Home

KGP-0109 DESIGN OF A WEARABLE BALANCE CONTROL

Contents

1. acceleration m s 7 8 9 1011 12 13 14 15 16 17 18 19 20 21 Time s F 3 Subject 4 Subject 4 Balanced Acceleration m s 2 Time s Subject 4 Unbalanced MAN AU M AAN ANAM Aw M M li WIN Vi VNWJM Ud y ty WAP WV I rn 12 14 16 Acceleration m s 2 Time s F 4 Subject 5 acceleration m s Subject 5 Balanced 8 9 10 11 12 13 14 15 16 17 18 19 20 Time s N D E c 2 i m o o o Subject 5 Unbalanced 9 10 11 12 13 14 15 16 17 18 19 20 Time s F 5 Subject 6 Subject 6 Balanced Axis Title 9 10 11 12 13 14 15 16 17 18 19 20 21 Time s Subject 6 Unbalanced N N E c 2 i S o o o 10 11 12 13 14 15 16 17 18 19 20 Time s F 6 Subject 7 Subject 7 Balanced Pom Creer nw md oc apy acceleration m s O 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 Time s Subject 7 Unbalanced hin Au Aug al acceleration m s 8 9 10 11 12 13 14 15 16 17 18 19 20 Time s F 7 Subject 8 Subject 8 Balanced acceleration m s 10 12 Time s Subject 8 Unbalanced acceleration m s 12 14 16 18 20 22 Time s F 8 APPENDIX G Data from Device Verification Table 1 Time between positive and negative peaks of shoulder width design verification trials
2. eee 114 Figure 57 X axis acceleration during Y axis tilt esses 115 Figure 58 Y axis acceleration during Y axis tilt 116 Figure 59 Z axis acceleration during Y axis tilt 116 Figure 60 Magnitude of Acceleration during Y axis Tilt 117 Figure 61 ADXL345 Data Sheet Og Bias Level eeeeeeeeeeeeeeeeennnen enne 117 Figure 62 Block diagram of algorithm eese nennen nennen nennen 118 Figure 63 Block diagram of device function sseeeeeeeeeeeeeeenne nne 119 Figure 64 Duino Balance A Wearable Balance Control Indicator eee 120 Figure 65 Placement of device on user ssesssseeseeeeeeeeeen ener nre nennen trennen entren 121 11 Figure 66 Arduino Duemilanove Microcontroller Board Arduino 2009 sess 122 Figure 67 Size of Arduino and its size relative to a 6 inch A to B USB plug Arduino 2009 123 Figure 68 ADXL345 Tri axis Accelerometer Sparkfun Electronics 2009 sse 124 Figure 69 CEM 1203 Buzzer Sparkfun Electronics 2009 sss 124 Figure 70 Rechargeable battery pack left and battery attached to Arduino right Allum 1999 Arduino 2009 p 125 Figure 71 9V Wall Adapter Plug Maker Media 20009 esses 125 Figure 72 On Off Slide Switch Digikey 20009 eese enn nen nennen 126 Figure 73 Prot
3. Subject 1 Trial Time s SW 0 80 SW 0 90 SW 1 10 SW 0 80 SW 0 70 SW 0 70 SW 0 80 SW 1 10 SW 0 55 SW 0 95 Table 2 Time between positive negative peaks of tandem device verification trials Subject 1 Trial Time s 0 45 0 40 0 50 0 60 0 35 0 45 0 70 0 55 0 30 a a4 4 4 4 4 4 4 4 0 75 The following shows a representative center of pressure plots for balanced black circle and unbalanced green circle sit to stand trials completed in design verification testing of subject 1 Yof FP in XofFP in The following show a representative balanced black and unbalanced green sit to stand trial for subjects 1 design verification testing Acceleration m s2 Acceleration m s2 Subject 1 Balanced 12 14 16 18 20 22 24 Time s Subject 1 Unbalanced d a en ae TY 10 12 14 16 18 20 22 24 Time s APPENDIX H Code for Device Parts of code were adapted from Arduino Forum 2009 and Faludi 2007 include lt Wire h gt define DEVICE 0x53 ADXL345 device address define TO_READ 6 num of bytes we are going to read each time two bytes for each axis Values specific to user define StartMinA 0 5 threshold acceleration define TE 0 6 1000 expected time between peaks in milliseconds define RingDuration 3 0 1000 time in milliseconds to ring buzzer define TestLength
4. 2 5 1000 total time in milliseconds for test TestStarted false TestFinished false RingBuzzer false BuzzerOff 0 TAA 0 time of max peak TAD 0 time of min peak AA 0 max acceleration AD 0 min acceleration byte buff TO READ 6 bytes buffer for saving data read from the device char str 512 string buffer to transform data before sending it to the serial port void setup Wire begin join i2c bus Serial begin 9600 start serial for output Turning on the ADXL345 writeTo DEVICE 0x2D 0 writeTo DEVICE 0x2D 16 writeTo DEVICE 0x2D 8 void loop int regAddress 0x32 x axis registers on the ADXL345 int regAddress 0x33 x axis register int regAddress 0x34 y axis register int regAddress 0x35 y axis register int regAddress 0x36 z axis register int regAddress 0x37 z axis register int x y Z readFrom DEVICE regAddress TO READ buff read the acceleration data from the ADXL345 each axis reading comes in 10 bit resolution ie 2 bytes thus we are converting both bytes in to one int H 2 x int buff 1 lt lt 8 buff 0 y int buff 3 lt lt 8 buff 2 z int buff 5 lt lt 8 buff 4 we send the x y z values as a string to the serial port sprintf str d d d x y z Serial print str Serial print 10 BYTE Delay is needed in order not to clog the port delay 1
5. Figure 81 As you can see there was a greater variation in the green COP which was consistent with an unbalanced trial After we completed and analyzed the ten shoulder width and ten tandem trials we compiled the results and determined the accuracy of our device as shown in Figure 85 The COP and KinetaMap data confirmed that all ten shoulder width trials were balanced so the device was not supposed to buzz at all during these trials However the device incorrectly buzzed and 140 produced a false alarm during one of the shoulder width trials As a result the device successfully identified a balanced STS activity for 9 out of 10 shoulder width trials Whereas for the tandem trials the COP and KinetaMap data identified all the trials as unbalanced thus the device was supposed to buzz for all ten trials But the device correctly buzzed and indicated an unbalanced situation during 8 out of the 10 tandem trials Table 18 Results of Device Testing Trial Type of Trials of Balanced of Buzzes Accuracy Shoulder Width 10 10 1 90 Tandem 10 0 8 80 Therefore the device was 90 accurate for balanced shoulder width STS trials and 80 accurate for unbalanced tandem STS trials 141 7 DISCUSSION Testing of our final device revealed that our device correctly identified 9 out of 10 balanced trials KinetaMap and COP data showed that all 10 trials were balanced therefore the device produced a
6. Measurement 24 R1 R20 Morisod J amp Coutaz M 2007 Post Fall Syndrome how to recognize and treat it Rev Med Suisse 2531 2536 Myllymake M 1996 Patent No 5 525 858 Nave R 2010 Center of Mass Retrieved 2010 from HyperPhysics http hyperphysics phy astr gsu edu HB ASE cm html 162 Nintendo 2009 Balance Games Retrieved April 10 2010 from WiiFit http wiifit com training balance games html Omega Engineering Inc 2003 Force Acceleration and Torque Retrieved 10 14 2009 from Omega Engineering Literature http www omega com Literature Transactions volume3 force2 html Omega Engineering Inc 2003 The Strain Gage Retrieved 10 14 2009 from Omega Engineering Literatue http www omega com Literature Transactions volume3 strain html Pai Y C amp Patton J 1997 Center of Mass Velocity Position Predictions for Balance Control Journal of Biomechanics 30 4 347 354 Pai Y C amp Patton J 1997 Center of mass velocity position predictions for balance control Journal of Biomechanics 30 347 354 Parker E B Fabeny B M Larson E C amp Monaco J F 2006 Patent No 6 997 882 Petelenz T J Peterson S C amp Jacobsen S C 2002 Patent No 6 433 690 Reithel B J 2010 Micro Electro Mechanical Systems Retrieved April 04 2010 from Faculty bus olemiss edu http faculty bus olemiss edu breithel b620s02 riley Micro_Electro_Mechanical_Systems
