April 5,2012 Dr. Julio Militzer Dalhousie University 6299 South
Contents
1. Seed pe E ge 19 6 A Bie SUR a o ee Ae mene 20 6 1 DEVIatION OF laboU em 20 6 2 M 20 6 3 ACCOSSIDINIY iii italia dte dada Eat Error Bookmark not defined 1 ISSUES wnet e LII 22 A A m E 23 81 Balancing and Operation 23 6 2 Demonstrative ASPE ln eoe one ued ge SR Ru Ee adu nance 27 8 3 Durability ca ea ERA R aa Base Aaa d 31 SA Battery a ELLE 32 9 Budsete roads 33 MEC cm 34 11 essi O O O O O O O 35 Appendix A Gantt Chart REM 36 Appendix B Technical Drawihgs root ec recte bx Cn VAK aaa 37 Appendix C SeBaRo REI EU 38 Page 2 of 39 Final Build Report Self Balancing Robot Group 10 Table of figures Figure 1 Internal pendulum to balance robot no s s 6 Figure 2 Motor driven balancing robot 7 Figure 3 Passive safety feature bumpers s s ss sss asas s s ss AR R ARR R snas nnn 8 Figure 4 Passive safety feature Kicketand 8 Figure 5 Student interaction possibilitieS s 3333333 333s 3333334333 aaas eia 9 Figure 6 Final designiof robots s e ee ou sea usd Ee EE 11 Figure 7 Exploded view of robort enne nennen nini nasse nenne tasa sanis sesenta sana 13 Figure 8 S2. The main menu consists of 4 options e About quick explanation of SeBaRo e Remote When this option is entered the remote on the android can be used to control the robot Tuning The different parameters of the robot can be changed here either the balance angle or the three different PID s balance position and motor e Output The outputs from the sensors are transmitted via Bluetooth to a computer to input data on the encoders voltage and angle Changing gains To change the different PID gains go into the menu and select which of the tuning parameters you would like to modify The three knobs on the side panel change their respective parameter The top is for the proportional gain the middle for the integral gain and the bottom for the derivative gain To tune the balance angle use the middle Integral knob To send the gains once they are chosen press the top button The different gains are only set to be tuned within a specific range Once a knob goes past the maximum value in its range it will go back to the lowest value of the range and vice versa This can be changed in the code on the Arduino Retrieving data To retrieve data select the output menu option and activate the appropriate program on your computer Ensure that Bluetooth is enabled on the computer and that the Bluetooth power switch is activated found to the right of the LCD screen on the side plate The data is saved in a txt file Batt
3. Dom SHEET HOF Dalhousie University Balancing Robot Team 10 ara Back Plate Gregory Ryan Units mm Units inches Angles Course Qty X XX 15 x 005 4 025 Mech 40 0 XX 25 x xx 01 Material n TE x 4 50 xx4 02 0 08 Aluminium Scale Unit Nov 15 II SHEET OF Break Corners 50 00 Break Corners uve Project aw mir Dalhousie University Balancing Robot Team 10 Back Plate bottom 5 Gregory Ryan Units mm Units inches Angles Course Qty X XX 15 x 005 4 025 Mech 4010 04 2 25 xo 01 m 7 50 x ae 008 Aluminium X Nov 15 III m Reference Sheet 2 of 3 Dalhousie University Balancing Robot Team 10 omg Top Plate m Gregory Ryan Units inches Angles Course Qty ob am 005 4 025 Mech 4010 ac e feet 008 Aluminum Nov 15 11 ee 12 mp Sheet of3 R S Iert Drawino Dwn By Dalhousie University Balancing Robot Team 10 5 Top Plate Vents Gregory Ryan Units mm Units inches Angles Course uu X XX 15 x XXX 005 4 025 Mech 40 0 X X 25 x xx 01 Material Am xal 02 0 08 Aluminum X nota l FT mm Feetzo Dalhousie University Balancing Robot Team 10 Units inches Angles Course Qty ob am 005 4 025 Mech 4010 ac a Material 008 Aluminum ME Nov 1511 Pee 1175 F pm heet 3 of 3 Dalhousie University Balancing Robot Team 10 Bottom Plate Gregory Ryan U
4. 2 Page 15 of 39 Final Build Report Self Balancing Robot Group 10 The non linearized equations were also built in Simulink to obtain the model This model would allow for a more realistic response by adding noise and step inputs to the system Figure 11 is the Simulink model of the system Wheel Pos Wheel Ang PosPen Acc Wheel Ang AccPen Ang Pos Pen Voltage To Workspace Acc Wheel Setpoint DC Motor Transfer Function PID Controller Ang Acc Pen System Transfer Function Figure 11 Simulink model of the model 5 2 Sensors Based on the specification sheets from the sensors the resolutions are calculated and can be seen in Table 2 Resolution increases with a more sensitive device or with a higher bit analog to digital converter For the final design both changes will be implemented to increase resolution Table 2 Theoretical angle resolution results Device 10 Bit ADC 12 Bit ADC Accelerometer LIS244ALH 0 842 0 18 ADXL203 0 842 0 21 Gyro 1 2500 0 016 0 0040 ADXRS610 0 0054 0 0013 Assumes a control loop frequency of 100Hz Experimental testing was performed to check resolution calculations the results are shown in Table 3 The accelerometer tested is LIS244ALH The experimental results match the theoretical results verifying the method of calculation Page 16 of 39 Final Build Report Self Balancing Robot Table 3 Exper
5. Page 19 of 39 Final Build Report Self Balancing Robot Group 10 6 Assembly This section explains the different aspects of the assembly of the robot including the division of labour how the parts are held together and how to easily disassemble the robot to reach the internal components 6 1 Deviation of labour The construction of the parts will all be done by Dalhousie Technicians Each part is fully drawn in Appendix C The majority of the electrical components are assembled by the team with the exception of the battery components which is done by Dalhousie technicians The assembly of all purchased and built parts is done by the team 6 2 Procedure The following is a step by step procedure to put the robot together Ensure that all components are accounted for and mounted on PCB if required Secure PCB to mounting plate using 4 X 3M screws Secure Motor controller to mounting plate 2X3M screws Secure mounting plate to side plate using 3 X 3M bolts Secure first electric motor to side plate using 3 X 3M screws Secure panel mount power switch Attach 4 X LED s 3 X push buttons 3 X potentiometer dials and 2 X USB panel mount receivers to the interaction plate Secure the interaction plate to the side plate with 2 X 46 38 thumb screws Secure second electric motor to side plate using 3 X 3M screws Secure panel mount switches Secure bottom rod bracket to bottom plate using 4 X 5M bolts Secure side plates to bottom plates using 8
6. their usability their expected lifetime and their inherent complexity The final design is described in three main parts the control system which acquires the angular position of the robot and uses it to direct the motors the drive system which mechanically balances the robot and the chassis which houses the entire robot Once the final design is completed analysis is carried out to the exact characteristic requirements of the different components This includes the sensitivities of the sensors the torque from the motors the power from the battery the shock absorption capabilities and the accessibility Finally testing is carried out on the final product to ensure its ability to meet the original criteria Testing is done on the performance of the robot in its ability to balance to demonstrate the concept of control systems and to function overall It is found that the robot functions as designed and meets the design constraints and criteria initially described at the beginning of the project With supplementary functions added to the robot it is concluded that the robot surpasses expectations A detailed budget and schedule is included in the report as well as the SeBaRo user manual Page 4 of 39 Final Build Report Self Balancing Robot Group 10 2 Design Requirements The design constraints and criteria that the robot is required to meet are as follows in order of importance 1 The robot shall balance via an internal c
