the system and it could be redesigned and re tuned to optimally fulfill this role Table 31 Pilot Suggestions for Potential Training Use Pilot Suggestions for Potential Training Use To train procedures emergency response and intrument flying For ab initio pilot training on circuit procedure and airspeed management To practice a flight before actually doing it for real To train inadvertent entry in instrument meteorological conditions used to recreate the associated disorientation To be used as a glider simulator for Air Cadets 5 5 Summary A pilot validation exercise was designed to study the flight simulation system Best practise from the previous studies involving pilots were incorporated into the proce
183. locities in degrees s An FSX connect request is a subscription to some flight simulator variables When the tagged variables change in the flight simulator the program notifies a registered handle that is created in this class include FSXConnect h ps FSXConnect Constructor ay FSXConnect FSXConnect quit false Set the end of connection flag to false connected false There is no open connection at the instantiation hSimConnect NULL Set the SimConnect handle to NULL for now ex The ports to be used for the UDP data transfer over to MATLAB NOTE Ports below 1024 are priviledged and are not a good choice udpDataPorts 0 10101 udpDataPorts 1 10102 udpDataPorts 2 10103 udpDataPorts 3 10104 udpDataPorts 4 10105 udpDataPorts 5 10106 udpDataPorts 6 10110 udpDataPorts 7 10111 pitchBank 0 0 pitchBank 1 0 udpSend Initialize 127 0 0 1 UDP protocol object instance FSXConnect pe The MyDispatchProcPDR method is where the data sent from the simulator is handled accordingly This method is listening for events or data from FSX In this case it is listening for a pause or unpause event of the flight simulator in which case null data will be broadcasted Next it is listening for accelerations and angular velocity data from FSX When it is determined that such data has been received from FSX it is broadcasted on UDP on 6 different ports By def
184. lso contains a summary and analysis of the results 11 Chapter 6 presents conclusion drawn from the research and describes area of future work to build on the results obtained herein The thesis is written in a logical fashion intended to lead the reader through the steps that were taken during this study 12 CHAPTER 2 MOTION DRIVE ALGORITHM AND 3DOF ADAPTATION 2 1 Introduction A flight simulator motion platform is constrained by its physical limits and thus cannot exactly reproduce the accelerations and angular rates experienced in a real aircraft In order to generate the suitable controls for a motion platform it is recommended 11 to use a WF algorithm that accounts for this motion envelope Although there are many versions of WF algorithms the classical WF algorithm as reported by Reid and Nahon 8 9 and 10 is used in this study because of its relative simplicity and extensive use in commercial flight simulator Other algorithms such as the coordinated adaptative algorithm and the optimal controller algorithm exist but are far more complex to implement and tune than the classical WF algorithm 12 Furthermore in a 6DOF flight simulator study by the same authors 4 the classical WF algorithm has been reported to achieve very good results when compared to the other algorithms In this chapter the reference frames and axis system employed in the formulation of the motion drive algorithm is introduced This chapter als
185. m this sample run the low frequency accelerations are attenuated and the general amplitude of the signal is reduced 2 4 2 Rotational Motion Channel In the rotational motion channel the angular rates are scaled down by a factor which is determined in the fine tuning phase As discussed in the literature 11 a scaling value of 0 5 usually represents a good starting point The angular rates are limited to 30 degree per second due to the physical limitations for safe operation of the motion platform as reported by the manufacturer SSSI 6 RMC s 3DOF motion platform is only capable of rotational motion about the x and y axis As such rotational motion about the z axis is not necessary in the calculation of the high frequency Euler angles Bsu for the 3DOF version of the algorithm The r component of Euler angle rates must be zeroed to prevent the propagation of z axis rotational motion further into the algorithm where the Euler transformation matrix Lisis generated This step is necessary for the 3DOF algorithm because matrices Lys and Ts are formed from the output of the rotational motion channel and are both intended to reflect the simulator s current position Thus the propagation of 21 a z axis rotational motion component could cause the matrices to report unfeasible simulator positions including yaw which would be detrimental to the simulation As can be seen in Figure 10 the unnecessary r component representing z axis
186. ment was implemented The most widely used motion control algorithm for flight simulation the classical washout algorithm was implemented in the Simulink environment and modified for use with a 3 DOF motion platform The use of a powerful modelling environment such as Simulink greatly simplified the traditional implementation process of the control algorithm and also simplified the fine tuning process In this environment a special method polarity switch was also developed and implemented as a mean to reduce undesirable motion platform overshooting often associated with the use of flight simulation control algorithms This led to significant improvement of the flight simulator response The use of an inertial measurement unit combined with traditional techniques allowed for the effective tuning of the motion platform response This approach also showed that there is great potential for improved algorithms using hardware in the loop configurations and that robust automated tuning technique could also be developed using an inertial measurement unit A pilot validation exercise where several pilots flew the flight simulator showed the importance of pilot evaluation by highlighting the shortfalls and positive aspects of the system A rating questionnaire was developed to assess the performance of the flight simulator in key areas The results of this pilot validation exercise provided valuable information that contributes the 10 local kn
187. mmand to the motion platform Despite the implementation of software safety measures 42 it is also necessary to implement hardware safety to maintain control in the event of a software failure As such a physical emergency shut down switch cutting power to the motion platform is necessary This safety switch should also be in close proximity to the simulator operator to ensure a quick response Next the simulator should be positioned at a safe distance from potential spectators as well as the operator s station The manufacturer suggests that a safe zone of 30 inches around the platform should be created 6 As such zone of operation should be clearly identified and personnel should be away from the simulator when it is in motion To ensure maximum safety personnel involved with the simulator should receive a security briefing as well as an outline of the emergency procedures The security briefing is printed and placed on a wall near the operator s station A copy of the security briefing can be found in Annex A Finally it is paramount to ensure the safety of any passenger in the flight simulator For this reason a harness system should be installed to prevent the ejection of personnel from the platform Only through the establishment of rigid software hardware and personnel safety measures can the risk of operating a motion platform be minimized to acceptable levels 3 3 Constraints The design of RMC s 3DOF flight simulator is subj
188. mming terminology do not require any knowledge of the other objects of the system They only need to be able to interact with the mediator Many to Many Relationship Mediator Design Pattern FSXConnect FSXConnect Interacts with Interacts with FSX FSX A GUI Model GUI Mediator Model Manages User r gt Interacts with Manages User t gt Controls Object lt gt Interacts with Inputs Motion Platform Inputs Interactions Motion Platform y i UDP UDP Interacts with Interacts with Simulink i Simulink Classical j Classical Washout i Washout Figure 41 The Many to Many Relationship and the Mediator Design Pattern 73 As a result the software components are uncoupled from each other and a modification to any of the components in the system does not affect the others 19 As can be seen in Figure 41 the Mediator is an object that defines and controls how the objects of the system in a many to many relationship interact with each other For example if a user commands a motion by clicking a button on the GUI this command is passed to the mediator which then forwards it to the model The model has specific knowledge on the motion platform and is responsible for communicating with the motion platform and executing commands Hence once the model receives a command from the mediator it simply executes it with
189. mount is static and positions the seat at a height of 20 cm above the platform s surface In addition a 92 racing harness was bolted through the motion platform frame to secure the passenger to the seat Next an adjustable mounting system was designed and fabricated to allow for the installation of the pilot controls The adjustable control mounts allow for the positioning of the joystick and throttle depending on the pilot s morphology and preferences Finally an adjustable sensor mount was fabricated to position the TrackIR 5 sensor at the appropriate height depending on the position of the pilot s head Both the TrackIR 5 and joystick mounts had to be fabricated at a local machining shop due to their increased complexity Figure 52 Installed Special Mounting Parts All mounting components were made from aluminum with the exception of the throttle mount block which was fabricated out of wood by the author of this study Figure 52 shows the 93 different mounting parts on RMC s 3DOF motion platform Red arrows illustrate the directions in which the flight simulator components can be adjusted using the mounting parts Figure 54 Comparison between Concept Draw and Actual Implementation 94 Figure 53 shows the head mounted display with attached TrackIR 5 and infrared sensor The system s key components were fastened to their respective mounting parts and cables were managed to prevent any snag during the operation Initia
190. mulator As can be seen in Figure 18 the design sub objectives associated with software portability are divided in three categories Software Portability Programming Language Single Language Features Languages Object Oriented Simplicity Figure 18 Software Portability Design Sub Objectives First the use of a mainstream portable programming language with a strong presence on the market is deemed very important Indeed the use of such a language ensures that the widest possible range of programmers can operate maintain and modify the software Additionally mainstream languages are usually the ones that offer the best cross platform compatibility It is also preferable that the programming language selected for the implementation of the system be object oriented Advantages of object oriented programming such as their modernity abstraction and inheritance facilitate the modification and porting of software 15 Secondly to ensure 38 enhanced software portability the selection of proper development tools is essential As it is for programming languages the selection of mainstream development tools allows a larger number of programmers to easily modify and port the software to other platforms When selecting development tools features and supported languages are factors that are taken into consideration The best development tools usually support several programming languages and offer many interesting features
191. n Forces for the training of pilots The flight simulators in use often feature 6 degrees of freedom motion platforms and exhibit high levels of realism In this study a low cost flight simulator was designed and implemented using a 3 degree of freedom motion platform with low cost commercially available components The objective of this work was to investigate the merits and shortfalls of such a system This document describes the design implementation and analysis of this system Central to this work was the classical washout filter WF algorithm a widely adopted motion cueing algorithm used to generate the platform s motion An inertial measurement unit was used to study and optimize the classical WF algorithm Finally the system was assessed by collecting pilot ratings through a pilot validation exercise Results showed that most pilots were satisfied by the system and that the addition of motion to flight simulation greatly enhanced flight simulation Pilots also identified areas for improvements and potential applications for the flight simulator developed in this study The empirical analysis of the system using an inertial measurement unit and the ratings obtained through a pilot validation exercise indicate that low cost 3 degrees of freedom flight simulator have potential for flight training vi R SUM La simulation du vol est couramment utilis e par les Forces Canadiennes pour l entrainement des pilotes Les simulateu
192. n from the GUI to the motion platform is depicted In essence the information travels from the GUI to the C implementation that is the core of the system From there each component can be reached by the C implementation which is able to bring them together into a cohesive system that eventually sends the proper commands to the motion platform Figure 38 provides a top level general picture of the software implementation that will be discussed in this chapter Specific interactions relationships and networking between each component will be thoroughly discussed in the sections to follow or in the annexes 4 2 1 Purpose and Use Cases In order to properly conduct the software implementation phase it is important to understand the purpose and use cases of the system The goal here is not to go into any details of the implementation or the software architecture but to examine the fundamental tasks that the users of the system are trying to accomplish After having clearly identified those use cases it is easier to implement a solution The purpose of the software in this research is to allow the simulator operator to interact with the motion platform in order to conduct flight simulation To empower the simulator operator to do this the software must provide several key functionalities Use case diagrams are commonly used to illustrate the functionalities of a system in terms of its users 17 In addition use case diagrams are a part of the