7. Trigger Bit A valise of 0 in the trigger bit links the trigger event of trigger mode to INTL and a value of 1 links the trigger event to INT2 Samples Bits The function of these bits depends on the FIFO mode selected see Table 20 Entering a value of 0 in the samples bits immediately sets the watermark status bit in the INT_SOURCE register regardless of which FIFO mode is selected Undesirable operation may occur if a value of 0 is used for the samples bits when trigger mode is used Table 20 Samples Bits Functions FIFO Mode Samples Bits Function Bypass None FIFO Specifies how many FIFO entries are needed to trigger a watermark interrupt Stream Specifies how many FIFO entries are needed to trigger a watermark interrupt Trigger Specifies how many FIFO samples are retained in the FIFO buffer before a trigger event 0x39 FIFO_STATUS Read Ont D7 D6 ps Ds o2 D2 or DO FO TmG Jo entries FIFO TRIG Bit A Lin the FIFO TRIG bit corresponds to a trigger event occurring and a0 means that a FIFO trigger event has not occurred Entries Bits These hits report how many data values are stored in FIFO Access to collect the data from FIFO is provided through the DATAX DATAY and DATAZ registers FIFO reads must be done in burst or multiple byte mode because each FIFO level is cleared after any read single or multiple byte of FIFO FIFO stores a maximum of 32 entries which equates to a maximum of 3
8. person to become more unstable As a result studies have used accelerometry to assess balance control and have shown that it is an accurate affordable alternative to using more expensive motion analysis system Therefore the team investigated accelerometry as a design alternative 2 5 2 Gyroscopes Gyroscopes are sensors primarily used to measure position tilt and orientation As a mechanical system a gyroscope is a spinning wheel with a high angular momentum The system is mounted within two rotors that make the system highly susceptible to external torque forces This design can be found in gyroscope toys MEMs gyroscopes such as those present in gyroscope ICs used in fall detection devices make use of a spinning disc built into a vibrating structure Because of the spinning disc configuration larger acceleration forces will act on the outside of the disc versus the inside of the disc Therefore changed in rotation will act with 58 greater force on the outside of the disc causing it to stray from its axis By using two discs as they stray from their axis the capacitance between them will change In order to accommodate the extremely small size of this complex system complex micromachining is necessary to create a gyroscope sensor Because of this these sensors tend to be costly Torrence 2008 Gyroscopes have been used to monitor the tilt of the body in particular the angle of the trunk at the hip and the angle between the body a
9. s Time s Time s Time s Time s Time s Time s SW 0 75 0 85 0 6 0 6 0 55 0 6 0 7 0 35 SW 0 8 0 7 0 1 0 1 0 5 0 65 0 65 0 15 SW 0 8 0 8 0 25 0 35 0 1 0 6 0 55 0 25 SW 0 9 0 9 0 6 0 3 0 4 0 4 0 65 0 1 SW 0 95 0 2 0 45 0 2 0 5 0 5 0 7 0 45 SW 0 9 0 3 0 48 0 55 0 2 SW 0 85 0 85 0 9 0 15 0 55 SW 1 15 0 95 0 5 0 25 0 5 SW 0 9 0 9 0 6 0 1 0 1 SW 0 9 0 65 0 4 0 5 0 4 SW 0 95 0 5 SW 0 95 SW 0 55 SW 0 8 SW 0 5 SW 0 1 SW 1 SW 0 85 SW 0 65 SW 1 SW 0 8 SW 0 95 SW 0 8 SW 0 75 SW 0 95 SW 0 85 SW 0 9 SW 0 95 SW 1 3 SW 0 75 SW 0 85 SW 0 9 SW 0 9 SW 0 85 AVG 0 85 0 71 0 71 0 59 0 55 0 75 0 74 0 35 STDEV 0 20 0 26 0 14 0 28 0 16 0 05 0 10 0 06 CV 0 23 0 37 0 19 0 48 0 29 0 07 0 13 0 17 D 1 Table 20 Time between maximum and minimum peaks of tandem trials T tandem AVG average STDEV standard deviation CV coefficient of variance Subject 1 Subject 2 Subject 3 Subject 4 Subject 5 Subject 6 Subject 7 Subject 8 Trial Time s Time s Time s Time s Time s Time s Time s Time s T 0 3 0 3 0 6 0 6 0 55 0 6 0 7 0 35 T 0 7 0 55 0 1 0 1 0 5 0 65 0 65 0 15 T 0 6 0 65 0 25 0 35 0 1 0 6 0 55 0 25 T 0 55 0 45 0 6 0 3 0 4 0 4 0 65 0 1 T 0 4 0 2 0 45 0 2 0 5 0 5 0 7 0 45 T 0 7 0 35 0 48 0 55 0 2 T 0 4 0 35 0 9 0 15 T 0 45 0 7 0 5 0 25 T 0 3 0 2 0 6 0 1 T 0 3 0 25 0 4 0 5 T 0 3 AUG 0 45 0 4
10. the oldest x y and z axes data are placed into the DATAX DATAY and DATAZ registers If a single byte read operation is performed the remaining bytes of data for the current FIFO sample are lost Therefore all axes of interest should be read in a burst or multiple byte read operation To ensure that the FIFO has completely popped that is that new data has completely moved into the DATAX DATAY and DATAZ registers there must be at least 5 us between the end of reading the data registers and the start of a new read of the FIFO or a read of the FIFO_STATUS register Address 0x39 The end of reading a data register is signified by the transition from Register 0x37 to Register 0x38 or by the CS pin going high For SPI operation at 1 6 MHz or less the register addressing portion of the transmission is a sufficient delay to ensure that the FIFO has completely popped For SPI operation greater than 1 6 MHz it is necessary to deassert the CS pin to ensure a total delay of 5 us otherwise the delay will not be sufficient The total delay necessary for 5 MHz operation is at most 3 4 us This is net a concern when using l C mode because the communication rate is low enough to ensure a sufficient delay between FIFO reads SELF TEST The ADXL345 incorporates a self test feature that effectively tests its mechanical and electronic systems simultaneously When the self test function is enabled via the SELF TEST bit in the DATA FORMAT register A
11. to a belt with a Velcro strap The device contains three biaxial accelerometers that can acquire data using six degrees of freedom and concurrently monitors heart rate using heart rate electrodes The accelerometers are attached to the belt so that one accelerometer placed over the midline of the back one accelerometer is placed over the right hip and the other is placed over the left hip Bluetooth technology is used to wirelessly transmit the acceleration and physiological data Parker Fabeny Larson amp Monaco 2006 One major benefit of attaching the sensors at the waist is the sensor s are close to the user s COM Another advantage of this design is the device is easy to use and does not require any manual activation of the device A 40 drawback of this device is potential false alarms as a result of activities of daily living such as bending forward and picking something up off the floor or reaching for something on a shelf Figure 13 Patent Number 6 997 882 Subject monitoring device and method Parker Fabeny Larson amp Monaco 2006 2 3 3 Current Fall Detection Technology Current methods of fall detection are summarized in Table 1 Of these methods there several commercialized products MyHalo Monitoring System BrickHouse Alarm System and Phillips LifeLine MyHalo consists of a strap that wraps around the user s chest a chest strap and sensing component Figure 14 Halo Monitoring 2009 The sensing compon