7. 29 1 gear motor w encoder Devantech 125mm Aluminum utility grade 3 56 thick 1x2ft Ball 5 8 10 pack Sheet 12 x24 30A Lipo safety sac Electronics board Dal and shell logo Bluetooth dongle Dalhousie Techs Nuts bolts rods 15 Supplier RobotShop com RobotShop com RobotShop com RobotShop com SparkFun com RobotShop com Greathobbies com RobotShop com RobotShop com Metals r us McMaster Carr Mcmaster Carr Mighty small cars BatchPCB Vinyl FX Robotshop com Dalhousie Page 33 of 39 Unit 1 R PF NY e NY 0 28 hr Cost 128 70 Incl w accel 34 00 103 00 79 99 41 19 79 99 200 00 41 99 28 34 22 25 0 99 40 30 20 00 128 50 37 60 6 65 50 00 100 00 Total 128 70 68 00 103 00 79 99 41 19 79 99 250 00 125 97 56 68 22 25 7 92 80 60 20 00 128 50 37 60 6 65 50 00 100 00 1 337 04 200 55 1 537 59 Final Build Report Self Balancing Robot Group 10 10 Results The design requirements are reviewed for evaluation 1 The robot shall balance via an internal control system for a minimum of 15 minutes while subject to no major disturbances V 2 The device shall be safe to operate in a classroom and laboratory setting V 3 The device shall balance on a surface area of 0 7 mx0 7 m V 4 Physical size of device shall not exceed 120 mm x 400 mm x 600 mm w l h V 5 Maximum mass of the device 10 kg V 6 The device shall be
8. 3 2 volts the battery should be replaced to avoid malfunctions Since there are multiple cells care needs to be taken to ensure that they are balance While they do not need to have the exact same voltage a range of 0 05 V is recommended Any more than that then the battery needs to be charged Page 22 of 39 Final Build Report Self Balancing Robot Group 10 8 Testing This section will go over the type of testing that will be performed The testing is to ensure that the criteria set early in the selection phase will be met and that the robot performs as expected Testing is also done to continually improve its performance by fine tuning the control system and its features 8 1 Balancing and operation The main criterion was for the robot to balance for 15 minutes without any external inputs It was determined early on in the testing of the robot that this was very easily achievable with nearly any PID input and therefore the majority of the balancing testing was done with an input force to the robot to destabilize it and see how it corrects itself These inputs or pushes are very difficult to gauge or measure and therefore were described as a small medium or large push The robot has the ability to communicate with a computer to output the data it reads internally This data includes the filtered angle from the gyroscope and accelerometer the encoder values and the voltage the motor controller sends to the motors From this data
9. Figure 2 is a drawing of the robot balanced by motors Page 6 of 39 Final Build Report Self Balancing Robot Group 10 Figure 2 Motor driven balancing robot 3 2 Angle measurement The robot needs to be able to determine its angular position to know whether it is balancing and if not how much it needs to maneuver to correct itself One possible way to accomplish this is to use a range finding sensor to measure the distance from the pendulum to the ground and obtain balancing position information This is however a roundabout way of finding the balancing point which is an angle A combination of a gyroscope and accelerometer is another more direct option for finding the balancing point The two sensors are required to obtain the angular position using sensor fusion filtering to combine measured angle from the accelerometer and integrated angular velocity from the gyroscope Due to the multiple inputs this is a more complicated design but it illustrates the robots balancing ability the best 3 3 Shock absorption The robot needs to be robust to handle balancing failures where the robot falls down either due to too strong of an input or bad system parameters therefore it requires suitable shock absorption to prevent damage to itself There are two routes to consider the first is passive safety features and the second is active safety features For passive safety features the casing of the robot itself could have shock absorption buil
10. REF 1 2 CH2 3 12 SDA H3 4 E 11 SCL 5 6 7 10 NC MG 8 CH6 Bluetooth Front DEEN RE DACH pueds w m W a 2 1102 Arduinox2 DR CR Hn WWW ARDUIND CC RESET 3 le RRDUTHD BAND 9 28 RESET GND 4 e 27 5v D2 5 6 26 Ao D3 6 e 25 A1 D4 og 6 24 A2 D5 8 23 D 9 22 A4 D7 10 6 21 A5 11 6 20 D9 12 q 6 19 010 13 l 18 AREF D11 14 e e 1 3v3 D12 15 l 16 D13 O O O Pololu TReX DMC pao GER gt a tle b WEE py um RS 232 TReX serial in RS 232 TReX serial out GND logic level TReX serial out logic level TReX serial in D RC or aux motor GND output sebr s d aux motor supply GND M2 VIN GND M1 outputs outputs TReX input output connection points
11. X 3M screws with spacers Attach shock absorption pad to front plate with 4 X 10M bolts with rubber stopper tops no nuts yet Secure front plate and absorption pad to side plates using the same 10M bolts as previous step and use the nuts to secure them lightly together Connect electrical components as per circuit diagram Attach bottom back plate to absorption pad with 2 X 10M bolts with rubber stopper tops Secure bottom back plates and lower portion of absorption pad to side plates using 2 X 10M bolts and secure lightly with nuts Slide upper back plate behind lower back plate Attach upper back plate to absorption pad and side plates with 2 X 10M bolts with rubber stopper tops and secure lightly with nuts Page 20 of 39 Final Build Report Self Balancing Robot Group 10 Place LIPO battery in battery bracket Secure battery bracket and rod assembly to top plate Connect battery to electronic components Secure the top plate to the side plates using 4 X 46 38 thumb screws Attach wheels using 3 32 Allan key 6 3 Accessibility To reach internal electronics once SeBaRo is fully assembled and in operation Make sure the power switch is in the off position zero side is pressed down Remove the four 46 38 thumb screws securing top plate to the side plates Carefully remove the top plate by pulling it straight up Disconnect the battery and place the top plate and the battery off to the side Remove the internal nuts securing the back plates upper
12. bolts Remove the back plates upper bolts Pull up the upper back plate and place it to the side The electronics should be fully accessible within a total time of less than 3 minutes Page 21 of 39 Final Build Report Self Balancing Robot Group 10 7 Safety Issues Lithium Polymer batteries which is what will be used for the robot have been known to explode on occasion if a short circuit occurs To mitigate this the assembly of the electrical components were supervised by the technicians Also the team will buy an anti explosion LIPO storage bag which can be used when charging the battery The safety bag would contain any fire or explosion that could occur if the battery is short circuited It will also be important to regularly carry out a visual inspection of the battery especially after removing it from the robot This inspection will be to look for any outside damage to the battery and especially to check for any damage to the insulation on the wires If damage is found it should be repaired immediately and if repairs are impossible the battery should be replaced The battery also has an ideal safe operating range of 3 2 to 4 2 volts The Mechanical Engineering department at Dalhousie University has a battery charger with built in features that prevent overcharging the battery beyond 4 2 volts The use of other chargers is not recommended should the battery be allowed rise significantly above 4 2 volts or to drop significantly below
13. less process to change the gains On the other hand there would be more components required for the robot increasing the cost and would be more complex to put together Controlling the gains by changing the in code values does not require additional hardware and is simple to do However it would be a longer process to change the gains and may require partially dismantling the robot Figure 5 is a drawing of the possible ways students can interact with the robot an LCD screen that shows the data and a remote controller where the PID gains are implementable Figure 5 Student interaction possibilities 3 5 Chosen Design The following design selection matrix Table 1 is used to aid in the design selection process Four criteria are scored with equal weight from 0 to 3 The criteria are illustration of control systems theory simple to construct safe and easy to use and long lasting A high score means the idea matched well with the criteria A low score indicates the idea did not match well against the criteria Cost is not considered a criterion due to the low variation in costs depending on the quality of the designed feature Page 9 of 39 Final Build Report Self Balancing Robot Group 10 Table 1 Design selection matrix Category Illustrates Design Safety amp Life Concept Complexity Usability time Balancing Positioning the 3 2 1 2 Method center of mass Positioning the 3 3 3 2 11 pivot point SS PA Angle Accelerometer 3 2 3 3