193. n light of the literature experimental results and pilot comments to formulate recommendations for future work on flight simulators at RMC 6 2 1 Classical WF Algorithm Improvement In this study the use of an inertial measurement unit was introduced to measure the accelerations and angular rates experienced on the motion platform However it is believed that 144 such a device can play a more active role as part of the WF algorithm Most WF algorithms assume a perfect simulator response to motion commands In reality this is not the case as small variations exist between the actual and assumed platform position Because the motion platform is constantly washing out to its central position there is no troublesome error accumulation and this formulation works relatively well In this case the true position of the motion platform can be obtained directly from the Galil motion controller and the motions can be measured with the IMU 320 It is believed that the classical WF algorithm or some other form of WF algorithm could be modified to use this hardware in the loop configuration to improve the motion simulation Intelligent approaches such as the use of human inner ear otholith models as measures of performance combined with fuzzy logic and genetical algorithm as reported by 31 could be studied and modified to develop an algorithm capable of automatically tuning its filter equations from the actual motions experienced on the motion platform
194. n platform frame Aircraft bank angle Euler angle rates of the rotational motion channel High pass filtered Euler angle rates of the rotational motion channel Individual actuator displacement Location of Fs with respect to F7 Low frequency tilt coordination Euler angles Simulator Euler angles 0 4 High frequency simulator Euler angles Low frequency simulator Euler angles Simulator low frequency Euler angles Angular acceleration Aircraft body axis reference frame Aircraft accelerations a ay az g Aircraft center of gravity reference frame Earth fixed inertial reference frame Low frequency aircraft accelerations Input of the tilt coordination channel Intermediate earth referenced accelerations Aircraft pilot s head reference frame Simulator pilot s head reference frame Simulator body axis reference frame Simulator accelerations a ay az g Acceleration due to gravity 0 0 9 81 ms Ixyz xviii Aircraft acceleration due to gravity gx 8y 8z Scaling coefficient of the scaling and limiting process Limiting value of the scaling and limiting process Individual actuator length Euler rotation matrix transforming accelerations from F4 to Fr Location of F4 with respect to Fg Location of Fpa with respect to F4 Location of Fps with respect to Fs Response variables for accelerations Response variables for angular rates Simulator translational displacement d dy dz Location of Fs with respe
195. nability can be seen in Figure 20 Flight Simulator Documentation Code Commenting Figure 20 Maintainability Design Sub Objectives A high level of maintainability allows others to adapt and evolve the software in accordance with their needs Considerations of maintainability must also be applied to the hardware used in the design of the flight simulator Indeed any equipment used as part of the flight simulation hardware should be thoroughly documented and understood by the users To promote this comprehension a flight simulator manual should be created for the end user Such manual should at least include sections on the characteristics and limitations of the system the operation of the system the maintenance and the safety considerations For this reason it is important to select equipment that comes sufficiently documented in order to increase its maintainability as part of the flight simulation system When acquiring equipment to be used in the flight simulator durability and low maintenance are also essential Selecting equipment that has high maintainability increases the maintainability of the system as a whole 3 2 6 Safety When designing systems that interact with humans safety is always a primary concern RMC s 3DOF motion platform is quite capable of generating extreme accelerations that could 41 damage equipment and injure personnel 6 As such safety is a very important design objective
196. o allow the passenger to rest his feet However as the motion platform is not capable of yaw motion it is deemed that the installation of rudder pedals is not necessary The flight simulation software can replace rudder pedals input as it allows for the configuration of automated rudder synchronization with the roll of the aircraft to produce turns Although Figure 34 shows a potential anchoring system it would be hard to 61 design a system that allows for the safe anchoring of the rudder pedals while providing adjustments As such rudder pedals are not part of the standard simulator configuration for this study Nevertheless should there be a need for rudder pedals the dimension of the feet holder as well as mounting holes allow for a quick installation of the latter The opinion of pilots with regards to the absence of rudder pedals can be found in Chapter 5 Finally an adjustable height pole supporting the infrared tracker is also present on the motion platform This pole is adjustable in order to allow passengers of various heights to sit with the tracker levelled to their head 3 4 4 Simulator Operation and Safety The operation of the flight simulator and safety are closely related The simulator is designed to be operated by an operator controlling the simulation at a PC based control station while the passenger undergoes the flight simulation The control station is connected to the motion platform controller by an Ethernet link An
197. o consist of a review of the 6DOF classical WE algorithm as developed and studied extensively at the University of Toronto Institute for Aerospace Studies UTIAS 4 8 9 10 11 and 12 As such it will be reviewed in detail for each channel Finally the modifications required to implement the classical WF algorithm to 13 RMC s 3DOF motion platform are discussed in this chapter as they are better understood in the context of the original algorithm 2 2 Reference Frames and Axis System When discussing any WF algorithm it is important to define the reference frames and axis system that are used in the formulation of the algorithm to describe the relative position of objects in space In this case the objects of interest are the aircraft and the flight simulator As such the reference frames and axis system as reported by 8 9 and 10 are used in this study and can be seen in Figure 3 Aircraft Fp Xp Simulator Xps Figure 3 Reference Frames and Axis System Frame F7 the inertial reference frame is Earth fixed with its z axis parallel to the earth gravity g Fr is also associated with both the aircraft and the simulator The right hand orthogonal axis system is used in this study In this axis system the x axis can be described as pointing towards the nose of the aircraft and the y axis orthogonal to the x and z axes as 14 pointing out the right wing of the aircraft The z axis is described as pointing downwar
198. o design a system using commercially available components at a very low cost that would achieve acceptable levels of performance for aircraft simulation To obtain maximum performance from RMC s 3DOF the underlying principles of the classical washout filter WF algorithm were carefully studied The classical WF algorithm was selected due its wide use in commercial simulation simplicity and favourable pilot response as reported in the literature The classical WF algorithm was then modified to be suitable for 3DOF use and implemented using the Simulink environment Microsoft s Flight Simulator X was selected as the virtual environment of choice due to its realism in all areas and its extensive software development kit A C executable was developed using VS2008 and makes use of a software design pattern the mediator pattern in order to interface and control the flight simulator components The entire system was designed using a design process where design objectives and constraints were studied This led to a final system which was implemented as envisioned in the original design Finally an experiment was designed to fine tune the classical WF algorithm and to collect performance data The thorough evaluation of the flight simulation capabilities of RMC s 3DOF flight simulator through an experimental study and a pilot validation exercise led to a better understanding of RMC s 3DOF motion platform as a tool to create flight simulation 143
199. on By controlling the length velocity and acceleration independently for each actuator the platform can be positioned in space and generate ranges of roll pitch and heave motions The motion platform displacement velocity and acceleration limits which determine the motion platform envelope are given in Table 1 These recommended motion platform limits are provided by the manufacturer SSSI 6 Table 1 RMC s 3DOF Motion Platform Limits Roll Pitch Heave Displacement 25 30 0 20 m Velocity 30 s 30 s 0 45 m s Acceleration 350 s 350 s 7 85 mes The displacement velocity and acceleration of the actuators are controlled by a high performance Galil Accelera Series 30x40 motion controller The motion controller comes with native C libraries and a COM interface that allows for the use of Windows programming languages such as Visual Basic VB and C It can operate in stand alone mode or it can be interfaced with a computer through Ethernet or RS 232 serial port cables The electric motors are driven by a Yaskawa F7 Drive acting as frequency converter with encoder feedback In this system the speed at which motors move is a function of the voltage frequency The position of each actuator is extrapolated from the encoder feedback feature of the Yaskawa F7 Drive To power the motion platform it is necessary to have a 240 volt AC 3 phase current 6 The Galil motion controller is powered by a regula
200. on request to FSXConnect Receive the data by UDP and pass it to the model if udpVelocities ReceiveDoubleArray velocities number_of_actuator 1 true Setting 1 for the timeout acts as a spinlock will wait undefinitely m gt velocity_command velocities Sending the velocities to the model WI if if udpAccelerations ReceiveDoubleArray accelerations number_of_actuator 1 true Setting 1 for the timeout acts as a spinlock will wait undefinitely m gt acceleration_command accelerations Sending the accelerations to the model Wif 209 if udpLengths ReceiveDoubleArray lengths number_of_actuator 1 true Setting 1 for the timeout acts as a spinlock will wait undefinitely m gt length_command lengths Sending the lengths to the model HI if MotionSimulation Model h pe Model h TIDE Visual C 2008 Express Edition author 2Lt Jean Le Bouthillier Latest Revision 11 May 2010 ek pragma once ifndef MODEL_H define MODEL_H include Galil h include amp lt string amp gt include amp lt iostream amp gt include amp lt sstream amp gt include amp lt math h amp gt using namespace std class Model public void change_speed int s string connect string ip string process_user_input string uInput void length_command double command void velocity_command double command void acceleration_command double command bool is_platform_moving private pee
201. one undividable packet of data When doing recv recvfrom you are waiting for one undividable packet of data Everything else is then completely dependent on the specific application protocol ek pragma once ifndef _SP_UDP_H_ define _SP_UDP_H_ include amp lt stdlib h amp gt include amp lt string h amp gt include amp lt winsock2 h amp gt include amp lt Ws2tcpip h amp gt include amp lt windows h amp gt typedef unsigned short PortNumber Convenience typedef for Port parameter class Udp public Udp Udp bool IsOkQ private pee 217 Flag used to hold the success of Winsock Initialization bool success j bool IsWsaReady Method to query Windows sockets status endif Udp cpp ee Udp cpp IDE Visual C 2008 Express Edition author Martin Vidovic Modified by 2Lt Jean Le Bouthillier Latest Revision 13 November 2009 ae Description A helper class to make web service adressing WSA Windows socket Winsock initialization and finalization easier and more reliable include Udp h pe WsaHelper Constructor Udp Udp WSAData wsaData WSAData used to store windows socket initialization informations pe Tnitialize windows socket 0 is returned if there are no errors in the initialization this gt success WSAStartup MAKEWORD 2 2 amp wsaData 0 WsaHelper px WsaHelper Destructor Udp UdpQ if this
202. orm Actuator Assembly 5 Figure 3 Reference Frames and Axis System eesesssseseresesseesressrserssresseseresresreserestessesetesresseeee 13 Fig r 4 Classical WF Algonthmi Glass cele tan anda nt nn 15 Figure 5 Example of Tilt COGTdINAIONS Een nn es tn nt Me es anis 19 Fieure 6 Scalhne and Limiting O2 anis nt nn nn ne ni 20 Figure 7 3DOF Translational Motion Channel 24 Figure 8 Translational z Component High Pass Filter wn 4 rad S C wo eeeeceesseeeteeeseeeeeeeenees 25 Figure 9 Translational Channel High Pass Filter Sample Run wn 4 rad sec s es 25 Figure 10 3DOF Rotational Motion Channels nanrntimiaeeianiite 27 Figure 11 Rotational Channel High Pass Filter wn 1 rad Sec eeseseeeeseeerererserererreesees 27 Figure 12 Rotational Channel High Pass Filter Sample Run wn 1 rad sec oe eens 28 Figure 13 Tilt Coordination Channel x Component Low Pass Filter wn 5 rad sec 29 Figure 14 Tilt Coordination Low Pass Filter x Component Sample Run 29 Figure 15 Objective Tree for RMC s 3DOF Flight Simulator 32 Figure 16 Realistic Simulation Design Sub Objectives ss 34 Figure 17 Cost Effectiveness Design Sub Objectives ccceesscessceceeneeceeeeeceeeeeceeeeecneeeeneeeee 36 Figure 18 Software Portability Design Sub Objectives ss 37 Xi Figure 19 Ease of Operation Design Sub Objectives 0 ieee ceeceeeseecseceneeeeeeeeaeecsaecneenseeeeaees 39 Figure 20 Maintainability De
203. ot drive the simulator to its limits as easily Another important role of the filters is to slowly and continuously bring the motion platform back to its central location 12 The classical WF algorithm uses both low pass and high pass filters to accomplish this task Equations 2 9 will be recognized as that of a 2 order high pass filter 2 10 as that of a 1 order high pass filter and 2 11 as that of a 2 order low pass filter 11 These equations describe the Laplace domain transfer functions of the different filters used in each channels of the classical WF algorithm 2 Translational channel filter Box 3 2 9 s2 2 wns w2 Rotational channel filter Box 10 2 10 StOn 22 On Tilt coordination channel filter Box 5 eee n n 2 11 Choosing the appropriate parameters for the filters can only be done with a trial and error tuning process combined with pilot feedback 12 Several false cues can be generated by improperly tuned filters 12 As such the motion platform designer must perform several tests by isolating each filter and ensuring a proper behaviour depending on the application The absence of a filter damping ratio a parameter used to characterize the frequency response of filters for first order filters makes them easier to fine tune Second order filters provide a much stronger attenuation power but can lead to other problems such as overshooting when the signal exceeds its final target value dur