12. with user s Control interactio n with pressure points Does not interfere with daily activity Securely ressur attached p e points According to the results of the safety PCC the order of importance of these objectives from most to least important was as follows 1 No pressure points 2 Accurate 2 Sensitive to balance control 3 Minimal interaction with user s skin m Securely attached 5 Light weight The team then assigned relative weights to each objective according to their importance For example a highly ranked objective was assigned a higher weight indicating that it was more 81 important to the design The objectives of each of the two PCCs were weighted as shown in Tables 5 and 6 Table 5 Weighted Objectives Marketability Easytoputon takeoff 6 6 1 7 7 36 0 19 Aesthetically appealing 0 0 1 1 1 36 003 Table 6 Weighted Objectives Safety No pressure points 6 6 20 0 30 Minimal interaction with user s skin 3 20 0 15 4 20 0 20 A numerical evaluation matrix was then used to assess how well each design alternative met the objectives The numerical evaluation matrix contained objectives and constraints of the design in the first column The second column contained the weighted percentage assigned to each objective The next three columns contained the preliminary designs Each design was then ranked on a scale from 0 1 in an increment
13. 1 idata rate Tarn on and wakeup times are determined by the user defined bandwidth At a 100 Hz data rate the turn on and wakeup times are each approximately 11 1 ms For other data rates the turn on and wake up times are gach appronimately t 1 1 In milliseconds where t 1 idata rata Riov 0 Pago 3 of 3 ADXL345 ABSOLUTE MAXIMUM RATINGS Table 2 Acceleration Any Axis Unpowered Any Axis Powered 10 000 g 10 000 g 03Vto 16V 03Vto 16V 03V to Voove 0 3 V or 36 V whichever is less 03Vto 3 6V Indefinite 40C to 105 C 40 C to 105 C Stresses above those listed under Absolute Maximum Ratings may cause permanent damage to the device This is a stress rating only functional operation of the device at these or any other conditions above those indicated in the operational section of this specification is not implied Exposure to absolute maximum rating conditions for extended periods may affect device reliability THERMAL RESISTANCE Table 3 Package Characteristics Package Type amp x DeviceWelght 14 Terminal LGA 150 C w S C w 20mg ESD electrostatic discharge sensitive device Chaves devices and emu bord can Jsdurge without detecuon Akheugh the product fextures patented o proprietary proteccon Cheuityy damage may neto om deven subjeted te high energy ESD Thaian proper ES precaution should be taken te avoid performance degradation or loss of functio
14. 101 Figure 42 Attachment of the KinetaMap device sees nennen nennen 101 Figure 43 Shoulder width foot position representing a balanced condition esses 102 Figure 44 Tandem foot position representing an unbalanced condition sees 102 Figure 45 Root Sum of Squares of Acceleration of the STS in relation to hip flexion and extension of lysine EE 104 Figure 46 Example of shoulder width black and tandem green acceleration curves showing the time duration measured between organge lines esses eene nennen enne nnne nnne 105 Figure 47 COP plots of tandem green and shoulder width black trials esses 105 Figure 48 Average time between peaks RL 1 2 3 4 5 6 7 8 scscssssssssrssssessscsessscersecsesescsestsesers 106 Figure 49 X axis acceleration during Z axis tilt eese nennen nee 110 Figure 50 Y axis acceleration during Z axis tilt eese nennen 111 Figure 51 Z axis acceleration during Z axis tilt sess nene 111 Figure 52 Magnitude of Acceleration during Z axis Tilt eee 112 Figure 53 X axis acceleration during X axis tilt eene 113 Figure 54 Y axis acceleration during X axis tilt seeeeeeeeeeeeeen eene eene 113 Figure 55 Z axis acceleration during X axis tilt nennen 114 Figure 56 Magnitude of Acceleration during X Axis Tilt
15. 5 4 Through interviews with Lauren Roberts Appendix C a physical therapist of Fairlawn Rehabilitation Hospital the design team established that marketability and safety were equally important to the design Therefore an additional Numerical Evaluation Matrix Table 9 was created to determine mathematically which alternative design best met the objectives of the design as a whole Table 9 Numerical Evaluation Matrix Overall Objectives DESIGN 2 Shoe Waist mounted Ankle Overall Weight Insole accelerometer Brace V Tact Objectives l 32 2 32 8 13 3 24 1 17 2 38 2 11 7 16 84 According to the Numerical Evaluation Matrices the waist mounted belt best satisfied the objectives as it met 65 55 of the safety objectives and 76 4 of the marketability objectives and thus fulfilled 71 of the overall objectives of the design Therefore the team chose the waist mounted conceptual design and determined the functions specifications and feasibility of the design 4 3 FUNCTIONS AND SPECIFICATIONS The two main goals of the design were to strengthen the user s awareness of their balance control condition and to proactively force the user to perform a rehabilitation exercise In order to verify that the waist mounted accelerometer preliminary design would be capable of fulfilling the goals of the project the team established three necessary functions and used research and client interviews to create specifications that would
16. ADXL345 tri axis accelerometer Arduino Duemilanove Microcontroller Board and a CEM1203 buzzer The device also includes a rechargeable battery pack and slide switch for powering the device as well as a protoboard for connecting and attaching all the components The primary component of our device is the Arduino Duemilanove microcontroller board Figure 66 In our device the Arduino microcontroller board is used to collect the data from the accelerometer process the data and sound a buzzer if necessary One of the key components on the Arduino board is the USB connector A to B plug which can be used to connect the board to a computer to program a code onto the board or to charge or power the board The board requires between 7 and 12 Volts of power to run properly Another important component is the 121 power connector which can be used to connect the device to a 9V wall adapter plug allowing the device to be charged through a wall outlet The digital analog and power inputs can be used to connect additional components to the board Lastly the main component of the Arduino Duemilanove board is the ATmegal68 Microcontroller which was used for the digital signal processing Arduino 2009 Digital Inputs ATmega168 Microcontroller USB Connector http www liquidware com shop show ARD Arduino Duemilanove Figure 66 Arduino Duemilanove Microcontroller Board Arduino 2009 The board can be programmed using Arduin