14. results for the different circumstances i e on a slope with a disturbance on the spot Jeremy wrote a cell phone app that lets us drive it around by remote control so yes No but with some good coding that should be possible but it is beyond our scope and abilities A survey of the class was taken at the end to gather the overall response of the class The survey consisted of a few yes or no questions with a comments and recommendation section Overall the feedback was mostly positive where the majority of the students learned something and found the demonstration interesting There were 58 students that filled out the survey the majority of the class and all were anonymous Page 29 of 39 Final Build Report Self Balancing Robot Table 6 Survey responses Survey Questions Did you like our demonstration find it interesting Did you understand what and why we were demonstrating Did you have a good grasp on systems control before this demonstration Did this demonstration increase your understanding of PID control Did this demonstration increase your interest in systems Was the robot easy to use Would having a hands on model like SeBaRo have helped you when learning control Did you learn something from our demonstration Group 10 Response yes 100 98 63 93 82 88 98 97 This survey shows that a large part of the class does not have a good understanding of the concepts of systems an
15. to specifications These specifications include durability interactivity and stability of the robot The different tests prove that SeBaRo has met all of the criteria that were listed back in September Some of these tests include a full demonstration done for the current Systems I class one on one interaction of SeBaRo with students performance comparisons of different PID gains on its balancing and multiple failures to determine its durability Recommendations are made on what can be done to add to SeBaRo in the future There are many different options to enhance the demonstration aspect that were past the scope of this project Page 1 of 39 Final Build Report Self Balancing Robot Group 10 Table of Contents e ee 1 Table OF COntents iii A bis 2 Table of TIgures ii a 3 MEE A ARNO 4 2 Design NEE e ET E 5 D si n PrOCGSS ii A A A ds 6 3 1 Balancingimethod edem eege dict 6 3 2 Anglemeasurermient u A 7 3 3 Shock aDSOrPtIO EE H 3 4 Student interactON EE 9 3 5 Chosen Desi iii aida 9 d Final Design nana s sim a aa dee Eegenen ct 11 41 Control System esseeri s ib bs inicia 11 4 1 1 gengt c 11 4 1 2 COMMUNICATIO EE 11 4 2 lt Drive SysteM aa a cda 12 413 CHASSIS e HP 12 Bi Design analysis s s ua apa sa d d 15 LEE lp R O EP O AEE ANE EEEE 15 5 2 e DC 16 5 3 Motor Torque CJ 18 54 e EE 18 5 5 Durabilit shock absorptiON ion tati
16. using 3 X 3M bolts Secure first electric motor to side plate using 3 X 3M screws Secure panel mount power switch Attach 4 X LED s 3 X push buttons 3 X potentiometer dials and 2 X USB panel mount receivers to the interaction plate Secure the interaction plate to the side plate with 2 X 3M thumb screws Secure second electric motor to side plate using 3 X 3M screws Secure panel mount switches Secure bottom rod bracket to bottom plate using 4 X 5M bolts Secure side plates to bottom plates using 4 X 5M bolts Secure 160X128 LCD screen to front plate Attach shock absorption pad to front plate with 4 X 10M bolts with rubber stopper tops no nuts yet Secure front plate and absorption pad to side plates using the same 10M bolts as previous step and use the nuts to secure them lightly together Connect electrical components as per circuit diagram Attach bottom back plate to absorption pad with 2 X 10M bolts with rubber stopper tops Secure bottom back plates and lower portion of absorption pad to side plates using 2 X 10M bolts and secure lightly with nuts Slide upper back plate behind lower back plate Attach upper back plate to absorption pad and side plates with 2 X 10M bolts with rubber stopper tops and secure lightly with nuts Place LIPO battery in battery bracket Secure battery bracket and rod assembly to top plate Connect battery to electronic components Secure the top plate to the side plates using 4 X 3M thumb screws Attach wheels
17. using Allan key Code and Spec Sheets Attached on CD rom Circuit diagrams The wires and connections in the following diagrams are the same colour as the wires they represent for convenience The interaction plate is shown below alongside the various devices that are present in the circuit It shows the naming convention used for the different components in the circuit diagrams based on their position on the interaction plate B1 B2 and B3 are the three Buttons for the LCD screen P1 P2 and P3 are the three potentiometers that we use to adjust the PID gains P1 Proportional P2 Integral P3 Derivative USB1 and USB2 are the two connections to the arduinos to update their codes USB1 connects to Arduino 1 and USB2 connects to Arduino 2 L1 L2 L3 and L4 are the four LED lights that are available to show error codes currently un coded SeBaRo Complete Circuit 10je n3 ay Se yor Je 0 13 07 1030 ZOP Z 49p02u3 Z JEJ T TE LT ep odu uA PCB components gt 2 E o N T OUIMPAY 03 XL T oumpuv 03 Xy r i mm x gt oo 2 2 m m N F z m w gt O x X NON 2 1Jepoou3 T DEAL T 1opo u3 my O eu is ra 2 AG AS 00 ON 11003en g gt S 2 zoumpiy 01 xy Z oulnpuy 03 XL s Interaction Plate ADC 14 VCC EH EH 13
18. 11 Measurement and Gyroscope d PA Range Finder 2 3 2 3 10 Sensor Safety and Kick Stand 3 2 1 1 7 Robustness Rubber 3 2 2 1 8 Stoppers Shock 3 3 3 3 12 Absorption Sw Material Student Adjustable PID 2 3 3 3 11 Interaction gains dx px Student 3 1 2 2 8 Implemented Control System Record data for 3 1 2 3 9 offsite analysis Sum Page 10 of 39 Final Build Report Self Balancing Robot Group 10 4 Final Design Figure 6 shows the outside shell of the robot with the main components labelled Top plate Side plate Front plate Interaction plate LCD Wheel Protrusion Figure 6 Final design of robot The control system includes most of the electronics the sensors that acquire data the microprocessors that analyses it and the motor controllers that receives commands from the microprocessors 4 1 1 Controller The main component of the controller is the microprocessor located in the Arduino the ATmega328 This device receives the voltages from the sensors and converts them to values the motor controller can read The motor controller then converts the values to voltages to power the motor the required amount to balance the robot The Proportional Integral Derivative controller is coded into the Arduino s microprocessor 4 1 2Communication There are two methods for the internal devices to communicate information to and from the user The first is via an LCD screen mounted in the chassis of the robot it di
19. April 5 2012 Dr Julio Militzer Dalhousie University 6299 South Street Halifax NS B3H 315 Dear Dr Militzer This is design team 10 s submission of the Final Design Report due April 5 2012 as requested in the Design Project Handbook This report is titled Final Design Report Self Balancing Robot The purpose of the report is to outline our final design decisions explain and analyse our testing and give out thoughts and recommendations for the project If you have any question concerning our project please contact any of the members of Group 10 Sincerely Luc Malo Jeremy Stewart Renske Ruben and Gregory Ryan Final Term Report Self Balancing Robot Group 10 April 5 2012 Luc Malo Renske Ruben Gregory Ryan Jeremy Stewart Dr Bauer Final Build Report Self Balancing Robot Group 10 Abstract The following report discusses the design selection assembly and testing of the self balancing robot SeBaRo project by group 10 The purpose of building SeBaRo is so that systems and II students have a hands on demonstration of how PID control works SeBaRo was built to be engaging and interactive as well as a safe and reliable classroom demonstration The design includes a shock absorption system for durability easily implementable controls for student interaction data collection options and intuitive mechanisms for easy adjustments disassembly and control A full set of tests were done to ensure the robot performs