204. ouble and send value over UDP Polarity is inverted for the roll since FSX uses different convention than the SAE right hand rule thisFS XConnect gt udpSend SendDouble thisFS XConnect gt udpDataPorts 3 double pS gt datum count value break case DATA_Q_ANGULAR_VELOCITY Cast into double and send value over UDP Polarity is inverted for the pitch since FSX uses different convention than the SAE right hand rule thisFS XConnect gt udpSend SendDouble thisFS XConnect gt udpDataPorts 4 double pS gt datum count value break case DATA_R_ANGULAR_VELOCITY Cast into double and send value over UDP Polarity is inverted for the pitch since FSX uses different convention than the SAE right hand rule thisFS XConnect gt udpSend SendDouble thisFS XConnect gt udpDataPorts 5 double pS gt datum count value break case DATA_P_AIRCRAFT_PITCH Cast into double and send value over UDP Polarity is inverted for the pitch since FSX uses different convention than the SAE right hand rule thisFS XConnect gt udpSend SendDouble thisFS XConnect gt udpDataPorts 6 double pS gt datum count value break case DATA_B_AIRCRAFT_BANK Cast into double and send value over UDP Polarity is inverted for the pitch since FSX uses different convention than the SAE right hand rule thisFS XConnect gt udpSend SendDouble thisFS XConnect gt udpDataPorts 7 double pS gt datum count value break default printf nUnknown d
205. out any descriptions for many of its variables This makes it hard to know which data is relevant for the motion simulation Furthermore the rate at which data can be extracted from the X Plane 9 environment sampling rate is fixed at 25 Hz This fixed sampling rate does not allow for any tuning in the event that this sampling rate is insufficient On the other hand FSX features a variable sampling rate that allows for tuning Overall FSX and X Plane 9 offer similar simulation capabilities but FSX is more developer friendly than X Plane 9 The implementation of RMC s 3DOF flight simulator warrants the writing of a plug in to extract relevant information from the simulation environment In this case the aircraft specific 51 accelerations are available without the gravity vector g fa a ay a in mes and the angular rates are available in degreess or radianes Since the aircraft accelerations are not available with the effect of gravity it is convenient to calculate the effect of gravity from pitch and bank information that are also available from FSX in degrees or radians The ability to specify the sampling rate in FSX is also an important feature For these reasons and because of its superior development capabilities FSX is the simulation environment of choice for this study 3 4 2 Visual Display and Sound Several options are available for the generation of visual and audio cues for the flight simulator However due to hard
206. out any knowledge or care for its origin Then the model returns information on the failure or success of the command to the mediator which sends it back to the GUI which is also unaware of the true origin of this information The use of a model in this implementation enables the GUI or any other components to send motion commands without any knowledge of the communication protocols required to communicate with the motion platform Because of this abstraction it is very easy to maintain and change the system with minimal effort The same condition exists for each of the four objects depicted in Figure 41 as they all benefit from the abstraction layer created by the Mediator design pattern A complete UML class diagram of the software implementation using the Mediator design pattern can be found in Annex D The complete C code can also be found in Annex K 4 2 4 Flight Simulator X Plug in To achieve a realistic simulation it is essential to provide the flight simulator passenger with high quality visual and audio inputs In addition it is even more important to generate aircraft translational accelerations and angular rates that are realistic and can be used to calculate the motion platform displacements After having decided that FSX is the flight simulation 74 environment of choice for this study it is necessary to find a way to extract the translational accelerations and angular rates in order to send them to the classical WF algorithm
207. owledge base on flight simulation and will allow for to the successful implementation of superior flight simulators at the Royal Military College of Canada Finally the work conducted in this thesis contributed greatly in displaying the capabilities of the Royal Military College of Canada to various organisations such as Defence Research and Development Canada DRCD and Directorate Land Synthetic Environment DLSE The Canadian Association of Computer Science organized a national Computer Science Education Week between Dec 5 11 2010 in which 26 Canadian Universities participated the Flight Simulation with Motion Platform out of this research was the leading event representing RMC to the nation 1 7 Thesis Structure The current chapter provides the reader with background information on flight simulation in general RMC s 3DOF motion platform and the purpose of this study The remainder of this thesis is structured as follows Chapter 2 reviews the classical WF algorithm in detail and explains the necessary adaptation of the algorithm from 6DOF to 3DOF Chapter 3 covers the considerations associated with the design of a flight simulator and explains the design of the flight simulator Chapter 4 describes the implementation of the flight simulator as well as the equipment and techniques used Chapter 5 describes the experiment conducted to optimize and study the flight simulation capabilities of the system This chapter a
208. param command a double array containing the three actuator velocity gi void Model velocity_command double command for int i 0 i amp lt snumber_of_actuator i command i conversion_factor Converting metric to 3DOF units for g gt command make3DOFCommand SP command Send the new speed to the controller velocity_command px Used to send an actuator acceleration command to the motion platform ek param command a double array containing the three actuator accelerations void Model acceleration_command double command for int i 0 i amp lt number_of_actuator i command i conversion_ factor Converting metric to 3DOF units command i fabs commandfi AC and DC must be positive only for make3DOFCommand AC command Send the new acceleration to the controller make3DOFCommand DC command Send the new decceleration to the controller acceleration_command pe Used to know if the motion platform is currently moving ek return amp lt true amp gt if the platform is moving and amp lt false amp gt if the platform is not moving bool Model is_platform_moving bool answer true By default we will say the platform is in motion string platformInfo String to receive the platform information if platformType 3DOF If connected to the 3DOF platformInfo g gt command TV Use the TV tell velocity command to get a reading if platformInfo 0
209. presented in both languages You must obtain RMC REB approval even if REB approval has been given at another institution You may be required to complete this application please contact the chair for advice Any questions regarding the necessity for ethics review should be directed to the RMC REB Chair You may need DHRRE approval if you or your human subjects fall outside of RMC jurisdiction Further information can be obtained from the Chair of RMC REB Please submit one electronic copy or three hard copies of the submission to the Chair or to the office of the Research Services The research project cannot proceed until ethics approval is obtained from the RMC REB 178 A GENERAL INFORMATION 1 Title of Project The Design Implementation and Study of a Low Cost 3 Degree of Freedom Flight Simulator using RMC S 3 DOF Motion Platform Indicate the anticipated commencement date for this project 25 October 2010 Indicate the anticipated completion date for this project 05 November 2010 2 Applicant Dept Ext Email 2Lt Jean Le Bouthillier Mathematics and 613 583 8778 jean le bouthillier rmc ca Computer Science Applicant s status Faculty Supervisor Graduate Student Dr Yawei Liang 3 Co Investigators Dept Ext E mail 4 Level of Project L Faculty Research gt lt Graduate Research C Undergraduate Research C Other specify 5 Funding Status Is this project currently fun
210. r 120 volts AC outlet The motion platform system comes with testing and demonstration software allowing the user to manually control the platform from a graphical user interface GUI This software is rudimentary and does not prevent a user from causing damage to the motion platform This lack of safety was experienced first hand and is discussed later in Chapter 4 Other than the demonstration package no control software or software documentation is provided with the system Additionally the motion platform system was not delivered with a simulation environment or motion cueing software However the Galil motion controller and its libraries allow for flexibility in interfacing with the platform 1 4 Motivation The Canadian Forces CF is currently operating several simulation systems for the training of pilots and drivers of a multitude of platforms Flight simulators are routinely used in the training and periodic skill checks of aircrew personnel Additionally there is an ongoing effort by the Directorate of Land Synthetic Environment DLSE a branch of the Land Force Doctrine and Training System LFDTS to assess develop and integrate new simulation technologies in the Army The use of simulation in the CF training system is cost effective safe and in most cases is also environmentally friendly According to the LFDTS it is believed that an increase in CF simulation capabilities will be essential to sustain the growth of a mo
211. rasePosition_clicked string status m gt execute erasePosition Send request to mediator if status Cannot execute command Check if the command can be executed Feedback status Let the user know the status ui gt erasePosition gt setDisabled true Disable the Erase Parked Position button ui gt markPositionParked gt setEnabled true Enable the Mark Position as Parked button Mif else If the command cannot be executed Feedback status Let the user know the status else on_erasePosition_clicked pe Event handler for the Mark Position as Parked button The request is passed to the mediator in this method oF void Main Window on_markPositionParked_clicked string status m gt execute markPositionParked Send request to mediator if status Cannot execute command Check if the command can be executed Feedback status Let the user know the status ui gt markPositionParked gt setDisabled true Disable the Mark Position as Parked button ui gt erasePosition gt setEnabled true Enable the Erase Parked Position button Mif else If the command cannot be executed Feedback status Let the user know the status else on_markPositionParked_clicked pe Event handler for the Center button The request is passed to the mediator in this method void MainWindow on_extend_clicked Feedback m gt execute extend Send request to mediator
212. ratings would be collected anonymously as pilot personal information is of no value for the evaluation A rating questionnaire developed at the UTIAS and modified at RMC will be used to collect pilot ratings References 180 1 Pouliot N A Gosselin C M Nahon M A Motion Simulation Capabilities of Three Degree of Freedom Flight Simulators Journal of Aircraft Vol 35 No 1 January 1998 pp 9 17 2 Reid L Nahon M Simulator Motion Drive Algorithms A Designer s Perspective Journal of Guidance Control and Dynamics Vol 13 No 2 1990 pp 356 362 3 Reid L D Nahon M A Response of Airline Pilots to Variations in Flight Simulator Motion Algorithms Journal of Aircraft Vol 25 No 7 July 1988 2 Methodology Procedures Provide a brief sequential description of the procedures to be used with human participants in this study Include a description of measures that will be used e g surveys questionnaires interviews focus groups audiotaping videotaping unobtrusive observations etc Attach a copy of all instruments and assessment tools or provide sample items if complete tools are unavailable Participants will be able to voluntarily choose to participate in the study by arranging an appointment at their convenience with the research staff A questionnaire Annex A will be used to collect ratings immediately after the conduct of test runs Then a short disc
213. rces X specify RMC Flying Club DND military personnel X DND civilian personnel X Canadian Government personnel E 182 School Boards Correctional Services _ Agencies _ Mailing Lists _ Businesses industries professions O Health care settings long term care facilities etc O Other specify _ d Describe how and by whom the potential participants will be recruited Attach a copy of any materials to be used for recruitment e g posters flyers advertisement letter telephone script Members of the RMC Flying club will be asked to participate DND employees and DND civilian personnel known to have flight experience will also be asked to participate Contact will be made in person by phone or by email in all cases No formal material will be used for contacting participants However solicitation will always include the following information 1 Description of the study 2 Tasks to be conducted by the participants 3 Time involved to complete the tasks C POTENTIAL BENEFITS TO PARTICIPANTS 1 If participants will receive any reward financial or other for participation provide details n a 2 Describe any other known or anticipated direct benefits to the participants from their involvement in the project Participants will have the opportunity to fly a realistic flight simulator featuring motion simulation and head tracking Interested participants will be