17. Go 0 1 pF unless otherwise noted Table 1 Specifications Parameter SENSOR INPUT Maasurement Range Nonlinozeity Intec Axis Alignment Error Cross Axis Sensitivi OUTPUT RESOLUTION All g Ranges 2g Range 4g Range 8 g Range 16 g Rang SENSTIVITY Sensitivity at Nour Your Zour Scale Factor at Xaur Your Zour Sensitivity at Xoor Your Zour Scale Factor at Kaur Your Zour Sensitivity at Nour Your Zour Scale Factor at Xart Your Zour Sensitivity at Kour Your Zour Scale Factor at our Your Zour Sensitivity Change Due to Temperature Og BIAS LEVEL Og Output for Nour Your Og Output for Zour Og Offset vs Temperature for x y Amos Og Offset vs Tomperature for z Axis NOISE PERFORMANCE Noise ox y Axes Noise z Axis OUTPUT DATA RATE AND BANDWIDTH User selectable Measurement Rate SELF TEST Data rate 100 Hz 20V s Wss 36V F s F aaa a ee ee eee AAAA ut A 23335 lt ia uu Al minimum and maximam specifications are quarantoed Typical spectications are not guaranteed Cross axis sensitivity Is defined as coupling between any two axes 1 Bandwidth is half the owtput data rata Self test change amp defined as the output jg when the SELF TEST bit 1 0n the DATA FORMAT register minus the output g when the SELF TEST bit O in the DATA FORMAT ragistor Due to device filtering the output reaches Its final value after 4 x t when enabling or disabling seif test where T
18. S EE A E E E E E eae lel doen edie a E reba ae 42 Figure 15 Philips LifeLine Personal Emergency Response System Koninklijke Philips Electronics PLN 42 Figure 16 Patent Number 5 919 149 Apparatus and method for determination of body sway Allum 1990 enon ette ethos face iet e dde odes te Pune eu laeta Fe aer a o eaae a a trasera RE EH 48 Figure 17 Patent Number 6 984 208 Method and apparatus for sensing body movement Zheng 2006 50 Figure 18 Patent Number 4 750 480 Posture correcting devices Jenness 1988 51 Figure 19 Patent 10 793 729 Method and apparatus for improving human balance and gait Harry Collins Prplata amp Kleshinkski 2004 soniais eaea ahi iia nnne nemen nennen ennt enne 52 Figure 20 MEMS accelerometer and muscle behavior model Fung 1993 sese 56 Figure 21 The Effect of Gravitation Acceleration on an MEMs Accelerometer eeeee 57 Figure 22 Hip fracture most costly and traumatic effect of falls sesssssseeeeeeeren 65 Figure 23 Objectives of the design ere teni rede ttr thea egi eti errato bn aide ea esae 67 Figure 24 Preliminary design 1 Shoe Insole eene emen 73 Figure 25 Preliminary Design 2 Waist mounted accelerometer eeeeeeeeen ene 75 Figure 26 Preliminary Design 3 Ankle Brace rennen renn
19. Te would classify as balanced and below the same expected time Te would classify as unbalanced In choosing Te the team chose a threshold acceleration that encompassed as many unbalanced situations as possible without having false errors Therefore the team chose 0 60 seconds the latter end of the tandem time range as the expected time If the device calculated a time equal to or less than the expected time then the device would buzz If the device calculated a time greater than the expected time then the device would reset as shown in Figure 63 sound D user is off buzzer balance user is balanced Figure 63 Block diagram of device function The device was created for subject 1 based on this conceptual design The final design is described in following section 119 5 FINAL DESIGN The purpose of the device is to detect an unbalanced STS in terms of acceleration and notify the user when an unbalanced STS occurs The team found that the STS activity produced an acceleration curve that contained a positive and negative amplitude of acceleration that corresponded to the forward and backward sway of the STS activity respectively Based on preliminary results the time between the positive and negative peaks of these amplitudes was significantly longer in shoulder width trials than in tandem trials Comfort ratings of all subjects and COP plots confirmed that shoulder width trials represented a balanced situation and tandem
20. Wire available device may send less than requested abnormal buff i Wire receive receive a byte i Wire endTransmission end transmission H 5 APPENDIX I Abstract Accepted to Northeast Bioengineering Conference A Wearable Balance Control Indicator Advisors Profs Krystyna Gielo Perczak and Yitzhak Mendelson Department of Biomedical Engineering Worcester Polytechnic Institute Worcester MA Abstract Each year ome im three elderly fall Studies show that many factors contribute to an elderly person s rik of falling but if the factor causing imbalance is improved a person s risk of falling may be reduced A device that detects and alerts the user of am off balance situation before the fall occurs could identify a specific need for improved balance control This paper describes the design of a prototype wearable device that can give the O INTRODUCTION Every year aa an result o falls 1 Injuries due to falls cost the United States 20 to 30 billion health care dollars each year and with an increasing elderly population this amount could rise to about 54 9 billion by 2020 2 Studies show that multiple factors increase the risk of falling in the elderly including muscle weakness and abnormal balance patterns 3 These factors can be improved to decrease a person s risk of falling Medical professionals often use a sit to stand STS test which requires the patient to rise
21. Y and Z components of acceleration need to be converted into m s using the procedure outlined in see Appendix J Subtract the average of the first 5 seconds of acceleration data from the entire data set in order to zero each plot Calculate the magnitude of the acceleration as outlined in Appendix J for each trial and plot the acceleration as a function of time Import the NetForce data files into BioAnalysis software to obtain the plots of COP 151 8 1 2 Determine the Difference Between Balanced and Unbalanced STS Next the team then needs to establish the difference between the subject s balanced and unbalanced STS This will be done by determining the expected time Te above which is considered balanced and equal to or below which is considered unbalanced In order to calculate this value first separate all trials into two groups one containing all data belonging to trials where the physical therapist rated the STS lt 3 unbalanced and the other containing all data belonging to trials where the physical therapist rated the STS gt 3 balanced There should be at least 10 trials in each category to prove that results are significant If there are less than 10 trials in either category more data needs to be collected Verify that the trials rated lt 3 are unbalanced and the trials rated gt 3 are balanced by qualitatively comparing the COP plots of each group Unbalanced trials are characterized by a larger variation of the COP t