20. SeBaRo while allowing it to return to its starting position It is important to keep in mind that the battery does not discharge linearly it will drain slower when it has more charge and as the charge goes down it will drain faster and faster This was seen when the battery was tested again after 30 minutes of the same type of testing immediately after and the average cell voltage was 3 It should be noted that if the voltage of the battery is below or near 3 5 it will deplete quickly The time to charge the battery of course depends on the voltage it starts at The one time the battery was completely depleted at 3V per cell 12 V total it took 90 minutes to completely recharge Page 32 of 39 Final Build Report 9 Budget Self Balancing Robot Group 10 The final budget is shown below the 200 that was given to our project to build the prototype is included The department granted the project 1500 for a total of 1700 SeBaRo therefore came out under budget by over 150 Table 7 Budget Part Accelerometer Gyroscope Micro processor Motor Controller LCD screen Bluetooth Battery Misc Electrical Motor Wheel Aluminum Rubber Neoprene Explosion bag PC Board Decals Bluetooth Machining Misc hardware Shipping Subtotal Taxes Total Name ADXL203 ADXRS610 Arduino nano Pololu dual 13A 16 6V Serial Graphic 160x128 BlueSMiRF Silver E flite 14 8V 4000mAh Wires capacitors resistors Pololu 12V
21. analysis can be done to determine many different aspects to the performance of the robot including e Maximum angle the robot can correct e Maximum speed the robot achieves e Typical oscillation period and span e PID performance Different PID s were inputted and then a small medium and large push was given to see how the PID gains affected the performance of the balance of the robot It is noted that the purpose of the project is not to find the optimal gains for the PID controller therefore this was not done during the testing The robot is meant to demonstrate how the PID works The robot was being shown to different students at one of these demonstrations the student was interested to see what would happen if the proportional gain was set to a very small number The student implemented the PID gains himself once a short explanation of how to do so was given The following figure shows the performance with a small medium and large push The PID gains implemented were 11 40 and 0 05 respectively Page 23 of 39 Final Build Report Self Balancing Robot Group 10 SeBaRo Post Control Analysis Angle Deg Position m T 0 10 20 30 40 50 60 70 80 time s 10 T T 1 T D 10 20 30 40 50 60 70 80 time s Figure 13 Student implemented PID gains 11 40 0 05 The filtered angle from the two sensors is shown in the first of the three graphs The posit
22. arn from it Each section was like a puzzle piece and the final design was chosen by taking the best of each section Brainstorming was done on each section to find different possibilities and each was given a score to compare 3 1 Balancing method There are basically two fundamental methods to balance an object shifting its pivot point below its center of mass or shifting its center of mass above its pivot point This section goes over the two methods and the pros and cons of each The other way to position the center of mass above the pivot point would be to essentially apply a force at the mass to shift its location This can be done by using a weighted pendulum powered by a motor As the pendulum swings to one side the center of gravity shifts with it because the pendulum is a large portion of the robots total weight This option would also give the opportunity to balance the robot at odd angles by leaving the pendulum extended Figure 1 is a drawing of a shifting mass pendulum b Figure 1 Internal pendulum to balance robot To shift the center of mass above the pivot point is to have the force acting on the pivot The pivot point could be moved slightly past the center of gravity essentially catching it by driving the wheels This method would require a motor to power the wheels in both the forward and backward directions By having motors attached to the wheels it also becomes possible to have the robot move around while it balances
23. ate From the position values the speed that the robot reached was calculated to be 1 8 m s This corresponds to wheel speed of 264 RPM when taking into account the wheel radius This value makes sense considering the free run speed of the motor is specified to be 360 RPM The position graph also shows that the robot required nearly 4 meters to correct itself 8 2 Demonstrative Aspect The main task of the robot is to be a hands on real life demonstration of what control systems are and how they work Therefore the most important test done was to have a demonstration to the systems I class This demonstration was not done at the best time during their semester since they had already completed that component of the course a few weeks previous The demonstration was done to teach something new to the class and increase their understanding in systems control The aim was also to increase their interest since interest in a subject aids in learning The demonstration was done during a regular class period for a duration of 50 minutes A quick introduction to the project was first given and followed by a review of PID control This opening presentation took 10 minutes so as to insure the class understood the purpose of the robot and how it Page 27 of 39 Final Build Report Self Balancing Robot Group 10 was controlled The majority of the demonstration was showing what the robot could do and how the PID effected its ability to balance The fi
24. capable of withstanding a minimum of 40 balancing failures V 7 lt shall be possible for students to implement their own control parameters V 8 The electronic components shall be accessible for repairs requiring less than three minutes to expose internal components V 9 The cost of SeBaRo is to be less than 1500 V Explanation in order of appearance e The robot will balance with a 3 cm amplitude for the duration of the battery life The battery will safely last over 3 hours if left undisturbed e The robot is an interesting and engaging demonstration to systems students in both a personal and classroom setting e The robot can be tuned to have a 3 cm amplitude and will stabilize a small push within 10 cm e The robot s size is 120 x 295 x 360 mm and weight is 4 3 kg e The robot has withstood over 40 failures on many different surfaces including tabletop e The robot can be disassembled within 50 seconds to reach internal components and reassembled in 90 seconds e SeBaRo s final cost was 1338 Page 34 of 39 Final Build Report Self Balancing Robot Group 10 11 Conclusion The robot was a success It is an interesting and engaging demonstration tool for systems students It has met every requirement defined in September and has gained many additional features The robot will be a great addition in both systems I and system II courses for students Some recommendations would be to insure the students have the opportunity to use the
25. d 93 of responders said that the demonstration increased it All of those surveyed said that they found the demonstration interesting Some of the comments and recommendations that were included on the survey are e That was pretty cool e Very interesting e Awesome demonstration Good job e Biggest value for me was better understanding of PID controls Really good job e It was a cool practical systems example e Well done Informative e think this is a great idea that will be useful in our next systems class e Cool project Systems sucks but you actually made it interesting being able to see a tangible use for the stuff we do in class was great e It looks tedious and I don t want to deal with it This list isn t the entirety of the comments given but do show the overall feedback from the class With the exception of one negative comment the comments were all positive The negative comment is the last listed above and is a valid argument This highlights the necessity of making the robot easy to use Page 30 of 39 Final Build Report Self Balancing Robot Group 10 8 3 Durability One of the original criteria was that the robot could sustain 40 failures These failures are when the robot falls overdue to improper gains or if too much of a force push was executed During routine operation and testing of the robot it has sustained well above the 40 failures it was designed for Addition assessments were done