214. rcraft The selection of a pilot seat is driven by the designs objectives and constraints To create a realistic and comfortable environment an adjustable seat suited for individuals of various heights must be installed on the motion platform This seat should be centered on the motion platform in accordance with the manufacturer s center of gravity recommendations In addition a lighter seat is preferable in order to limit the forces it will generate as the motion platform accelerates It must be fastened to the motion platform in a way that guarantees it can resist the strong accelerations of the latter Furthermore this seat must provide a harnessing system to effectively secure the occupant on the platform and to prevent potential ejection under high acceleration manoeuvres Seats that are compatible with the aforementioned requirements fall under three categories actual aircraft seat car seats and seats specifically designed for aircraft simulators The cost of actual aircraft seats is very high and they are difficult to obtain Actual aircraft seats 55 are also likely to be bulky and quite heavy depending on their origin The design of a custom seat for this study is not a feasible option due to the cost and unavailability of manufacturing equipment On the other hand car seats are readily available and come in various configurations Designed with the safety of passengers in mind car seats feature strong anchor points structural integri
215. re implementation by means of technical and user reliant testing will be discussed 42 Software Implementation The software implementation of the flight simulation system encompasses four main software components and utilizes both UDP and TCP networking protocols to establish communication between certain components oa Software Hardware We Flight Simulator X KL SUIS SimConnecth k C Implementation lt lt includes gt gt mainwindow h 4 FSXConnect cpp Galil h 2a i i A LT TN ae lt lt includes gt gt 6 3 lt lt includes gt gt a i A 7 ee Motion a Graphical User Interface 7 1 gt mainwindow cpp 2 Mediator cpp 11 gt Model cpp 12 Platform Ss PA ca Ken UdpSend cpp 10 UdpReceive cpp I J lt cincludes gt gt lt lt includes gt gt i Po Figure 38 General Interaction of Software Components and Flow of Information ooo Washout Algorithm Simulink g Se J DA 68 The four main software components are the GUI Flight Simulator X the classical WF algorithm and the C implementation Together these software components form the flight simulation system In Figure 38 the general interaction of the four main software components can be seen In addition the typical flow of informatio
216. reate an optimal degree of fidelity 5 3 5 Test Run 3 The configuration of the third test run can be found in Table 23 where changes from the second test run are highlighted in red The main objective of this test run is to attempt to gain fidelity on pitch motions which were identified as underpowered in the previous run Table 23 Classical WF algorithm Configuration Test Run 3 Parameters Scaling Limiting TUA Misc Parameters ax 1 wn 5 0 1 ay 1 On 8 1 _ _ Overshooting me met ae Management p 0 6 wn 1 q Os On 1 r Wn 1 Tilt _ 3 _ Coord maisi 4 5 deges Gain x 5 Gain y 6 As per the protocol the author flew several flight manoeuvres to study the motion response of the algorithm A summary of pilot observations can be found in Table 24 Overall it is felt that an increase of the scaling coefficient for x axis angular rates and tilt coordination had a positive effect on the motion generation Table 24 Pilot Observations Test Run 3 Motion Channel Observations Translational Z axis accelerations are not too powerful and contribute nicely to the overall simulation Tilt Coordination Tilt coordination 1s present and proportional Rotational Rotational motion is proportional for both pitch and roll motion Pitch motions are in tune with z axis accelerations 131 The response variables were calculated and
217. released the motion stops The request is passed to the mediator in this method void MainWindow on_swayLeft_released Feedback m gt execute stopMovement Send request to mediator on_swayLeft_released px Event handler for the Sway right button As long as this button is pressed the platform will sway right The request is passed to the mediator in this method void MainWindow on_swayRight_pressed Feedback m gt execute swayRight Send request to mediator on_swayRight_pressed px Event handler for the Sway right button When the button is released the motion stops The request is passed to the mediator in this method void MainWindow on_swayRight_released Feedback m gt execute stopMovement Send request to mediator on_swayRight_released pe Event handler for the Yaw left button As long as this button is pressed the platform will yaw left The request is passed to the mediator in this method ef void MainWindow on_yawLeft_pressed Feedback m gt execute yawLeft Send request to mediator on_yawLeft_pressed px Event handler for the Yaw left button When the button is released the motion stops The request is passed to the mediator in this method void MainWindow on_yawLeft_released Feedback m gt execute stopMovement Send request to mediator on_yawLeft_released
218. rom little endian PC format to big endian network format this gt mSourceAddr sin_port htons nSourcePort pee Our socket can be bound to any IP address The Datagram will be assigned a source IP address from the egress network interface this gt mSourceAddr sin_addr s_addr inet_addr nSourcelp ex Bind source address to our socket 1 int bindResult bind this gt mSocket sockaddr amp this gt mSourceAddr sizeof this gt mSourceAddr if bindResult 0 If bind failed for some reason this gt SocketCleanup Close the current socket return false Return failure Mif return true Initialize px This method reads one UDP datagram from the network and puts the information from this datagram in a data buffer at a specified memory adress 4 param nData pointer to a valid data buffer to receive datagram bytes param nDataSize contains the number of bytes available in nData buffer param nTimeoutMs sets the timeout of this method if no datagram is available 1 is infinite timeout return amp lt true amp gt if nDataSize bytes were successfully read amp lt false amp gt otherwise bool UdpReceive ReceiveDatagram void nData size_t nDataSize long nTimeoutMs fd_set readSet Initialize the fd_set structure used for waiting FD_ZERO amp readSet Initializes readSet to have zero bits for all file descriptors FD_SET this gt mSocket amp readSet Set the bit for mSoc
219. rotational motion channel high pass filter can be seen in Figure 12 where the results of a sample run are shown As can be seen from this sample run the amplitude of the Euler angle rate signal is attenuated It can also be observed that the attenuation is less than that of the high pass filter in the translational motion channel This can be attributed to the slope of the gain magnitude plot which is less pronounced for this filter when compared to the filter used in the translational motion channel 2 4 3 Tilt Coordination Motion Channel The tilt coordination channel relies on rotation about the x and y axis which RMC s 3DOF can generate As such the tilt coordination channel does not require any modification to work with RMC s 3DOF motion platform In this study the tilt coordination coefficient is 29 determined experimentally and kept below the suggested threshold values 11 This process is covered in Chapter 5 The x component low pass filter Bode plot can be seen in Figure 13 Bode Diagram D D roy o T T zi Magnitude dB for O T 80 L y bor p peek M TE por peek r ror pope 100 45h 90 Phase deg 135 180 FE TTE LITT 10 Figure 13 Tilt Coordination Channel x Component Low Pass Filter w 5 rad sec Frequency rad sec 1 ter L T L a T T T T ja a ft fi f 1 T 05 nn G bo E j l c
220. rotational motion of Euler angle rates By is discarded in the high pass filter step Box ten Ben p q 0 Figure 10 3DOF Rotational Motion Channel In the rotational motion channel Euler angle rates are filtered using a first order high pass filter As discussed in Section 2 4 1 the high pass filtering in the rotational motion channel is also performed in frame Frin order to minimize offsets and errors Figure 11 shows the Bode plot for the first order high pass filter used to filter Euler angle rates Bode Diagram Or es n_e 20 i 30 al Magnitude dB 40 a 50 r poop peek r por op peek r poor op ph r por re peek 90 TE eee MERE 45h Sg 4 Phase deg OE ror ee ee ror rrrrrr ee eee aaa 10 10 10 10 10 Frequency rad sec Figure 11 Rotational Channel High Pass Filter w 1 rad sec 28 3 FT T T L T T T 3 T T R 2r Y Q 2 1 AN TD Se g O D J D lt I 5 1b j 4 3 J Lu 2 Input LS Filtered 3 L T T T T T i T tc i ra 0 2 4 6 8 10 12 14 16 18 20 Time sec Figure 12 Rotational Channel High Pass Filter Sample Run wp 1 rad sec As can be seen from the gain magnitude plot a negative gain is applied to the signal as its frequency lowers For this reason fine tuning with the scaling values of the rotational channel is indicated to obtain the overall desired scaling effect The general behaviour of the
221. rs utilis s ont la plupart du temps des plateformes de mouvement avec 6 degr s de libert et ont un niveau de r alisme tr s lev Dans ce m moire un simulateur de vol faible co t fut developp avec une plateforme de mouvement avec 3 degr s de libert ainsi que des composantes disponibles commerciallement Le but de cette tude fut d investiguer les m rites et inconv nients d un tel syst me Ce document d crit le d sign l impl mentation et l analyse de ce syst me L algorithme classique de filtrage un algorithme de mouvement tr s populaire fut utilis pour g n rer les mouvements de la plateforme Un senseur de mouvement fut utilis pour tudier et optimiser l algorithme classique de filtrage Finalement les valuations de pilotes furent receuillies par l entremise d un exercise de validation Les r sultats de cet exercise ont d montr s que l addition de mouvement la simulation du vol a un impact tr s positif Les pilotes ont galement formul s des suggestions pour am liorer le syst me et ils ont identif s des applications potentielles pour le simulateur de vol d velopp dans cette tude L tude empirique du syst me avec un senseur de mouvement ainsi que les valuations obtenues par l entremise d un exercise de validation indiquent que les simulateurs avec 3 degr s de libert faible co t ont un potentiel pour l entrainement au vol vii TABLE OF CONTENTS Page A
222. ry modifications for the implementation of the classical WF algorithm Additionally the scaling limiting filtering and actuator motion calculation steps were discussed for the 3DOF version of the algorithm With an understanding of the algorithm and limitations of RMC s 3DOF motion platform a flight simulator can be designed and implemented 31 CHAPTER 3 DESIGN 3 1 Introduction With an understanding of flight simulation RMC s 3DOF motion platform and the classical WF algorithm a plan can be developed to build a flight simulator that takes advantage of the available hardware As this study is aimed at determining the capabilities of RMC s 3DOF motion platform and its associated classical WF algorithm it is necessary to design a flight simulator that facilitates the conduct of the study while taking advantage of the motion platform As such the design of RMC s 3DOF flight simulator is done through a systematic approach keeping in mind a series of pre defined design objectives that are conducive to the accomplishment of this goal Furthermore as resources are limited and the equipment has specific limitations the design process is also influenced by a set of constraints that are inherent to the system In the design process available equipment design objectives and constraints are taken into consideration in order to implement a flight simulator that effectively balances all criterions while meeting the study goals In th
223. s for each motion channel are discussed in the order in which they were conducted As predicted in the experimental design several runs were required to reach a satisfactory level of performance The optimization of each motion channel was accomplished and lead to a higher fidelity in motion simulation Due to the sequential nature of this optimization process the analysis of results was conducted immediately after each test run to derive the coefficients required to improve performance 5 3 1 Reproducibility and Repeatability Considerations The experiment in this study is essentially designed to measure and improve the accuracy of the flight simulator However before measuring accuracy the flight simulator s precision must be measured to ensure the trueness of data collected during the experiment In this case precision or reproducibility represents the degree to which the flight simulator will generate the same motions for each run of an identical flight profile Ideally the motion simulator should always generate exactly the same motions between each run of an identical flight profile If it is the case the experimenter then knows with certitude that recorded variations between each run are caused solely by changes to the coefficients of the system that are modified and not by randomness or instability in the system In reality small variations make their way to the system 109 and the motions generated by the flight simulator are sligh
224. s outfitted with a spring that creates resistance as the yoke is moved away from its neutral position Figure 33 Logitech s G940 Flight System Of the system reviewed the Flight System G940 as can be seen in Figure 33 was judged to be the best input system for this study The G940 features high quality throttle joystick and rudders components Each component is sturdy and outfitted with mounting holes to allow them to be bolted to a surface It is felt that mounting holes are very useful in this context because they ensure the safe anchoring of the components to the motion platform Next the force feedback feature of the G940 makes it one of the most realistic consumer level joysticks The absence of force feedback or resistance in the pilot control column is a problem because it allows the pilot to easily bring the controls to their limits which is translated into extreme motions of the simulated aircraft This could be acceptable on a motionless flight simulator but for a full motion flight simulator the extremes flight motions are felt on the motion platform Without force feedback the passenger of the flight simulator could feel that his control inputs bring about disproportionate motion of the flight simulator The joystick and throttle controllers come with many buttons that are fully programmable to map various secondary controls of the simulated 59 aircraft Without a full cockpit this becomes a useful feature to map second
225. sary to collect an unbiased pilot feedback sample 5 2 2 Measures of Performance and Response Variables When conducting an experiment whose aim is the optimization of a well understood system such as the classical WF algorithm it is necessary to establish robust indicators of performance 25 In this case the performance of the classical WF algorithm is measured in two ways First the quality of the motion simulation is evaluated by the author who flies the simulator between each test runs Using this method the pilot compares his notion of expected motion stimuli to the motion stimuli experienced on the system during each flight This is useful to identify problems that may not be easily interpreted using numerical methods Second the quality of motion simulation is evaluated by measuring the fidelity of the simulator motions when compared to that of the simulated aircraft in FSX An improvement of fidelity between each test run is an indicator that the motion response is improving and the opposite is also true In essence pilot evaluation is used to target problematic areas of motion simulation while numerical methods provide the tools to truly identify and solve problems 98 In order to numerically measure performance during the experiment response variables that reflect the performance of the process must be selected The response variables are dependant variables of the process and provide an explicit measurement of the performance of