22. a laboratory force platform and can be used to assess standing balance Clark Bryant Pua McCrory Bennell amp Hunt 2010 The Wii Fit video game has balance mini games that are also intended to improve balance through an interactive gaming experience These games involve moving the user s COP and COM to complete various tasks in the video game and have been used in rehabilitation programs to improve balance control Clark Bryant Pua McCrory Bennell amp Hunt 2010 2 3 2 Fall Detection Patents Numerous patents exist for devices that detect falls and the methods by which these devices detect falls vary greatly These patents can be classified into three categories based on their method of detection ambient sensors active protection garments and accelerometer based 2 3 2 1 Active Protection Garments Active protection garments Figure 8 detect certain conditions that could predict a fall and upon these conditions deploy an airbag intended to cushion the impact of the fall The design pictured in Figure 8 contains airbags within non gas porous pockets The airbags deploy under the following conditions a rotation rate between sensors on the waistband or torso and at the bottom of the leg exceeding 45 degrees in 0 1 seconds nearly weightless condition for a period of 0 1 seconds and lateral and vertical accelerations meeting certain parameters with respect to each other and with respect to normal values Buckman 2006 These condit
23. account for 61 of the costs Among fractures hip fractures are the most costly and traumatic fall injury as 1 in 5 who suffer a hip fracture dies within a year Women are 2 3 more likely to sustain a fracture due to an unintentional fall than men and 72 of elderly admitted to the hospital due to a hip fracture were women Centers for Disease Control amp Prevention 2009 64 Fractures e 19 billion US e Hips Health Care e 61 of costs Legs Hip Fracture 1in5 will die 18 000 hip fracture dollars or 12 Billion Feet e Women 2 3 e Arm hand Nonfatal more likely Fractures in Injury women Figure 22 Hip fracture most costly and traumatic effect of falls As a result the team identified the causes of hip fractures and specified a cause that could be ameliorated via the present balance control indicator The leading causes of hip fracture were tripping and slipping but numerous variables e g lighting surface obstacles can contribute to tripping and slipping Kerr White Barr amp Mollan 1997 Fractures were found to be prevalent among people with lower body weakness problems with gait and balance and chronic diseases e g Parkinson s Diseases arthritis The National Council on the Aging 2005 Studies showed that falls can be prevented by increasing strength and balance control by exercises such as Tai Chi and strength training The National Council on the Aging 2005 In addition clinical
24. advantages of this design were that the accelerometer has been shown to accurately monitor balance control especially when positioned on a person s waist Ward Evenson Vaughn Rodgers amp Troiano 2005 Therefore the device could be placed in an accessible location easy for the user to locate and attach the device The location would also enable the user to wear it at any time of day unlike the Preliminary Design 1 shoe insole 74 Leeelecomete vy Breed XEM dieXecv er mounted ok Mgrs korso GO ophwrwaction cuneo geduren dese Leune enouet oxteleromeler Control Quer cadk comm utt N AM olin Vell al de guise Status edet oC eos lese 3 Pome toe Ayer ua Cee V gt belt Che Lm p ec Cosme akk Gc aches 4 ele wee E Com Oen WC d Ux x cam Figure 25 Preliminary Design 2 Waist mounted accelerometer 4 1 3 Preliminary Design 3 Ankle Brace Accelerometer Gyroscope The preliminary design shown in Figure 26 is based off the concept of vibration therapy described in the vibrotactile shoe insert patent mentioned previously Harry Collins Prplata amp Kleshinkski 2004 and studies that show changes in the angle and acceleration at the ankle joint help maintain 75 balance The design was an ankle brace that contains an accelerometer or gyroscope positioned at the ankle joint of the brace and removable vibrating components located near tendons of the ank
25. because they don t want to get fractures and they know it might be the last straw for independence If you need this then you probably need people around you anyway don t you think I mean like if people are confused and they re on medicine and they re home alone and they might just forget an move quick and maybe it d be nice to have it on at night can you have it on overnight Or maybe it could just be on when you re about to get up And would it be better if it was on overnight I think bedtime would be a good time to put it on and then have it start working as soon as you get up and get on the move Do you think we picked the best way to notify the person I think a noise would be best some quick sound that would alert you to sit for a while I m trying to think of any other times I think post op patients when they re on medication and they re asleep you get up and you try to do things automatically you know you just try to get up and I think a tone Do you think it would be helpful for otherwise healthy people Like if they re mentally well and physically well but maybe not so physically well I think of my sister you know she s had a few falls and I think she can be careless so I think of Rose C 5 She can be careless and get up without thinking And with her recuperating I think it would be helpful What s she recuperating from She had a fall she was up in the middle of the nig
26. design that utilizes vibrotactile feedback to enhance sensory function in the foot thereby improving human balance Vibration actuators or electrodes are contained within a wearable system such as a sock or shoe insole Figure 19 These actuators or electrodes provide electrical stimulation to the mechanoreceptors in the foot and ankle to increase their sensitivity and ability to transmit sensory information to the central nervous system Harry Collins Prplata amp Kleshinkski 2004 By increasing the sensory performance of the mechanoreceptors in the foot the device may improve balance in the user However this device is located in the user s shoes which is a difficult location for the elderly to use Also the vibration from this device could startle the elderly user and cause them to fall This device is 51 being pursued as a therapeutic tool by a small business Afferent Corporation however it is not currently on the market Figure 19 Patent 10 793 729 Method and apparatus for improving human balance and gait Harry Collins Prplata amp Kleshinkski 2004 2 4 2 Current Devices for Assessing Balance Control There are two devices the iShoe and the Wii Fit Balance Board and gaming system that are currently on the market for assessing and monitoring balance control A summary of these devices can be found in Table 2 The iShoe is a shoe insert that contains pressure sensors to determine abnormal pressure patterns in th