26. e made that there were two medium pushes as seen between 20 and 30 seconds The large push began at 42 seconds SeBaRo Post Control Analysis 2 i i timg s o 10 20 30 40 50 60 70 80 Position m time s 60 70 80 time s 0 10 20 30 40 50 60 70 80 Figure 15 PID gains of 25 25 0 1 At first the gains looked like successful gains Minimal oscillations at equilibrium are seen for angle 0 5 degrees and position 2cm The robot finds stability relatively easily with the small and medium pushes However when a large push is inputted the robot becomes unstable As seen around the 50 second mark the robot overshot quite a bit and returns even farther than originally after the push The change in slope seen just prior to the 60 second mark is when the robot is caught to help it stabilize While the angle never became as large as the first experiment reached it could not stabilize Obviously these gains were not very good The graph below shows the max angle the robot was able to achieve and regain balance at 6 54 degrees when taking into account that the robot was balancing around one degree This angle also corresponded to the max voltage output and speed Page 26 of 39 Final Build Report Self Balancing Robot Group 10 SeBaRo Post Control Analysis 10 1 Angle Deg Position m Figure 16 Maximum balancing angle attained to d
27. eam 10 7 Rod Bracket Gregory Ryan Units mm Units inches Angles Course uu X 15 x XxX 005 4 025 Mech 40 0 XX 25 x xx 01 Material Am xX l 02 0 08 Aluminum X Fyov r F mm 125 43 115 43 R S Iert Drawino Dwn By Dalhousie University Balancing Robot Team 10 7 Rod Bracket Flattened Gregory Ryan Units mm Units inches Angles Course Qty XXX 4 45 xl 005 4 025 Mech 4010 XX l 25 X xx 01 Material Am xal 02 0 08 Aluminum X noto FT m Pez Dalhousie University Balancing Robot Team 10 ans Rod Bracket Bottom Gregory Ryan Units mm Units inches Angles Course Qty X XX 45 x 005 4 025 Mech 40 0 X l 25 XX 01 Material u j 405 14502 mere 0 08 Aluminum Nov 15 11 PT HI nts ym Preet 1 of 2 R S Iert Drawino Dwn By Dalhousie University Balancing Robot Team 10 7 Rod Bracket Bottom Flattened Gregory Ryan Units mm Units inches Angles se Mech 4010 X XX 15 x 005 4 025 40 0 XX 25 X XX D Material 4 50 xx 02 0 08 Aluminum X Nove m PT Fm pesirs Dalhousie University Balancing Robot Team 10 we Sensor Bracket Gregory Ryan Units mm Units inches Angles Course Qty X XX 15 x 005 4 025 Mech 40 0 X l 25 XX 01 Material u oc mere 0 08 Aluminum dano P ns ym Preet 1 of 2 Dalhousie U
28. eels were chosen to be larger than the depth of the chassis to avoid the chassis making contact with the ground during balancing These wheels were found to have extremely hard rubber wheels that adversely affected the ability for the robot to balance so a foam rubber tread was secured to the wheel This tread increases the wheel contact area and allows the robot to balance over a wider range of gains The Pololu motors require 12 volt and 300mA when in free run we have thus selected the GENS ACE 5000mAH 451 14 8V 25C Lipo battery Lipos are very lightweight rechargeable and have long life expectancies if properly cared for 4 3 Chassis The robots chassis is made up of three main sections the casing the shock absorption system and the internal brackets The casing is meant to serve as the frame work for the robot holding all of the internal components in place The shock absorption system is to protect the robot from sudden impacts due to high inputs or balancing failures The internal brackets hold the control system electronics and battery holding them securely to the casing Figure 7 is an exploded view of all the different parts of the robot Page 12 of 39 Final Build Report Self Balancing Robot Group 10 Figure 7 Exploded view of robot The casing is made of seven parts to make up the six sides of the robot The front plate is where the LCD screen display of the control system response is secured The backside is split int
29. ernal brackets are designed to secure the internal components of the robot to the casing Figure 9 shows the internal components of the robot The sensors and processor are mounted to a thin aluminum plate connected the side plate without the controls The battery housing secures the battery to the top plat Battery holder Accelerometer mount Sensor mount Motor Base plate Wheel Figure 9 Internal parts Page 14 of 39 Final Build Report Self Balancing Robot Group 10 5 Design analysis Analysis for the different components was done to ensure the design requirements were met Numerical calculations were made on the motor requirements and on the sensitivities of the sensors Simulation testing was done when for the geometry of the robot using the model found in section 1 5 1 Model A free body diagram of the system is shown in figure 10 where the two main components of the body are considered separately The transfer function is derived to create a simulation for the system on Matlab where a fine tuned PID controller is found Figure 10 Free body diagram of system A Newtonian approach was taken to derive the equations where the sum of the forces and moments were used Once the equations of motion were found for the pendulum and wheel the output of the motor was implemented Bulls My m n Tim 6 MpyTyLcos 0 6 mpr Lsin 6 udy 1 6 I myl 2T mpgLsin 0 m Lcos 0p
30. erometer Angle S Gyro Angle A Filtered Angle Angle dec a 5 pa A O lcd X Time s Figure 12 Test of Accelerometer Gyroscope and Kalman filter 5 3 Motor Torque The requirements of the motor are calculated by finding the maximum force the pendulum would have on the wheel The max angle that the robot would have that can be corrected is six degrees and the max weight is 4kg and the radius of the wheel is 0 2 m The max torque requirements are then calculated as T mx 3 T 4 kg sin 6 0 2 m 4 T 8kgcm 5 This is what the chosen motor supplies as max torque and is therefore sufficient 5 4 Power While both the Arduino and the motors are being powered by the battery the Arduino takes nearly no current therefore is negligible compared to the motors The motors required a 12V supply and have a 5A stall The required battery then needs to be at least 12 V and max torque used is 5A The Li Po batteries allow for easy recharging and a four cell is cost efficient and supply 14 8V The ampere chosen is Page 18 of 39 Final Build Report Self Balancing Robot Group 10 5000mAh which would allow for at least an hour of continuous run time if the robot was at full torque the whole time This is sufficient to demonstrate during a full class 5 5 Durability shock absorption Analysis on the required shock absorption is done to ensure when the robot fails and falls over i
31. ery Always check the voltages of the battery cells before and after operating the robot To check the voltages of the battery insert the dongle picture below and wait to see the 4 different voltages of each cell When the robot is not in use store the battery at 3 8V in the charging bag The figure above shows how the dongle is supposed to be connected to the battery The number 1 pin bottom pin as shown in the two figures above is inserted into the black wire input on the battery The digits show are the charges of each cell followed by the total battery voltage Use the safety charging bag when charging the battery Do not leave the battery unattended when charging A full charge will take approximately 90 minutes Disassembly Quick Access To reach internal electronics once SeBaRo is fully assembled and in operation Make sure the power switch is in the off position zero side is pressed down Remove the four screws 10 screws securing top plate to the side plates Carefully remove the top plate by pulling it straight up Disconnect the battery and place the top plate and the battery off to the side Remove the internal nuts securing the back plates upper bolts Remove the back plates upper bolts Pull up the upper back plate and place it to the side Complete Assembly Ensure that all components are accounted for and mounted on PCB if required Secure PCB to mounting plate using 4 X 3M screws Secure mounting plate to side plate