226. sececeeececsesssseaeeececeeeessseaeeeseceeeeeas 16 2 3 2 Rotational Motion Channel est nn name na hi n nl cua EiS 17 2 3 3 Tilt Coordimation Channel 18 2 3 4 Scaling and imitin sisreine esse EES A Min en A ne 20 2 3 5 FILME pense nn Mennientn nn rte ee e nan sn intra en testo step heat Ne 21 24 Three Degree of Freedom Adaptation ss 23 2 4 1 Translational Motion Channel ss 23 2 4 2 Rotational Motion Channel sense 26 2 4 3 Tilt Coordination Motion Channel ss 28 2 5 Summary 28 MR arm re nom rm Make a seb teehee donee cated ere dr mn ios oy 30 CHAPTER 3 DESIGN nent tnt not stein rfi t an iis Nentuectacintienti nate 31 3 1 Introductio M es EU ARR Rte th bees cheese cede dots oualeet nes Mise ee 31 3 2 D sisn ODJECUVES sir Mn tits vu eee re eet eee ee NL 32 3 2 1 Achieving a Realistic Simulation 33 3 2 2 Cost Effectiveness rs ineens ees east tee eRe tei rentre ade ea dr done pale dues 35 3 2 3 Software Portabiltye 2234 rss artnet seen ces A sent sn cet A he 36 3 2 4 Fase Of Operation fi element nt Et ip ets rt tn Ggugbectetotabotiagoelecsiieuk teense 38 3 2 5 Maintaimability isenimin e aa a e E eS nn dieu hee 39 3 2 6 Safety Are or a mt rte ARE nn E E T a hat Reet E nt ste 40 3 3 Constraints ss eee St aes E RE se me A eee ee 42 3 3 1 Hardware Constraints uinnnninmes ei MAL p a ee ie 43 3 3 2 Software Constants aaea tente ds uaa nn OR tn neat es tro tir a 46 3 4 Flight Simulator Components ss 47 3 4 1 Simulation Environm
227. sed void on_pitchForward_released void on_rollLeft_pressed void on_rollLeft_released void on_rollRight_pressed void on_rollRight_released void on_speedSlider_sliderMoved int void on_stop_button_clicked void on_surgeBackward_pressed void on_surgeBackward_released voi on_surgeForward_pressed on_surgeForward_released void on_swayLeft_pressed void on_swayLeft_released void on_swayRight_pressed void on_swayRight_released void on_yawLeft_pressed void on_yawLeft released void on_yawRight_pressed void on_yawRight_released void qt_ metacall QMetaObject Call int void int qt_metacast const char void MainWindow FSXConnect _ UdpReceive connected hSimConnect mSocket quit mSourceAddr udpDataPorts udpSend Initialize const char PortNumber bool ReceiveDatagram void size_t long bool FSXConnect FSXConnectRequest void FSXDisconnect void MyDispatchProcPDR SIMCONNECT_RECV DWORD void static void isConnected bool simulation void ReceiveDouble double long bool ReceiveDoubleArray double size_t long bool SocketCleanup void UdpReceive UdpReceive Mediator accelerations f
228. ses the pilot is forced to rely on his instruments and not on the motion cues The introduction of motion cues could train pilots to ignore them and focus on their instruments RMC s 3DOF flight simulator is also perfectly suited for the study of advanced algorithms using hardware in the loop configurations Its low cost and simplicity allows for the testing of ideas in short periods of time Such ideas could include the simulation of other vehicles such as tanks and armoured personal carriers The use of software already in use within the CF such as Virtual Battle Space 2 VBS2 33 could provide the virtual environment for military 147 vehicules or aircrafts This would allow for the simple and seamless integration of motion simulators to a networked environment of VBS 2 In light of the low cost of RMC s 3DOF motion platform it might be interesting and also economically feasible to network multiple simulators within this environment In essence with the right virtual environment it would be possible to use motion in many applications Depending on the task finding suitable virtual environment without building them from the ground up can be a challenging consideration when exploring the simulation of new vehicles ANNEXES 148 Annex A Motion Platform Security Briefing 149 150 RMC s 3DOF Motion Platform Security Briefing RMC s 3DOF motion platform is quite powerful and if used improperly could cause injuries to p
229. sical WF Algorithm Implementation The classical WF algorithm can be challenging to implement as it contains a fair amount of mathematical operations such as continuous time integrations derivations and matrix multiplications Reid and Nahon demonstrated that it is feasible to implement the classical WF algorithm using traditional programming languages such a FORTRAN 8 They also highlighted the difficulties associated with the conduct of continuous time integration and derivations with 78 an acceptable level of mathematical precision while keeping the processing times very low 8 Indeed they highlighted the importance of fast and accurate integration methods to maintain an overall low latency in the flight simulation system 8 More than 25 years after the ground breaking work of Reid and Nahon 8 9 and 10 new technologies such as Mathwork s Simulink have emerged and make it easier to deal with concerns of accuracy and processing time when performing mathematical operations Simulink is an environment for multi domain simulation and Model Based Design for dynamic and embedded systems It provides an interactive graphical environment and a customizable set of block libraries that let you design simulate implement and test a variety of time varying systems including communications controls signal processing video processing and image processing 22 The use of Simulink allows focusing on the actual time varying system instea
230. sign Sub Objectives 00 eee eeseseeeeeneecsseceeeeseeesaeecaecsseeseeeeaees 40 Figure 21 Safety Design SUb OD CCH Ves esse it Gi 2G hate ne er aaa te te Maan Cg ete etre 41 Fieur 22e D sren Constant nis Se ed en en TN rm nd 43 Figure 23 Design Hardware Constramts 2 taime 43 Figure 24 Serious Damage Sustained by RMC s 3DOF 45 Figure 25 Hardware Collision Caused by a Design Flaw 00 cceessceeesceceeeeceeeeeceeeeeeeteeeeeeeeees 45 Figure 26 Design Software Constraints yee RGA BE URIS 47 Figure 27 Interaction of Components in the Final Flight Simulator Design 48 Fig re 28 Z600 3D Vis aca ede eal lee oak eda Re RS EE cates Sage tin 52 Figure 29 TrackIR 5 with TrackIR PRO CDs te Me eee talent 53 Fig t 30 Fatality Sound Headset te sine nt ne nm a a a Raa 54 Figure 31 Corbeau A4 Racing Seat and Dimensions ssessseeesesrsesssresseseresressesererrerseseresreeseese 55 Figure 32 Corbeau Four Points Harness Bell net 56 Figure 33 Logitech s G940 Flight System sigs sn sn ranasn menant denied 58 Figure 34 Concept Drawing of the Pilot Seat and Controls 00 eee eeeeeeeeeeeereeessecneeeeeeeeenees 59 Figure 35 Close up on Control Adjustment Mechanisms Concept 60 Figure 36 SquareD Heavy Duty 60A Safety Switch 62 Figure 37 Pinal GUI DESERT ne mn me ten me nn nn testes tah 63 Figure 38 General Interaction of Software Components and Flow of Information 0 67 Figure 39
231. simulator system an important experimental phase began to fine tune the classical WF algorithm This would help to study the performance of the flight simulator with the intent of improving the fidelity of motion simulation The purpose of this chapter is to describe the experiment that was conducted to study and improve the performance of RMC s 3DOF motion platform used as a flight simulator The experiment provided the author with a tangible way to assess the fidelity and limitations of the motion sensation generated by RMC s 3DOF motion platform when using different parameters for the classical WF algorithm After having completed the experiment and fine tuning of the motion platform with the intent of achieving its full motion simulation potential a validation exercise was conducted with private commercial and military pilots This exercise provided insight into the pilot perceived fidelity of this particular flight simulation as well as feedback on the unique design of RMC s 3DOF flight simulator 5 2 Experimental Study of the System The flight simulator described in Chapters 3 and 4 was run through several select flight profiles that reproduce the most prevalent manoeuvres during flight This provided insight into the fidelity of motion the simulator generated when using different parameters for the filtering and scaling blocks in the classical WF algorithm The data collected during this experimentation 96 phase allowed for the fine
232. ssage W else if else if uInput park DOF CODE HERE feedback Parking the platform Feedback message else if else if uInput stopMovement 6DOF CODE HERE feedback Movement Stopped Feedback message else if else if uInput pitchForward DOF CODE HERE feedback Pitching forward Feedback message else if else if uInput pitchBackward DOF CODE HERE feedback Pitching backward Feedback message else if else if uInput rollLeft DOF CODE HERE feedback Rolling left Feedback message else if else if uInput rollRight 6DOF CODE HERE 215 feedback Rolling right Feedback message else if else if uInput heaveUp DOF CODE HERE feedback Heaving up Feedback message W else if else if uInput heaveDown DOF CODE HERE feedback Heaving down Feedback message W else if else if uInput yawLeft DOF CODE HERE feedback Yawing left Feedback message else if else if uInput yawRight DOF CODE HERE feedback Yawing right Feedback message l else if else if uInput swayLeft 6DOF CODE HERE feedback Swaying left Feedback message W else if else if uInput swayRight DOF CODE HERE feedback Swaying right Feedback message W else if else if uInput surgeForward DOF CODE HERE feedback
233. still required for the tilt coordination channel Lis faa a ay a Die g Zy g Tilt Coordination Figure 7 3DOF Translational Motion Channel As can be seen in Figure 7 the unnecessary accelerations are discarded in the high pass filtering step Box three but all the accelerations make their way to the tilt coordination channel Since the x and y translational acceleration components are not required only the z component is high pass filtered using a second order filter In addition it can be seen that the vector Jyyz is added to Block 1 This step is necessary for this study since the aircraft accelerations that are obtained from Microsoft Flight Simulator X FSX the virtual environment used in this study do not include the force of earth gravity The algorithm is designed to receive aircraft accelerations that include the effect of gravity Additional information on this topic and the solution to this problem is available in Section 4 2 4 Filtering of the low frequency accelerations only occurs after the accelerations are brought to the earth fixed reference frame F using the Euler transformation matrix Lys High pass filtering can be done in both the Fs or F reference frames but each location has its advantages and disadvantages When filtering in either Fs or Fr false 25 offsets or incorrect motion cues will be generated However according to the study conducted by 8 filtering in F71s preferable as the adverse ef
234. system as shown in Figure 27 At the heart of the flight simulation system a control station is the interface between the software and hardware components A detailed wiring diagram for the hardware components is available in Annex C 49 In the sections to follow each of the key components used in the system as well as their intended role are discussed In addition the reasons for the selection of those components are explained in each section 3 4 1 Simulation Environment As stated in the design objectives the selection of a simulation environment is an important decision in the design of a flight simulator Three commercially available simulation environments were evaluated for RMC s 3DOF flight simulator Microsoft Flight Simulator X FSX was selected as the most suitable for the design of RMC s 3DOF motion platform As can be seen in Table 3 FSX surpasses the other flight simulation products on many of the characteristics thought relevant to this application Table 3 Cost Benefit Analysis of Flight Simulation Software Microsoft Flight Simulator Flight Simulator X Plane 9 Flight Gear v2 0 Programming Language C C C Cost Around 50 Around 50 Free Physics Quality Accurate Accurate Approximate Graphics Quality Very Good Good Average Development Tools Yes Yes No Sample Rate Variable 1 to 100 Hz 25Hz Variable Access to Relevant Flight Data Very Good Poor Poor User Documentation Very
235. t and rate the potential for using RMC s 3DOF as a training aid for pilots under training The pilot rating questionnaire that was used to collect the pilot ratings can be found in Annex H Volunteers were directed to observe silence during the tests when other pilots were present This measure was taken to prevent any discussion of the simulator s performance which could lead to a bias in the group responses The focus was put on collecting individual and unique evaluation from each pilot 136 5 4 2 Pilot Ratings and Comments A total of 12 pilots with varying degrees of experience conducted the pilot evaluation exercise The following section will discuss the pilot ratings and comments obtained during the validation exercise The pilot validation exercise benefited from a wide range of pilots with different qualifications Pilot flight time ranged from 85 hours to more than 10 000 hours Qualifications ranged from private pilot to fully qualified multi engine commercial instructors It was also possible to obtain the input of a military test pilot who had previous experience in simulator evaluation Relevant information such as pilot flight time qualifications and aircraft flown can be found in Annex I The ratings collected through the pilot rating questionnaire were separated into four categories relating to hardware motion flight simulation impact and training potential Table 27 shows the average ratings collected for the