27. designed and tested 63 From the original client statement the team identified the main goal of the project to design an early and wearable balance control indicator for the elderly Three steps that were identified to achieve this goal as established by the original problem statement are as follows 1 Use a wearable sensor to collect data during daily movement and situations when individuals can lose balance 2 Analyze the data 3 Design an early balance control indicator based on the data analysis A key piece of information that was missing from the original problem statement was the purpose for creating a balance control indicator To establish the purpose of the design the team further defined the problem by using a strategy of design thinking called decomposition where the larger problem was broken down into smaller subproblems Dym amp Little 2004 To narrow the problem the team researched the root cause of falls in the elderly population Numerous risk factors can contribute to the likelihood of falling Since an elderly person can obtain any combination of these risk factors an infinite amount of situations could lead to a fall Therefore the team identified the most costly and traumatic effects of falls as shown in Figure 22 Nonfatal fall injuries were found to account for the majority of healthcare expenditures due to falls Centers for Disease Control amp Prevention 2009 and of these nonfatal injuries fractures
28. enable the device to achieve each function The team also identified constraints that could prevent the design from achieving the goals To strengthen the user s awareness of their balance conditions our device should notify the user if they experience an unbalanced situation Specifically our device should detect and notify the user of an unbalanced situation before a fall occurs In order to proactively force user rehabilitation operation of the device should require the user to perform a rehabilitation activity Specifically the rehabilitation activity should be a daily activity so that operation of the device is integrated into the daily life of the user Constraints of the design were the varying disabilities of the elderly such as osteoporosis decreased muscle strength and decreased range of motion The device needed to remain easy to use and to operate regardless of disabilities due to normal aging Two other constraints were the available testing equipment at WPI and affordability The team only had access to a force platform and the cost of the device should remain equivalent or less than the fall detection technology on the market so that the device can be afforded without the 85 aid of health insurance Figure 28 outlines the goals red functions purple specifications green and constraints grey STRENGTHEN USER S AWARENESS OF BALANCE CONDITION Notify user of unbalanced situation DAAE situation before fall o
29. fall before it occurs Limited transmission range Study not conducted on elderly but by students intentionally falling The overlying disadvantage of the aforementioned devices is that they do not detect a fall before the fall occurs Therefore these device are no proactive and do not help assess a user s risk of falling 2 3 4 Limitations of Current Technology The main limitation common among all commercial fall devices is that they do not detect or signal a fall prior to it happening Therefore the user does not have a decreased risk of falling 46 by wearing the device The devices containing a call button are not useful if the user is unconscious However the main downfall of devices on the market is that they do not decrease the risk of falling 2 4 METHODS FOR ASSESSING BALANCE CONTROL There are patents and devices on the market that are used for assessing balance control The patents are vibrotactile based and are used to alert a user of abnormal posture or sensory function There are two devices currently on the market the iShoe Trafton 2008 and the Wii Fit Balance Board and game Nintendo 2009 2 4 4 Vibrotactile Based Patents Vibrotactile based patents can be used for monitoring user movements and to correct abnormal posture The main difference between all of the vibration based feedback devices is the location of the vibration tactors There are no vibrotactile based patents designed specifically
30. false alarm for one trial We noticed that when the KinetaMap shifted during preliminary testing this caused a shorter time between the maximum and minimum acceleration peaks Therefore the false alarm during the device verification testing could have been caused by shifting of the device during the STS activity Testing of our final device revealed that our device correctly identified 8 out of 10 unbalanced trials Kinetamap and COP data showed that all 10 trials were unbalanced therefore the device produced a false alarm for two trials We noticed that when the KinetaMap shifted during tandem preliminary testing it caused a peak due to noise to have a more negative acceleration than the actual peak that occurred during the backward sway Therefore the two false alarms during the device verification testing could have been caused by shifting of the device during the STS activity Due to these results one improvement for the device would be creating an attachment that keeps the device secure and eliminates shifting during the STS activity The functionality of the Duino Balance balance control indicator device is fairly unique in that no commercial device has been released to perform the same task either by the same or different means One current device which bears the most similarity to the balance control indicator is the Wii Fit videogame system which is roughly based on a force platform with an interactive videogame interface that tests
31. for balance analysis that has been proven through testing to be as effective as a conventional gait monitoring system for the process of monitoring the sit to stand motion This eliminated the need for costly motion detection systems markers and force plates for the user to conduct a simple at home balance test 155 For future development of the device a larger population study would improve the viability of the time duration between peaks idea With a larger sample population including the elderly more support could be made for the observed pattern of a longer time duration between peaks in acceleration This would also confirm that this pattern was observable for elderly subjects With a larger population it could be either confirmed or refuted that there is significant variability between elderly subjects necessitating a reprogramming of the device between subjects Additionally improvements could be made to the device to make it more effective in a home or clinical setting For instance the added capability of the device to log number of STS performed and the number of balanced unbalanced trials and the accessibility of this information by a clinician would help a clinician or physical therapist to be sure that a subject was performing the activity and reporting correct results Finally a user interface for the computer to the device could allow not only for better clinician accessibility of this information but also for an easier mea