32. gure below is a snapshot of the demonstration done Figure 17 Class room demonstration A few things that should be added changed for future presentation are listed below Have students handle the robot themselves and try different gains e Smaller groups of students to insure all have the ability to see try the robot e Longer question period e Initially start with better gains to show its ability to correct a push The demonstration ended with a question period A highlight of the questions asked is summarized below with their responses Page 28 of 39 Final Build Report Self Balancing Robot Group 10 Table 5 Summary of questions asked during class demonstration Question Why do you need multiple PIDs Can it balance on a slope What is the max angle it can recover at How do you pick the gains Do you have the ability to move it a certain distance and have it reach equilibrium there Can it tune its gains itself Response To control multiple things to keep the two motors in sync we use one set of PID gains different PID gains are used for the position control based on how far away the robot is from its starting point and then depending on the location slope of the ground temperature surface friction etc the gains needed to balance the robot change Yes but it requires different PID gains the balancing on a flat surface From the testing done at least 6 5 degree angle You tune them for the desired
33. hock absorption system internal view 14 Figure 9 internal a iia 14 Figure 10 Free body diagram of system 15 Figure 11 Simulink model of the model 16 Figure 12 Test of Accelerometer Gyroscope and Kalman 18 Figure 13 Student implemented PID gains 11 40 0 05 24 Figure 14 PID gains of 20 30 and 0 2 respectively 25 Figure 15 PID gains of 25 25 0 Londa ama aeree eo area certat a bo ARAR zao od ek za Aa ed 26 Figure 16 Maximum balancing angle attained to date 27 Figure 17 Classroom demonstrationis ore eer eee cs O WO Aldaia 28 Page 3 of 39 Final Build Report Self Balancing Robot Group 10 1 Introduction The objective of our design is to clearly present an application as well as demonstrate concepts and theory of control systems SeBaRo will balance itself using a control method taught in MECH 3900 and MECH 4900 at Dalhousie University The design will provide students with the opportunity to interact with the robot by adjusting control parameters The effect of the adjustments will be obvious by the changes in the robot s ability to balance The design process began with determining the basic criteria and requirements that the robot needs to abide by Brainstorming then followed to find different methods of meeting those requirements The final design was chosen by comparing the different options through their ability to illustrate concept their safety
34. imentally determined resolution Consecutive Angle Measurements 0 84 09 0 85 1 69 525537 7 62 8 47 9 33 10 18 Measured Resolution 0 84 0 85 0 84 0 84 0 85 0 86 0 85 Group 10 The accelerometer zero g offset and sensitivity constants are experimentally determined in table 4 below These values match the values found in the specifications verifying measuring technique and methods The specifications list the Zero g offset as 1 65 V and the sensitivity as 222 mV g assuming a Vcc of 3 3V Table 4 Statistical calculation of zero g offset and sensitivity for the LIS244ALH accelerometer Acceleration 8 8 Samples 694 1042 Max V 1 877 1 452 Min V 1 896 1 432 Mean V 1 883 1 440 STD V 0 003 0 003 Precision Uncertainty V 0 003 0 003 Resolution Uncertainty mV 3 22 3 22 10 Bit ADC Zero g Offset V Sensitivity V g 1 662 V 0 0044 V 95 0 222 V 0 0044 V 95 Three angle measurements are plotted in the figure below the accelerometer angle the gyro angle and the complimentary filtered angle The results verify the presence of unwanted acceleration measurements and gyro drift due to integration Although the figure below shows the filter to be working further development on the filtering process is required to ensure a clean and accurate angle measurement Page 17 of 39 Final Build Report Self Balancing Robot Group 10 d Accel
35. ion of the robot relative to the location it started at is the second this is found by the output of the encoder and knowing the circumference of the wheel The third graph is the voltage the motor controller sends to the motors The three different pushes are best distinguished by the position graph as pointed out by the arrows The oscillations seen are because the robot does not balance perfectly at zero degrees this is impossible due to the many external variables wind air pressure unbalances vibrations in the floor These oscillations depend on the gains implemented and at these specific ones it can be seen that the angle is generally between 1 5 degrees and requires 2 V to sustain it The robot is also programed to return to its starting position once equilibrium is reached This also depends on the gains implemented for example after the medium push the graph shows that it took over 30 seconds to return to its original position after it found balance The reason for such a long delay is that the balance and position controllers are fighting each other the balance controller will not allow the robot to destabilize to return to the original position All three graphs together show many different things First the max angle reached was due to the large push and was 6 3 degrees The distance travelled due to the large push was quite disproportional to the Page 24 of 39 Final Build Report Self Balancing Robot Group 10 two smalle
36. iversity Balancing Robot Team 10 7 Side Plate Interaction Motor Mount Gregory Ryan Qty Units mm Units inches Angles Course Xxx 15 xx 4 005 4 025 Mech 4010 X l 25 x 01 Material 4 50 wx al 02 an 2mm Alluminum Nov 15 ili Sheet 2 of 3 E 7 ject Drawing Dwn By Dalhousie University Balancing Robot Team 10 um Side Plate Interaction Flattened Gregory Ryan Units mm Units inches Angles Course Qty mb md 005 4 025 Mech 4010 XX l 25 x xx 01 Material Am xal 02 0 08 Aluminum X BE Jon 612 PI mm Pete ject Drawing Dwn By Dalhousie University Balancing Robot Team 10 Control Panel Gregory Ryan Unifs mm Units inches Angles Course Qiy xo md 005 d 0250 Mech 4010 XX 25 x xx 01 Material T x 4 50 xxd 02 0 08 Aluminum BE jan 612 PP rm Peer j rawing Dwn B Dalhousie University Balancing Robot Team 10 ana Battery Bracket Gregory Ryan Units inches Angles Course Qty ob am 005 4 025 Mech 4010 755 rt 008 Aluminum Scale Unit me Nov 15 rir PPH mm Bheet of 2 R ject Drawing Dwn By orv Rvan Dalhousie University Balancing Robot Team 10 Battery Bracket Flattened Gregory Ry Units inches Angles Course Mech 4010 X 005 4 025 X xx 01 Material 0 08 Aluminum Unit noto FT m Pez R S Iert Drawino Dwn By Dalhousie University Balancing Robot T
37. mage the internal components but could dent the chassis over time making disassembly and reassembly more difficult and are thus not recommended Page 31 of 39 Final Build Report Self Balancing Robot Group 10 8 4 Battery The battery mainly affects two of our design criterion the weight and the length of time the robot can balance for We determined the balancing duration was slightly more important and calculated that the required voltage to operate all of our hardware could be met by a 14 8 volt battery in section 5 4 above Next looking at weight we chose to use a lithium polymer battery because of their high energy density high energy to weight ratio This choice minimized the weight impact of the battery while still granting us the required voltage and gave us a 5000 kWh capacity battery In theory with 5000kWh operating at maximum torque the robot could last for one hour but it was pointed out to us by Dr Swan that this was in fact incorrect and as the torque went up the current drawn from the battery would decrease and the battery would likely last a great deal longer During a day of testing that used the robot for approximately eight hours non consecutively tuning the gains and pushing the robot did not significantly deplete the battery During a 20 minute testing session where the robot was pushed repeatedly the battery started at 3 7 volts per cell and ended at 3 67 volts This was a testing session with constant inputs to
38. n days Project close out turnover Final Project report due Duration 9 days 15 days 8 days 3 days 2 days 2 days 3 days 0 days 40 days 7 days 22 days 7 days 12 days 0 days 22 days 16 days 14 days 4 days 4 days 4 days 3 days 1 day 14 days 14 days 3 days 7 days 30 days 7 days 5 days 14 days 0 days 15 days 10 days 10 days 4 days 1 day 7 days 1 day 2 days 3 days 0 days Start Sun 1 1 12 Sun 1 1 12 Sun 1 1 12 Mon 1 9 12 Mon 1 9 12 Mon 1 9 12 Fri 1 13 12 Mon 1 16 12 Mon 12 5 11 Mon 12 5 11 Mon 12 12 11 Mon 1 2 12 Mon 1 2 12 Fri 1 13 12 Mon 1 16 12 Mon 2 6 12 Mon 2 6 12 Tue 2 7 12 Sat 2 11 12 Wed 2 15 12 Sun 2 19 12 Wed 2 22 12 Thu 2 23 12 Thu 2 23 12 Fri 2 24 12 Mon 2 27 12 Thu 2 23 12 Thu 2 23 12 Mon 3 5 12 Sat 3 10 12 Fri 3 16 12 Mon 3 26 12 Mon 3 26 12 Mon 3 26 12 Thu 4 5 12 Mon 4 9 12 Mon 3 26 12 Mon 4 2 12 Wed 4 4 12 Fri 4 6 12 Mon 4 9 12 Finish Mon 1 9 12 Sun 1 15 12 Sun 1 8 12 Wed 1 11 12 Tue 1 10 12 Tue 1 10 12 Sun 1 15 12 Mon 1 16 12 Fri 1 13 12 Sun 12 11 11 Mon 1 2 12 Sun 1 8 12 Fri 1 13 12 Fri 1 13 12 Mon 2 6 12 Tue 2 21 12 Sun 2 19 12 Fri 2 10 12 Tue 2 14 12 Sat 2 18 12 Tue 2 21 12 Wed 2 22 12 Wed 3 7 12 Wed 3 7 12 Sun 2 26 12 Sun 3 4 12 Fri 3 23 12 Wed 2 29 12 Fri 3 9 12 Fri 3 23 12 Fri 3 16 12 Mon 4 9 12 Wed 4 4 12 Wed 4 4 12 Sun 4 8 12 Mon 4 9 12 Sun 4 1 12 Mon 4 2 12 Thu 4 5 12 Sun 4 8 12 Mon 4 9 12 ber 2011 Janua