236. te buttons on the GUI a void MainWindow checkConnections while 1 When there is no motion simulation if m gt isReceivingUDP amp amp m gt isFSXConnected If both FSX and Simulink connection are open ui gt SimulinkConnectLabel gt setText Simulink Connection Established ui gt FSXConnectLabel gt setText FSX Connection Established if motionSimulation ui gt stop_button gt setEnabled true Enabled the Stop Motion Simulation button ui gt begin_button gt setDisabled true Disable the Begin Motion Simulation button ui gt FSXConnect gt setDisabled true Disable the Start FSX connection button ui gt FSXDisconnect gt setDisabled true Disable the Stop FSX conenction button Wif else if motionSimulation ui gt begin_button gt setEnabled true Enable the Begin Motion Simulation button ui gt stop_button gt setDisabled true Disable the Stop Motion Simulation button ui gt FSXConnect gt setDisabled true Disable the Start FSX connection button ui gt FSXDisconnect gt setEnabled true Enable the Stop FSX conenction button else if break HI if else if m gt isReceivingUDP amp amp m gt isFSXConnected ui gt begin_button gt setDisabled true Disable the Begin Motion Simulation button ui gt stop_button gt setDisabled true Disable the Stop Motion Simulation button ui gt FSXConnect gt setEnabled true Enable the Start FSX connection
237. tered signal Indeed it was observed that most undesirable overshoots found in the filtered signal exhibits an inverse polarity to that of the unfiltered signal As such by comparing the polarity of the unfiltered signal with the filtered signal and by switching the filtered signal to zero when polarities are different most undesirable overshoot are neutralized A test run featuring this overshoot management technique can be seen in Figure 68 As can be seen this technique is very effective and reliable in most cases This can be seen when comparing Figure 66 with Figure 68 where a definitive improvement is noted The only shortfall of this technique is that sometimes it will fail when the unfiltered signal wanders and acquires the same polarity of an overshoot Fortunately it is seldom the case where both the overshoot and the unfiltered signal have the same polarity Equation 5 5 describes the conditional statement required to implement this switch on the filtered signal Essentially the signal is automatically zeroed every time the polarity of unfiltered and filtered is different Filtered es If SGN Filtered SGN Unfiltered 5 5 Else After having studied the phenomenon of overshooting on 2 order high pass filter a technique using the polarity of unfiltered signals is used to remove most overshooting from filtered signals This technique is implemented in Matlab and will be tested in the second test run 123 5 3 4 Test Run
238. the angular rate response variables won t carry much information on the performance of the channel In addition the yaw motion channel response variable is never relevant for this study since RMC s 3DOF is not capable of generating yaw motions Thus the response variables of interest are selected depending on the channel under study Matlab code was developed to automatically calculate the response variables from a test run This response variable calculation code can be found in Annex E Finally the visual inspection of motions against time plot between experimental and recorded flights can provide additional insight into the quality of motions simulation The motions of the flight simulator are captured using an inertial measurement unit The Crossbow IMU 320 dynamic measurement system is specifically used to measure the translational accelerations and angular rates of the motion platform As can be seen in Figure 55 it is installed at the simulator reference frame Fs which is located in the geometrical center of the motion platform below the pilot seat As discussed in Chapter 2 reference frame Fs must be in the same location with respect to the pilot in both the simulated aircraft and the motion platform As such the measurements provided by the IMU 320 fe and Wss are taken at the same location as hia and w44 the inputs to the classical WF algorithm In order to conduct the experiment with ease and precision the IMU 320 was interfaced
239. the simulation is paused in order to set all outbound FSXConnect cpp data to zero and ensure that the motion platform stops moving It is important to highlight the fact that data is only received from FSX when it has been modified in the flight simulation In other words if there is constant x axis acceleration the acceleration value will be published only once The next publish by the SimConnect server will only occur when the acceleration value has changed in the simulation The differences between the axis systems used by the classical WF algorithm and FSX can be seen in Figure 43 It is 77 important to note that the axis orientation and axis rotation is different between the two environments Classical Washout Algorithm Flight Simulator X _ VA Ses x Z i Axis Rotation X Axis Rotation Right Hand Rule Left Hand Rule Z i w Figure 43 Different Axis Systems between Classical WF Algorithm and FSX Thus the FSX axis system is converted from FSX to classical WF algorithm axis system inside FSXConnect cpp No officially published data could be found on the FSX axis system As a result it was necessary to extrapolate the axis system by conducting an experiment where specific manoeuvres known to generate motion along or about a specific axis were performed to uncover the orientation and axis rotation of each axis More information on the specific C implementation of FSXConnect cpp can be found in Annex K 4 2 5 Clas
240. the simulator s apparatus needs to be positioned at this location and remain below the maximal payload weight In theory if the motion platform respects all the aforementioned parameters and remains within its motion envelope Table 1 it is safe to operate However as can be seen in Figure 24 during the conduct of this study serious damage was caused to the motion platform during a routine session After investigating the cause of this incident it was concluded that SSSI s specifications were incorrect and that damage could occur even when following the prescribed limits Indeed when following a specific series of manoeuvres the upper platform frame can be brought in a position where it collides with one of the motion platform drive motors An example 45 of this kind of collision can be seen in Figure 25 where the motion platform is in a full pitch backward combined with a roll left Figure 25 Hardware Collision Caused by a Design Flaw The possibility of a collision between the upper platform frame and two of its drive motors creates concerns for the safe operation of RMC s 3DOF However such collisions can be 46 avoided with the appropriate modifications Through testing it was found that it is possible to prevent a rear axis drive motor collision by limiting the rear actuators three centimetres higher than originally specified or by cutting the rear corners of the platform This can be accomplished by implementin
241. the size of the double array for accelerations Mif else if platformType 6DOF number_of_actuator 6 lengths new double number_of_actuator Setting the size of the double array for lengths velocities new double number_of_actuator Setting the size of the double array for velocities accelerations new double number_of_actuator Setting the size of the double array for accelerations else if return platformType connect pe The execute method is called to perform an action This method simply passes the request to the model 208 which will evaluate it and send the proper motion commands to the motion platform k param ulnput a string containing the command request return a string containing feedback on the success or failure of the request string Mediator execute string uInput string status if uInput FSXConnect If the GUI is asking for an FSX connection f gt FSXConnectRequest Pass the FSX connect request to FSXConnect Mif else if ulnput FSXDisconnect If the GUI is asking for an FSX disconnection f gt FSXDisconnect Pass the FSX disconnect to FSXConnect else if else For any other GUI requests status m gt process_user_input uInput Pass the command request to the model else return status process_user_input pe The change_speed method is called to change the motion platform speed This method simply passes the informa
242. the system s frequency is 50 Hz The signals initially collected from three identical runs were smoothed to evaluate the effectiveness of this technique to improve performance As can be seen in Table 15 the smoothing of Crossbow IMU signals has a positive effect on the overall precision of the system Even if the angular rate signals were the most affected by noise smoothing was applied to all Crossbow IMU signals As a result a gain on precision was also made on the translational motion outputs of the Crossbow IMU 112 Crossbow Output p deg s wo Raw Crossbow Signal 1 L 1 Extremity Error Smoothed Crossbow Signal 1 4 1 1 1 1 1 1 1 3 4 5 6 1 7 8 9 10 0 1 2 3 4 5 6 8 9 10 Figure 61 Smoothing of a Typical Crossbow Angular Rate Signal This gain on precision can be easily visualized in Figure 62 where margins of errors of raw and smoothed signals are compared Indeed using this smoothing technique precision is at least doubled on each channel Table 15 Margin of Error for Three Identical Runs Smoothed Signals Motion peat Standard Margin of Error 95 Confidence ean Error Channel Deviation Interval Translational 3 F j 0 0096 mes 0 0144 mes 0 0384 mes x Translational P 5 0 0043 mes 0 0073 mes 0 0188 mes y Translational 5 0 0041 mes 0 0097 mes 0 0236 mes Z Angularp 0 0754 deg
243. the user if the controller IP Address would be detected automatically This problem is caused when two types of commands absolute and jog are executed simultaneously Protective code is added to prevent the user from doing this The program crashes when commanding any motion during a parking extend of center manoeuvre A logic error in a Try Catch routine The program connects to RMC f used to detect to which motion 6DOF when there is no controller i platform the software is connected connected was corrected A connection monitoring algorithm is implemented to ensure that motion simulation can only occur when all required connections are established The program does not detect a lost of connection from either Simulink or FSX and alive There is not h explanation f ee S EEA E ee a The GUI was updated with further the Erase Parked Position R ca explanation to pr functionality A user could damage r f r inadvertent incorrect use of this the platform not knowing about this advanced feature functionality If the user forgets to park the platform after a session manual reconfiguration is required Most users won t know how to reconfigure the platform for safe usage If the user forgets to park the motion platform it will automatically park itself when closing the software 4 4 Hardware Implementation Following the acquisition of all the required components select
244. this button is pressed actuator 4 will move up The request is passed to the mediator in this method ii void MainWindow on_m4up_pressed Feedback m gt execute actuator4Up Send request to mediator on_m4up_pressed J Event handler for the Motor 4 Up button When the button is released the actuator stops The request is passed to the mediator in this method void MainWindow on_m4up_released Feedback m gt execute stopMovement Send request to mediator on_m4up_released px Event handler for the Motor 4 Down button As long as this button is pressed actuator 4 will move down The request is passed to the mediator in this method void Main Window on_m4down_pressed Feedback m gt execute actuator4Down Send request to mediator on_m4down_pressed px Event handler for the Motor 4 Up button When the button is released the actuator stops The request is passed to the mediator in this method af 205 void MainWindow on_m4down_released Feedback m gt execute stopMovement Send request to mediator on_m4down_ released ee Event handler for the Motor 5 Up button As long as this button is pressed actuator 5 will move up The request is passed to the mediator in this method i void MainWindow on_mSup_pressed Feedback m gt execute actuatorsSUp Send request to mediator on_mS
245. tilt angle signals of the channel This effect can be attributed to the taking of the integral when the derivative is equal to zero In order to solve this problem while keeping the essential limiting block of the algorithm intact a gain coefficient was introduced after the tilt limiting block 124 Tilt Rate deg s 0 5 10 15 20 25 30 Time s Figure 69 The effect of taking the Derivative in the Tilt Limiting Block This gain is unique to this implementation of the classical WF algorithm and ensures that the tilt coordination channel exhibits a signal sufficiently strong to move the motion platform as desired Table 19 shows the coefficients used for the second test run where modifications are highlighted in red Most scaling coefficients were adjusted and the filter damping ratio of the heave channel was increased as recommended by 12 to reduce overshoot Although an overshoot management segment was implemented the adjustment of the filter Equation was beneficial for the case where the overshoot management segment fails To begin the test run the author sat in the motion simulator and flew several manoeuvres to evaluate the simulator response under various conditions In essence the response was felt to be much more realistic for all motion channels A summary of the pilot observations collected for all motion channels during this test run can be found in Table 20 125 Table 19 Classical WF algorithm Configuration Test
246. tion Annex B Flight Simulator Components and Cost 151 Approx Cost Component Manufacturer CDN Role 2 Inch Harness Belts Corbeau 85 00 Pilot seat safety harness A4 Racing Seat Corbeau 350 00 Pilot seat Control Shaft and Woodman 1 216 35 Custom harness for TrackIR 5 Mount Machine Products co pilot controls sale une Creative 55 00 Audio Headset Flight Simulator X Microsoft 40 00 ae ven Soalan environment Pilot input system Flight System G940 Logitech 300 00 controls Heavy Duty 60A Motion platform safety Safety Switch SANE 7309 switch EMCS DOF Monon SSSI 55 000 00 Motion cues generation Platform Trac HR vand NaturalPoint 175 00 6DOF Head tracking Trackclip Pro USB Extension 32 ft Sabrent 18 00 Extension for TrackIR 5 USB Extension 32 ft Sabrent 18 00 Extension for controls USB Extension 32 ft Sabrent 18 00 PDA Tor A300 3DVisor VGA Extension 25 ft Ultra 36 00 Eon Or AOR 3DVisor VGA Extension 25 ft Ultra 36 00 FRANS N 3DVisor Z800 3DVisor eMagin 1 800 00 Visual display Total 59 222 35 Annex C Flight Simulator Wiring Diagram 153 154 Figure 73 is a simple diagram showing the specific layout of cables used to bring the system together It provides additional information for anyone trying to understand the implementation of the system using the components selected during the design Several universal serial bus USB video graphics