32. for fall detection or prevention but some are used to detect balance issues Patent number 5 919 149 Figure 16 is a wearable device used to detect balance problems mainly abnormal postural sway of the upper torso and to provide feedback to the user to help with rehabilitation The device collects angular velocity and body tilt angles using two gyroscopes One of the gyroscopes 12A is mounted in the middle of the chest for the purpose of measuring side to side movements or roll of the trunk The other gyroscope is attached on the side of the chest 12B and is used to measure forward and backward movements or pitch In this design feedback is provided to the use through visual and vibration feedback Visual feedback is provided by projecting an image of the body sway angle and angular velocity of the trunk onto a pair of eye glasses 24 Vibrotactile feedback is provided to the user by two vibration tactors that activate when the user has exceeded a particular tilt angle or velocity 47 Allum 1999 One benefit of this design is it can be worn under the user s clothing so it does not attract any additional attention to the user that could cause embarrassment This design also attempts to provide the user with feedback about the postural sway which could potentially prevent the user from falling or allow them to anticipate the fall and catch themselves assuming that the stimuli was activated soon enough and that the user was able to resp
33. htm Robert F Buckman J A 2006 Patent No 7017195 United States Rose D J 2005 The Role of Exercise in Reducing Falls and Fall Related Injuries in Older Adults Falls Free Promoting a National Falls Prevention Action Plan pp 19 27 Rubenstein L Z 2006 Falls in older people epidemiology risk factors and strategies for prevention Age and Ageing 35 S2 1137 1141 Runge M amp Schact E 2005 Multifactorial pathogenesis of falls as a basis for multifactorial interventions Journal of Muscoloskeletal Neuronal Interact 5 2 127 134 Schillings A M Mulder T amp Duysens J 2005 Stumbling Over Obstacles in Older Adults Compared to Young Adults Journal of Neurophysiology 94 2 1158 1168 Services U D 2007 October 31 HHS Strategic Plan Fiscal Years 2007 2012 Strategic Plan Retrieved September 20 2009 from United States Department of Health amp Human Services http aspe hhs gov hhsplan 2007 HHSStrate gicPlan07 12 htm content Shumway Cook A amp Woollacott M H 2007 Motor Control translating research into clinical practice Philadelphia PA Lippincott Williams amp Wilkins 163 SparkFun Electronics 2009 KinetaMap Product Info Retrieved from SparkFun Electronics http www sparkfun com commerce product_info php products_id 8725 Sparkfun Electronics 2009 Sparkfun Electronics Retrieved December 4 2009 from Triple Axis Accelerometer Breakout ADXL345 h
34. me a mug of beer And my other fall was with the dog I was standing talking with my son in law whatever all a sudden he turns around to go out in the backyard and the dog goes to chase him I didn t realize the leash was behind me laying on the ground picked me right up off the ground landed on my back and I had two fractures which weren t found for quite a while by MRI finally of my sacrum Took so long to heal because I wasn t aware of the fracture that s been painful How has that effected your daily life Well I can t carry heavy things weighty When I m bringing in my bundles grocery shopping which I always used to do I could do it but it causes pain in my lower back Even today I just had 5 Ibs of flour 8 Ibs of groceries and I could feel it So I tire more easily and I was always very active I can t do as much as I used to I like to be active What type of therapy or medicine or PT have you done or tried to do after your falls to help you I went to therapy they had me do exercises leg exercises laying down picking up your butt off the floor with straight knees leg raises laying flat with bent knees For my arms it was mainly range of motion exercises stretches you know like rattle I tried to do more they d be like no you re burnt you can t do any more and it was like once you re done you knew you could not The hardest one was laying flat on my left side raising that arm up gravity working again
35. of 0 1 on how well it met the objective 0 it does not met the objective and 1 it completely met the objective Each team member ranked the design on how well it mets each objective The rankings for each objective were averaged A total percentage of how well the objective was met by the design was calculated by multiplying the ranking by the corresponding weighted percentage Two numerical evaluation matrices were completed one determined how well 82 the preliminary designs met the objectives of marketability Table 7 and the other determined how well the alternative designs met the objectives of safety Table 8 The Preliminary Design that met the highest percentage of the objectives is highlighted in red Table 7 Numerical Evaluation Matrix Marketability DESIGN Waist Constraints Weight Shoe Insole mounted Anki V Tact 96 Brace amp accelerometer Objectives 6 6 17 6 4 4 4 4 7 6 15 2 3 8 3 8 1 7 13 6 3 4 3 4 5 6 11 2 4 2 7 0 life 6 6 IA 4 4 5 5 appealing 2 1 1 5 1 5 1 5 2 4 4 8 3 2 4 0 1 8 4 8 1 8 2 4 TOTALS 100 34 4 76 4 23 3 83 Table 8 Numerical Evaluation Matrix Safety DESIGN Weight Shoe AU Ankle Constraints mounted 9c Insole Brace amp accelerometer Objectives 23 2 14 5 5 8 14 5 with user s skin 11 2 11 2 1 4 11 2 9 5 15 2 9 5 9 5 control 13 3 15 2 1 9 15 2 3 0 2 5 2 5 1 5 with daily activity 2 5 2 45 3 0 2 0 5 4 4 5 5 4
36. of the program is include lt wire h gt function which was necessary to use the Wire library functions within the Arduino software The next several lines of code are used to define the address of the ADXL345 accelerometer and numbers of bytes in this case 6 that the device reads using the define function The device address found in the ADXL345 datasheet is 0x53 Analog Devices 2009 Next the C communication is initialized using the Wire begin function and the serial output is set to 9600 bits second Arduino Forum 2009 The next section of the program was used to read and write the data from the ADXL345 accelerometer The ADXL345 is turned on using the WriteTo function and the power control register address Ox2D which is used to take the accelerometer out of sleep mode Finally the regAddress function is used to read the registers for each of the three axes on the ADXL345 Arduino Forum 2009 The ADXL345 data sheet states that the address for the x axis is 0x32 and 0x33 the y axis is 0x34 and 0x35 and the z axis is 0x36 and 0x37 Analog Devices 2009 The data from each axis is squared using the sq x function then each square of each axis is added together using the operator and the square root of this sum is taken using the sqrt x function Arduino 2010 to represent the overall magnitude of the acceleration The third section of the code was used to process the accelerometer data and is where the parameters defined at th