39. nits mm Units inches Angles Course Qty X 15 x 005 4 025 Mech 40 0 XX 25 x xx 01 Material x 4 50 xx4 02 0 08 Alluminum Scale Unit Nov 15 II SHEET OF 10 Screw X 2 Reference Sheet 2 of 3 N Dalhousie University Balancing Robot Team 10 ana Side Plate Gregory Ryan Units mm Units inches Angles Course Qty X XX l 45 X 005 4 025 Mech 40 0 XX 25 x xx 01 Material n x 4 50 91 02 a 0 08 Aluminum dora e P Mem Bheet 1 of 3 Dalhousie University Balancing Robot Team 10 we Side Plate Motor Mount Gregory Ryan Qty Units mm Units inches Angles Course xl 15 xl 005 4 025 Mech 4010 X l 25 x xx d 01 Material 4 50 wx al 02 an 2mm Alluminum Nov 15 ili Sheet 2 of 3 7 Project Drawing m Dalhousie University VT Balancing Robot Team 10 ws Side Plate Flattened 9 Gregory Ryan Units mm Units inches Angles Course Qty X 15 x 005 4 025 Mech 4010 x4 25 xl 01 mi gt bu ew TT 008 Aluminum dano 12 59912 SHEET OF 10 Screw X 2 Reference Sheet 2 of 3 mm l ME Dalhousie University Balancing Robot Team 10 une Side Plate Interaction Gregory Ryan Units mm Units inches Angles Course uu X XX 15 x 005 4 025 Mech 40 0 XX 25 x xx 01 Material x 4 50 xxal 02 0 08 Aluminum minsi 912 nm Bets Dalhousie Un
40. niversity Balancing Robot Team 10 ane Sensor Mount Flattened Gregory Ryan Units mm Units inches Angles Course Qty X XX 15 x 005 4 025 Mech 40 0 X l 25 XX 01 Material u 54 00 34502 yerli 0 08 Aluminum don PP nts m Sheet 2 of 2 Project Drawing Dwn B Dalhousie University Balancing Robot Team 10 m Plate Assembly Exploded Gregory Ryan Units inches Angles Course Qty X XXX 005 4 025 Mech 40 0 XX l Material Various Scale Units ject Drawing Dwn By Dalhousie University Balancing Robot Team 10 7 5 Bracket Assebly Exploded Gregory Ryan Units mm Units inches Angles Course Qty XXX 4 45 xl 005 4 025 Mech 4010 XX l 25 x xx 01 Material x 4 50 xx l 02 0 08 Alluminum TO PH y SETI Final Build Report Self Balancing Robot Group 10 Appendix C SeBaRo Manual Page 38 of 39 SeBaRo Self Balancing Robot Safety Please read and follow all safety instructions prior to operating the robot e Check battery capacity prior to turning the robot on Ensure that all cells are well above 3 2 and below 4 2 volts and that cells are within 0 05 volts of each other e When inserting and removing top plate battery housing be careful with the wires from the battery so that they do not get pinched e Always put the battery in the safety bag when charging e If operating over a long range be mindful that the r
41. o two plates one designed to be removed easily to allow maintenance access and one secured to keep the rear plate together during operation The top and bottom plates are where the bracket system is secured and they are in turn secured to the two side plates The motors shock absorption system interaction devices and casing plates are fastened to the side plates which also act as a portion of the shock absorption system The shock absorption system is made of two main parts the bumpers and the shock absorption material Figure 8 shows a section view of the shock absorption parts The bumpers are simple bolts with rubber tips to prevent damage to the surfaces the robot falls on These bolts are slid through mounting holes in the front and rear plates as well as the side plates The bolts are free to move axially The shock absorption material is placed between the side plates and the front rear plates so that when the robot faces an impact the bolts push against the face rear plates which compresses the shock absorption material and distributes the load against the side plates The bolts are secured with nuts using just enough tension to secure the front and rear plates without compressing the shock absorption material Page 13 of 39 Final Build Report Self Balancing Robot Group 10 Front plate Top plate Protrusion Neoprene rubber Back plate Flange Left Side plate Figure 8 Shock absorption system internal view The int
42. obot can become unstable and travel at fast speeds e The robot needs to be turned off when turned or lifted off the ground as well as when it falls over because the sensors become confused Operating instructions The following image is a picture of the control panel on the side of the robot The main features are pointed out as they are described in the instructions below Down Enter Back Up Send Derivative gain Integral Proportional gain gain To turn robot on off The following steps need to be taken to start operating the robot e Hold the robot in the up near balancing position steady e Turn on the power switch Wait until all four LED s on the panel light up e When the motors start to move slowly let go of the robot The power switch is found on the left side of the robot when facing the LCD screen The off position is the zero on is the one To reset the gains to the coded values turn the robot off and on The robot needs to be turned off when lifted from the ground or if it falls over This is because the sensors are sensitive and these sharp motions resemble a large angle change and cause the code to output large voltages to the motors Using the LCD screen and menu The three black buttons control the menu on the LCD screen The top button moves up or updates the PID gains the bottom moves down and the middle button is Enter or Return The buttons function is determined by the menu screen you are on
43. ontrol system for a minimum of 15 minutes while subject to no major disturbances The device shall be safe to operate in a classroom and laboratory setting The device shall balance on a surface area of 0 7 mx 0 7 m Physical size of device shall not exceed 120 mm x 400 mm x 600 mm w l h Maximum mass of the device 10 kg The device shall be capable of withstanding a minimum of 40 balancing failures It shall be possible for students to implement their own control parameters 00 ir St UT de W ei The electronic components shall be accessible for repairs requiring less than three minutes to expose internal components 9 The cost of the project is to be less than 1500 Various reports and memos are to be submitted to adhere to the project guidelines listed on the course website The following deliverables have been submitted e Design Requirements memo October 34 2011 Design Selection memo November 7 2011 e Build Report November 21 2011 e Fall Term Report December 7 2011 e Individual Lab book December 7 2011 WWW webpage December 7 2011 e Final Build report January 16 2012 Final Term report April 5 2012 Page 5 of 39 Final Build Report Self Balancing Robot Group 10 3 Design Process To simplify the process of choosing the final design the ideas were broken intro four main sections how the robot will balance how it measures its angle how to ensure it is a safe and robust device and how the students will le
44. r ones as seen not only by the position graph but also the angle it reached and voltage the motors received The voltage outputted by the motor controller was 10 V to support this versus the 5 V from the medium push An interesting aspect of the voltage graph is that the voltage required to reach equilibrium after the small push was only marginally larger than what is used to keep the robot balanced A few changes were done to the gains to see the effects The gains used for this experiment was 20 30 and 0 2 respectively SeBaRo Post Control Analysis 4 T l 1 Angle Deg Position m Voltage Figure 14 PID gains of 20 30 and 0 2 respectively The first difference between the two is the angle at which it oscillates instead of a three degree range as with the previous gains it is approximately a one degree range This same phenomenon can be seen for the position plot where the oscillation distance is quite small approximately three cm Another difference is that from the large push the robot overshot its original position prior to gaining equilibrium We can also see that the robot did not require the full ability of the motor controller to do this as it only drew a peak of 6 V Page 25 of 39 Final Build Report Self Balancing Robot Group 10 The next figure is using another set of gains 25 25 0 1 and shows the importance of the derivative control for overshoot A note should b