247. tion to the model which will interpret it properly and change the motion platform speed ek param s the new speed of the motion platform from 1 to 100 void Mediator change_speed int s m gt change_speed s Pass the speed change request to the model change_speed pe The isReceivingUDP method checks if the mediator js receiving any UDP packets from MATLAB ek return amp lt true amp gt if UDP packets are received amp lt false amp gt if no UDP packet are received bool Mediator isReceivingUDP0 if udpVelocities ReceiveDoubleArray velocities number_of_actuator 10 true ll udpAccelerations ReceiveDoubleArray accelerations number_of_actuator 10 true ll udpLengths ReceiveDoubleArray lengths number_of_actuator 10 true return true If there are any packets received return true Mif else return false No packets were received return false else isReceivingUDP pe The isReceivingUDP method checks if FSXConnect is connected to FSX through SimConnect ek return amp lt true amp gt if there is an active FSX connection amp lt false amp gt if there is no connection bool Mediator isFS XConnected return f gt isConnected isFSXConnected J2 The simulation method receives the motion simulation information by UDP from Simulink and forwards them to the model for processing void Mediator simulation f gt simulation Pass the simulati
248. tion was created This documentation allows anyone to quickly find the role of classes methods and fields variables in the code Most notably the documentation explains which parameters are necessary for each method and explains what is returned Although it is possible to create documentation in many formats the Javadoc format from Sun Microsystems was selected for this study because of its simplicity and ease of use Indeed the Javadoc format is built upon the HTML standard and is widely used in the industry even to document code in languages other than Java In addition to documenting the code with a Javadoc each line of the C implementation was commented to ensure that a complete stranger would be able to understand any and all lines of code that were produced include Model h Used to establish a connection with either RMC s 3DOF or 6 DOF platform Some parameters are also initialized depending on which platform the connection is established with param ip the ip adress of the controller to which a connection is attempted return a string containing either 3DOF or 6DOF y string Model connect string ip ipAddress ip platformType 3DOF Default platform type 3DOF try Attempts to connect to the 3DOF g new Galil ipAddress Creates a Galil Object g gt timeout ms 1000 Allow for a 1 second timeout g gt connection Connect to the controller try Figure 51
249. tive process where one tries to find errors inside his software As a result it can sometimes be difficult to be truly critical of one s own work and find all the errors 24 For this reason submitting the software to others for testing can be an effective way of uncovering the most hidden and insidious bugs of software 24 In this validation exercise three volunteers in the field of Computer Science were given the task of testing all three intended use cases of the software with a mandate to attempt to crash or find bugs in the process The goal was to unearth potential problems and to collect commentaries on the look and feel of the interface The results of this validation exercise are summarized in Table 9 This short and simple exercise proved to be most valuable for this study as it unearthed important problems in the software Despite the author s best effort at debugging some problems were still unearthed by third party reviewers This shows the effectiveness of user validation when developing software Finally this exercise also provided valuable information on the expectations of the software user a notion that is not always obvious to the software developer 90 Table 9 Observations and Problems Discovered during User Validation Ob ti d Probl serva ee and Froblems Action Taken Discovered The standard controller IP Address is automatically typed in the appropriate field for the user It would be easier for
250. tly different between each run of an identical flight profile With that in mind a small experiment was conducted to measure the loss of precision that can be expected for each motion channel Using Simulink a flight profile was recorded and played back three times through the classical WF algorithm For each run the motions generated by RMC s 3DOF motion platform were recorded using the Crossbow IMU An average motion dataset was then generated from the three recorded sets Next the absolute error between each of the recorded sets and the average set was calculated These absolute error dataset for each run were averaged to create an absolute mean error dataset Finally the mean error for each motion channels of this mean error dataset was calculated along with the standard deviation As can be seen in Table 14 a margin of error with a confidence interval of 95 was calculated from this data Table 14 Margin of Error for Three Identical Runs Unsmoothed Signals Motion Standard Margin of Error 95 Confidence Mean Error Channel Deviation Interval Translational 5 A 0 0227 mes 0 0315 mes 0 0857 mes x Translational 2 0 0142 mes 0 0148 mes 0 0438 mes y Translational 5 4 gt 0 0223 mes 0 0256 mes 0 0735 mes Z Angular p 0 2218 deges 0 1367 deges 0 4952 deges Angular g 0 2430 deges 0 1598 deges 0 5626 deges Angular r 0 1682 deges 0 10
251. to aid development Lastly to facilitate portability of the software components a simple architecture must be designed The use of a single programming language for all the software component of the flight simulator promotes simplicity Indeed if a system were to use several different programming languages it reduces its portability to other platforms and creates interface problems between those languages 3 2 4 Ease of Operation Ease of Operation represents the ease with which users can operate the flight simulation system to accomplish their goals This includes all phases of operation involved in a typical flight simulation run Users interact with the flight simulation system by means of a software interface For this reason great care must be taken to design a GUI that promotes ease of operation As can be seen in Figure 19 several factors must be taken into consideration to design a GUI that allows the flight simulator to be easily operated When properly implemented a GUI guides the actions of a user and reduces errors that could crash the system The GUI must be designed in such a way that it does not allow the user to perform a series of actions that could lead to a crash Additionally the GUI must be forgiving if a user commits an error when controlling the flight simulator 39 Ease of Operation Graphical User Interface Awareness Feedback Forgiving Protection Abstraction Simplicity Figure 19 Ease of
252. to flight simulation Flight Training Potential If you have experience with flight instruction 171 Do you see a potential use for this kind of flight simulator as a flight instruction aid Please elaborate Please include a rating from I to 5 on the training potential of RMC s 3DOF flight simulator Attribute Rating Response RMC s 3DOF flight RMC s 3DOF flight Potential simulator has great 4 simulator has no potential for flight potential for flight training training Additional Comments Feel free to enter any comments with regards to the flight simulation experience Annex I Pilot Information and Ratings 172 Table 32 Pilot Information 173 Total C 172 Pilot Flight Flight Experience Aircrafts Flown Time hrs Time hrs Commercial Instructor Various commercial aircrafts up to l 19009 000 Multi Engine IFR Boeing 737 Military helicopter pilot 2 2600 None Standards officer Test CH146 CH136 CH118 pilot Commercial Instructor 3 800 100 Multi Engine IFR C 150 C 172 Zlin 242 Commercial Multi C 172 PA 28 140 Piper Seminole TAR an Engine IFR Beech Duchess 5 103 9 50 RONA Pri mary cred C 150 C 172 Grob 120A training Air Force 6 110 50 Private C 152 C 172 7 100 70 Glider Private SGS233A C 172 8 400 75 Commercial Instructor C 150 C 172 9 280 150 Commercial Instructor C 150 C 152 C 172 10 225 90 Glid
253. tor 1 feedback Actuator 1 moving down Feedback message else if else if uInput actuator2Up g gt command JG speed Actuator 2 up g gt command BGY Begin motion on actuator 2 feedback Actuator 2 moving up Feedback message W else if else if uInput actuator2Down g gt command JG speed Actuator 2 down g gt command BGY Begin motion on actuator 2 feedback Actuator 2 moving down Feedback message l else if else if uInput actuator3Up g gt command JG speed Actuator 3 up g gt command BGZ Begin motion on actuator 3 feedback Actuator 3 moving up Feedback message l else if else if uInput actuator3Down g gt command JG speed Actuator 3 down g gt command BGZ Begin motion on actuator 3 feedback Actuator 3 moving down Feedback message W else if W if pe 4 6DOF MOTION CODE TO BE INSERTED BELOW Specific 6DOF commands to be inserted for each motion type ek rs else if platformType 6DOF If we are connected to the DOF if uInput erasePosition DOF CODE HERE feedback Parked position erased Feedback message Wif else if uInput markPositionParked 6DOF CODE HERE feedback Position marked as parked Feedback message else if else if uInput center DOF CODE HERE feedback Moving platform to central position Feedback me
254. trol experience listic feeling Excellent creates a U ireaisue termi Audio Pu l 5 detrimental to the very realistic feeling experience Ergonomic very Uncomfortable kpn b detrimental to the on comfortable 5 e e ta o experience Overall Excellent 5 Extremely poor Motion Simulation Rating Please indicate your assessment of motion simulation based on the following rating scales 170 Attribute Rating Response Extremely smooth seu Smoothness comparable with fixed HAS ere vlimi of tolerance base Sense Denne LL Totally reversed aircraft Amplitude No motion experienced Teast twice Mat expected Phase Lag None experienced At least 180 Cannot continue Discomfort None experienced manoeuvre Disorientation None experienced Connor pertoni Manoeuvre Overall Excellent Extremely poor Motion Simulation Impact Please indicate your assessment of motion simulation impact based on the following rating scale Attribute Rating Response Impact Motion simulation has a very positive impact on the flight simulation Motion simulation has no impact or is even detrimental
255. tudent pilot to fly on his own Early aircraft operators acknowledged this reality and quickly foresaw a need to replace and supplement some of the real world flying exercises with ground based training exercises Ground based flight training systems were first introduced during the First World War in an attempt to train a larger number of pilots 1 Most early aircraft simulation systems gave mixed results as it was unclear whether they had a positive effect on pilot training 1 Successful aircraft simulation systems really made their appearance when computing technology allowed for basic flight equations to be solved in real time 1 Since then the flight simulation industry has grown significantly and there has been sustained development in ground based flight simulators Nowadays a myriad of aircraft simulation systems are in use by commercial and military organizations for the training of pilots 1 1 2 2 Advantages of Flight Simulation Flight simulator operators benefit from a reduction in pilot training time and a massive reduction in training costs 1 For the training of emergency procedures flight simulators allow an unparalleled level of convenience and safety Additionally flight simulators are ideally suited for the training of instrument flight rules IFR flights and the conduct of re currency checks 1 2 Furthermore typical flight simulator environment allow for the recording of exercises and provide the users with enh
256. turn to its central position in order to accommodate subsequent motion demands The work completed in this research is intended to allow others to understand the underlying principles that govern RMC s 3DOF flight simulator With that in mind this thesis is crafted in anticipation that it will be used as a reference for other motion simulation projects to come This research is accomplished using a 3 step process The first step is to review the literature and to adapt the 6DOF classical WF algorithm for RMC s 3DOF motion platform Secondly a flight simulator is designed and implemented to take advantage of the motion platform capabilities Finally an experiment is conducted in order to optimize the simulator s fidelity and to evaluate the quality of sensations it creates 1 6 Contributions This thesis introduces an architecture that provides a way to enable flight simulation using a low cost motion platform with commercially available components The approach utilized showed that an acceptable level of realism can be reached for flight simulation without the need for specifically engineered components The flight simulator developed in this thesis shows the effectiveness of this approach in displaying the feasibility qualities and shortfalls of such a commercially available component based system More specifically robust and flexible software designed to allow for the use of different motion platforms and virtual flight environ
257. ty seatbelt compatibility and the possibility of adjustments for the passenger It is felt that racing type car seats are the most suitable for this application as they provide harness compatibility and protruding sides to help keep the passenger inside the seat Most racing seats are also ruggedized with wear patches and are designed for tight spaces Almost all commercially available racing seats boast similar features between brands For this reason a good quality seat was selected based on availability and pricing Weighing only 26 pounds the Corbeau A4 racing seat was selected for the design or RMC s 3DOF flight simulator It features solid anchor points harness installation holes and a reclining backseat The seat and its associated dimensions can be seen in Figure 31 CORBEAU Figure 31 Corbeau A4 Racing Seat and Dimensions The Corbeau four point harness belt used to secure the passenger to the motion platform can be seen in Figure 32 This harness is bolted directly to the motion platform upper frame to secure the passenger on the seat As highlighted in the design objectives pilot controls must be 56 selected in order to provide realistic control of the simulated aircraft to the passenger The realism and quality of pilot controls increases the overall realism of the flight simulation One of the most important controls for an aircraft pilot is the control column that commands motions of pitch and roll by moving the aircra
258. udio video tapes DT Data will be retained indefinitely in a secure location C Other 4 If there are conditions under which privacy of participants or confidentiality of data cannot be guaranteed provide details n a G DECEPTION 1 If deception is to be used in this study provide a justification n a 2 If deception is to be used outline the process to be used to disclose the full nature of the research to the participants Attach a copy of the information sheet describing the true nature of the research n a Is consent sought again following disclosure of the full nature of the research to the participants for the use of their data Yes No APPLICANT AGREEMENT I have read the Guiding Ethical Principles in the Tri Council Policy Statement and agree to comply with the conditions outlined in the Guidelines In the case of student research as Faculty Supervisor my signature indicates that I have read and approved the application and proposal and I deem the project to be of high scientific ethical and scholarly standards Signature of Applicant Date Signature of Faculty Supervisor if applicable Date 186 INFORMED CONSENT LETTER Good day The purpose of this study is to evaluate the fidelity of RMC s 3 DOF flight simulator RMC s 3 DOF flight simulator is a low cost flight simulator when compared to other commercially available flight simulator The study invol
259. un Center the motion platform if required Ensure that the appropriate flight profile is set to be 3 read in the Simulink Playback model Set the simulation time accordingly 30 seconds Ensure that the Inertial measurement unit recording 4 block is powered on and properly configured to save the data under the appropriate filename Begin the motion simulation using the Playback Simulink model Once the simulation is complete save the test run results 7 Calculate the response variables and analyze the data This procedure represents the actions taken during one test run for each channel and the overall system In accordance with the sequential approach of this experiment the data and response variables for each test run are analyzed in order to determine the scaling and filtering coefficients of interest for the next run Unlimited flight testing by the author is also permitted to 108 collect a subjective measure of performance in order to better target problematic areas of motion cueing Observations are collected in writing and kept as experimental data 5 3 Conduct of the Experiment After having carefully designed the experimentation used to optimize the classical WF algorithm experimental test runs were conducted Prior to the experimental test runs due diligence was exercised by thoroughly testing the precision of the system This precision was measured through a series of control flights In this section test run
260. up_pressed ee Event handler for the Motor 5 Up button When the button is released the actuator stops The request is passed to the mediator in this method ie void MainWindow on_mS5up_released Feedback m gt execute stopMovement Send request to mediator on_mSup_released Event handler for the Motor 5 Down button As long as this button is pressed actuator 5 will move down The request is passed to the mediator in this method void Main Window on_m5down_pressed Feedback m gt execute actuatorsDown Send request to mediator on_mSdown_pressed ee Event handler for the Motor 5 Down button When the button is released the actuator stops The request is passed to the mediator in this method af void MainWindow on_mSdown_released Feedback m gt execute stopMovement Send request to mediator on_m5down_released px Event handler for the Motor 6 Up button As long as this button is pressed actuator 6 will move up The request is passed to the mediator in this method el void MainWindow on_m6up_pressed Feedback m gt execute actuator6Up Send request to mediator on_m6up_pressed ee Event handler for the Motor 6 Up button When the button is released the actuator stops The request is passed to the mediator in this method void MainWindow on_m6up_released Feedback m gt
261. ussion will be held to collect additional comments and thank the participants for their time The only record kept for a session is the rating questionnaire The conduct of a session is summarized in the following table Step Description 1 Participant arrival Introduction to the flight simulator and introduction to the pilot validation procedure 2 Briefing Explanation of the rating questionnaire as well the manoeuvres used for rating 3 Initial run The participant performs a circuit without motion to get used to the controls the airplane and the manoeuvres 4 Validation run The participant performs the same circuit with full motion to evaluate the flight simulator 5 Rating The participant fills the questionnaire to rate the flight simulator 6 Debriefing The participant is thanked and additional feedback is collected trough a short discussion 181 3 Research Setting If procedures involve direct contact with participants describe the setting in which the research will take place Indicate who will be present during the process Describe any known or foreseeable risks for the researcher or others involved in collecting the research data What safeguards are in place to protect against these risks In person contact with participants will be made during the pilot rating exercise Staff will welcome the participants and guide them throughout the exercise Staff will also be responsible for oper
262. ves flying the simulator through two circuits with and without motion Then you will be asked to fill a short questionnaire to rate the flight simulator It will take approximately 30 minutes of your time to complete these activities 2Lt Jean Le Bouthillier of the Department of Mathematics and Computer Science is conducting the study Participation is voluntary You are free to participate to the extent that you feel comfortable and can decide to stop participating in the study at any time by alerting the research staff Strict guidelines will be followed to protect your privacy The ratings you provide about the flight simulator will be anonymous Information on your flight experience will be collected to study the impact of flight experience on the results It will not be possible to link the results with your participation in this study and only group data will be reported 2Lt Jean Le Bouthillier will informally keep track of who has completed the study to ensure participants do not participate more than once This research project has received ethical approval by the Royal Military College Research Ethics Board If you have concerns about the ethical nature of this study please contact the Chair of this Committee Dr Daniel Lagace Roy Daniel lagace roy rmc ca 613 541 6000 ext 6572 Questions regarding this study should be addressed to 2Lt Jean Le Bouthillier jean le bouthillier rmc ca tel 613 583 8778 The results of this study wh
263. ware constraints not all options are suitable for the design of RMC s 3DOF flight simulator The main constraint limiting the selection of a visual display is the size of the motion platform and the requirement for maintaining any load in the vicinity of its center of gravity Table 4 Cost Benefit Analysis of the Three Major Types of Display Head Mounted Projection Visual Display Display Flat Screen s System Cost 500 50000 200 500 Each 500 5000 Weight on Platform Very Light Heavy Heavy Must be Secured to Platform No Yes Yes Immersive Without Closed Yes No No Cabin IR Head Tracker Compatible Fully Partially Partially Geometrically Centered Yes No No Requires Special Requirements No Mounting Requires Screen Z800 Head Mounted Display Available in the Department 52 For this reason a convenient option is to utilize a head mounted display combined with a head motion tracker to achieve an immersive effect by providing unlimited field of regard A comparison between the three major types of visual display system can be seen in Table 4 The use of flat screens or a projection system is incompatible with the design objectives and the constraints imposed by RMC s 3DOF motion platform Indeed those types of visual displays are most effective when the user is enclosed in a cockpit or a cave without outside references Due to the motion platform size and limitations it is not practicable to
264. wn to be within the physical limits of the motion platform being used and limited in function of the motion platform physical limits additional information can be found in Section 2 3 4 The scaled down and limited accelerations are represented by the intermediate variable f that is used later in the tilt coordination channel In Box two the accelerations are transformed from reference frame F4 to reference frame Fr using the Euler rotation matrix Lys given by Equation 2 1 8 The intermediate earth referenced accelerations derived from Box two are represented by bs 17 cos cosy Sin sin cosp cos sin8 cosy cos sin sing siny Lis sing sin sinp coso sin siny 2 1 cast Su coso cosp sing cosy sin0 sing cos0 cos cos The local gravity vector g is then added to the accelerations producing intermediate specific accelerations represented by az In this formulation 5 represents the scaled down and limited aircraft specific accelerations in the earth reference frame F Next a is obtained with the addition of the local gravity vector to fe As can be seen in Box twelve the Euler rotation matrix Lys is formed using the current simulator orientation given by Euler angles represented by Be In Box three and four the accelerations are high pass filtered and integrated twice to obtain Sy the simulator s translational displacements High pass filtering is necessary to eliminate the low frequency accelerations t
265. y be achieved by carefully identifying the sub systems of a flight simulator that bring the greatest contribution to realism that will provide the greatest benefits The specific design sub objectives relating to cost effectiveness can be seen in Figure 17 It is important to thoroughly investigate the commercially available subsystem components to gain a market perspective on their quality cost and availability Finally an aspect of equipment that brings added benefits to a design is the polyvalence of the equipment selected 36 Cost Benefit Analysis Contribution of Commercially Polyvalence of Subsystem to Available Subsystem subsystem Quality Components Components Figure 17 Cost Effectiveness Design Sub Objectives Indeed in a research environment the purchase of equipment that can be used for the conduct of other experiments or within other projects is regarded as a benefit For this reason the reusability of equipment and its ability to serve multiple purposes is an evaluation factor of the cost benefit analysis that is also taken into consideration The careful execution of the cost benefit analysis allows the spending of funds on the most useful and competitively priced equipment thus achieving cost effectiveness in the short and long term Three cost benefit analysis are performed in this study for the following components the flight simulator software the visual display and the pilot input system A cost benefit
266. y standard computer with a network card and the appropriate control software can serve as the control station In addition the flight simulation and head tracking software is run on the control station As the passenger concentrates on the actual simulation and flying of the aircraft the operator monitors the simulation software and the motion hardware The operator is explicitly responsible for maintaining safety by overseeing the simulation managing any spectators and stopping the simulation if required The operator is also responsible to assist the passenger with the boarding and tie in procedure prior to the flight simulation Risks to safety when using the flight simulator emerge almost exclusively from the use of the motion platform As such the operator must be able to quickly de energize the motion platform when a dangerous situation arises The surest way to guarantee absolute termination of the motion under emergency circumstances is to install a safety switch in line with the main 62 power supply of the motion platform The heavy duty SquareD safety switch that was installed in close proximity to the control station can be seen in Figure 36 Figure 36 SquareD Heavy Duty 60A Safety Switch Additional safety measures include the creation of an exclusion zone around the motion platform Finally software limiting and shutdown functionalities are implemented in the control software Indeed a Stop Motion Simulation button is
267. ycle as it reduces dependencies and coupling between each software components When there is a need to modify a specific component or to add functionality the idea is that only the required components are modified or new ones added without the need to modify the entire system Figure 40 illustrates the difference between a coupled approach and the functional approach used in this study In the coupled approach we can see that four software functionalities are implemented as a single software component 71 Coupled Approach Functional Approach Manage User Inputs E l Interact with FSX Interact with Motion Platform Washout Algorithm l Algorithm Manage Useri Interact with PUS LA Motion Platform Figure 40 Comparison between Coupled and Functional Approach This makes such a system very resistant to change or the addition of new functionalities In the functional approach four software components are created for each of the four distinct software functionalities As a result each of these software components are now independent from each other and change can be made on any of the components without affecting the others For example if we wanted to change the flight simulation environment to X Plane 9 instead of FSX we would only need to change the Interact with FSX software component This change would be effected with minimal or no disturbance at all to the other software components Finally since the software is now divi
268. ynamic array for receiving the actuator lengths double lengths rere Double dynamic array for receiving the actuator velocities double velocities pee Double dynamic array for receiving the actuator accelerations double accelerations pee UDP server object for lenghts UdpReceive udpLengths pre UDP server object for velocities UdpReceive udpVelocities pee UDP server object for accelerations UdpReceive udpAccelerations ex 207 Flag to differentiate between 3 and 6DOF string platformT ype ex Value containing the number of actuators int number_of_actuator endif MEDIATOR_H Mediator cpp pe Mediator cpp TIDE Visual C 2008 Express Edition author 2Lt Jean Le Bouthillier Latest Revision 11 May 2010 ek Description This class is responsible for providing a centralized location where interactions between all the object of the system are controlled As per the mediator pattern this class effectively encapsulate how all the object of the system interact Coupling is reduced by this approach as it prevents objects from explicitely calling or reffering to each other they must go through the mediator The advantage of de coupling is that the modification of one of the components of the system has little impact on the other components More specifically the mediator abstracts the motion platform from the GUI the

MSc Thesis: The Design, Implementation and Analysis of a

Contents

Download Pdf Manuals

Related Search

Related Contents