45. robot themselves and implement their own gains to learn from them Page 35 of 39 Final Build Report Self Balancing Robot Group 10 Appendix A Gantt Chart Page 36 of 39 N w I RIR I A nr R o o Do H k LI RI RI Gad k RI FU RI RIR GI k RI a Li k RI RI 4 I 54 EX 54 E Task Name Prototype testing improvements Work on Final build report Revise drawings of robot Finalise drawings with technicians Update Budget Update Gantt Chart Final build design report for review by supervisor Final build design report due Begin building Procurement of parts material Machining Assemble electric circuit Assemble entire robot Complete assembled robot Controller Design Initial testing Design of the design Motor power function Controller performance Robustness testing Possible additions to design Inspection of working device by supervisor Adjustments to design Misc building assembly machining Addition procurement Controller improvements Final testing Final design refinement Robustness testing Performance evaluation Inpection testing require by date Write final build report Document final design Organize testing results Final Project report for review by supervisor Lab books due Organize presentation Practice rehearsal presentation Oral presentatio
46. ry 2012 Dec 4 11 Dec 11 11 Dec 18 11 Dec 25 11 Jan 1 12 SMIT TFISSMTWT FSSMTWT FSSMTWT F SISMTWTIFIS E MEI Page 1 February 2012 March 2012 Page 2 012 May 2012 June 2012 TIFISISMTWIIF S S MITWT F SIS MT WT F S SIMT T F SS MTVVT F S S M TVV T F S 5 MIT T FS SMTVVT F S SM TVVT F S S MTWT F S SM TVVT F SSSMTVVT F r q 49 Page 3 Final Build Report Self Balancing Robot Appendix B Technical Drawings Solid Edge Schematics in order of occurrence Face Plate Back Plate Back Plate Bottom Top Plate Top Plate Vents Top Plate Flattened Bottom Plate Side Plate Side Plate Motor Mount Side Plate Flattened Side Plate Interaction Side Plate Interaction Motor Mount Side Plate Interaction Flattened Control Panel Battery Bracket Battery Bracket Flattened Rod Bracket Rod Bracket Flattened Rod Bracket Bottom Rod Bracket Bottom Flattened Sensor Bracket Sensor Bracket Flattened Plate Assembly Exploded Bracket Assembly Exploded Page 37 of 39 Group 10 Break Corners I 10 X4 I D 3X4 I 122 00 PA I O Dalhousie University Balancing Robot Team 10 ar Front Plate Gregory Ryan Units mm Units inches Angles Course Qty X 15 x 005 4 025 Mech 40 0 X l 25 XX 01 Material TN 4 50 wx al 02 222 2mm Aluminium Dat Scale Unit Now 15 11 P
47. splays the PID controller gains in real time to help the user when tuning with the potentiometers or simply to have an easy way to check them without opening up the code The LCD screen also displays the various menus that allow the user Page 11 of 39 Final Build Report Self Balancing Robot Group 10 to change some of the control systems or put the robot into one of several modes The actual controls are mounted to the side plate as a set of three buttons and three knobs The second method of communication from the robot to the user is via Bluetooth This allows for wireless communication between the robot and another Bluetooth device both a computer and a cellular phone The wireless connection allows the user to change certain variables values in the code while the robot is running The wireless connection also allows for the user to control the position of the robot This means that the user can drive the robot around to get to different locations 4 2 Drive System The drive system consists of the different parts that will contribute to the robots movement when it is balancing This includes the motor wheels and batteries The self balancing robots wheels will be driven by a Pololu 12V 29 1 gear motor This motor was chosen because it is relatively low cost compared to other similar motors and it has enough torque to control the pendulums estimated weight The motor will be driving a pair of Devantech 125mm wheels The Devantech wh
48. t would not damage any of the internal components The forces involved were assumed to be the highest possible for a balancing failure with a safety factor of two The force was assumed to be similar to dropping the robot straight down F S F 2 mxg 6 The mass of the robot m is estimated at 3 5 kg leading to a total force of 68 67 N VVith this force and a designed spring material thickness of one centimeter to reduce bulk vve assumed a compression of one half Using these numbers a modulus of elasticity can be found FL L Es p Ax Ee 7 Where AL is the change in length L is the original length F is the force involved A is the area the force is applied to and E is the modulus of elasticity The area the force is applied to is determined by the geometry of the robot side plate reference draft dravving side plate The modulus of elasticity vvas found to be 1 3734E Pa Most elastomer materials used in shock absorption do not advertise the modulus of elasticity of the given material but do relate it to the Shore hardness or durometer rating The Shore hardness can be found vvith the follovving equation vvhere E is in MPa ig 0 54936 E 34 92627 0 74785 E 0 3492627 8 where S is the Shore hardness of the material It is found to be 38 5931 which rounds to 40 A durometer rating of 40 is very common in shock absorption elastomers and neoprene rubber is also commonly used for this application
49. t into it via energy absorbing materials or devices such as rubber springs etc The components within the robot would need to be placed so that they do not interfere with the case s ability to compress while absorbing impacts and one side of the casing would not be able to be directly connected to the wheels Page 7 of 39 Final Build Report Self Balancing Robot Group 10 and would need to use some medium to connect to the rest of the case Figure 3 is a drawing of bumpers with spring to take the shock of falling e P pnl Ze Ba KZ gt 1 l d 12 AB SET Figure 3 Passive safety feature bumpers An active safety feature could be to have a kick stand come out when the robot reaches its point of no return the angle past which it is no longer able to catch itself when falling This kick stand would catch the robot before it hits the ground reducing the amount of impact it would feel This feature would need to be controlled by the robot itself once it realizes it will not be able to catch itself and then the kick stand would be deployed Figure 4 is a depiction of the feature Figure 4 Passive safety feature Kickstand Page 8 of 39 Final Build Report Self Balancing Robot Group 10 3 4 Student interaction A potentiometer would be easy to use and understand In real time the students would be able to see how the system reacts when the gain is altered It would also be a quick and pain
50. to see what was required to achieve failure and trials on different surfaces was carried out Surfaces included e Hardwood e Tile e Ceramic e Thin carpet e Tabletop It is noted that while the robot is designed to be able to balance on top of a classroom podium it is not supposed to withstand a fall from such a height All falls tested were from the balancing position and from no higher Any demonstration of the robot failing should be done from the balancing position in a place where it will fall flat on the ground and not over a ledge Throughout every fall that the robot withstood the only damages caused were to the wheels and when the rubber stoppers The damage to the wheels was the thin foam we attached to them was peeled off The rubber stoppers were detached after a fall which ended with SeBaRo dragging along the floor for a short distance The wheels were quickly and easily fixed with measures implemented to insure it could not happen again The rubber stoppers have also been reattached and we have ordered new stoppers which we believe will withstand dragging along the floor better These rubber stoppers are only meant to stop SeBaRo s protruding bolts from damage the surfaces it falls on after failures and are not needed as a part of the shock absorption system The robot has also survived high velocity impacts with standing objects such as chairs and table legs with no visible damage These types of collisions should not da
Download Pdf Manuals
Related Search