37. regular chair will be set up as shown in Figure 84 The force platform will be placed directly on the ground and a wooden platform the same height as the force platform will be placed adjacent to the force platform The chair will be placed on top of the wooden platform and in front of the force platform so that when the subject sits in the chair their feet rest comfortably on the force platform The KinetaMap will be attached with Velcro to an adjustable belt The belt will be positioned so that the KinetaMap is mounted on the right side of the subject externally adjacent to the iliac crest Figure 85 Figure 84 Chair and force platform experimental set up 149 Figure 85 Attachment of the KinetaMap device Using the NetForce Software create a subject account for each subject Press subject and then add record Enter the subject information Select the first subject Press weigh and then tare and ask subject 1 to stand still with feet shoulder width apart on the force platform Click weigh Their weight will be collected and saved in NetForce software With the KinetaMap attached properly to the subject s right hip ask the subject to sit in the chair Start NetForce data collection and press tare to start data collection Ask the subject to turn on the KinetaMap Wait for the KinetaMap s LED to start blinking blue which signifies that the accelerometer has started collecting data Ten seconds after the LED starts blinking blue te
38. related to post fall healthcare costs would also be welcomed as well as the obvious benefits to the quality of life of individuals that can avoid the devastating effects of a fall Despite advancements in treatment available to fall victims the simplest means to avoid lasting fall related disabilities and quality of life limitations would be to avoid falling and injuring oneself in the first place Because of this the device would succeed in addressing improvements to the quality of life of subjects who used it properly to address balance control issues before the issues resulted in a fall By not only helping to prevent the lasting and 145 sometimes devastating effects of a fall injury but also mitigating the financial ruin that can be resultant of a costly hospitalization and rehabilitation following a fall the device serves to address the need for the elderly to live a good and satisfying life This device would have a clear influence on the health and personal safety of its users as it would attempt to remediate a potentially dangerous situation of an elderly person lacking proper balance Because of a simple and condensed list of components and the low cost nature of the device it is very manufacturable The prototype was built entirely from off shelf components and could easily be optimized to minimize cost as well as reduce size and weight The lithium ion rechargeable battery pack contained within the device reduces the need for
39. shown in Figure 33 Ward Evenson Vaughn Rodgers amp Troiano 2005 Tilt was eliminated because it was not sensitive to balance control Mathie Coster amp Lovel 2004 Through research the team identified that acceleration of COM was a key factor in determining a person s ability to stay balanced Pai amp Patton 1997 In addition the team found that pressure under the foot can indicate where a person s COM is located However research showed that location of COM along the BOS was not the factor that determines a person s balance condition but the acceleration of a person s mass can dictate if a person is balanced Pai amp Patton 1997 In addition acceleration was shown to be the most accurate measure of balance control during the STS activity and is often used in research to assess balance control Winter D 1995 Gross Stevenson Charette Pyka amp Marcus 1998 Acceleration is most accurately monitored on the trunk which would be an accessible location Ward Evenson Vaughn Rodgers amp Troiano 91 2005 On the other hand pressure under the foot would require the device to be located at the feet It has been shown that elderly have trouble putting on shoes and socks so a device similarly worn on the foot could be difficult for the elderly to use Dunne Bergman Rogers amp Rivara 1993 L1 p Figure 33 Signals to monitor balance control pressure acceleration tilt Therefore as
40. side of the hip and attaches to the user by a clip Two biaxial accelerometers with a sensitivity of 2G are enclosed in the monitoring device These accelerometers are used to measure the angle and acceleration of the body in order to determine if the user is horizontal for longer than two seconds and if any of the threshold accelerations or angles are exceeded The patent claims to detect falls with a 95 accuracy rate Petelenz Peterson amp Jacobsen 2002 An advantage of the design is it can automatically detect fall events and rapidly send an alert to a caregiver or call station without any input from the user which is important if the user were to become unconscious as a result of the fall By alerting a caregiver or call station of the fall the faller is able to receive help quickly and reduce long lie injuries and further injuries that could occur as a result of not being able to get up off the floor The small size of the device allows the user to wear the monitor without interfering with their daily activities The drawback of the design is that the fall is not detected until after the person is lying horizontal on the ground Another drawback is that attention needs to be paid to how the device is worn as false alarms are more likely to occur if the device is not aligned properly with respect to the vertical axis 37 Figure 10 Patent Number 6 433 690 Elderly fall monitoring method and device Petelenz Peterson
41. squares represented the forward sway and backward sway of the STS respectively 103 Flexion of Hip Forward Sway T I E c z M 2 o o lt Extension of Lower Limbs Time s Backward Sway Figure 45 Root Sum of Squares of Acceleration of the STS in relation to hip flexion and extension of lower limbs Through quantitative assessment the team observed that the time between the two peaks was longer in the shoulder width trials than in the tandem trials as shown by Figure 46 This was consistent with research that showed if a person accelerated quickly during the STS they became more unbalanced and the duration of the STS got shorter On the other hand if the person accelerated slowly during the STS they can maintain balance more easily and the duration of the STS got longer In addition the team observed that each plot had baseline noise and that in some occasions the final baseline did not equal zero This is due to the change in tilt of the device from the start to the end of the activity 104 Shoulder width n E c 9 5 c 2 v o o lt Acceleration m s Time s Time s Figure 46 Example of shoulder width black and tandem green acceleration curves showing the time duration measured between organge lines The COP plots of the tandem and shoulder width trials verified that the shoulder width trials
42. stand movement Medical amp Biological Engineering and Computing 43 265 272 Janssen W G Bussmann J B L D H amp Stam H H 2008 Validity of Accelerometry in assessing the duration of the sit to stand movement Medical Biological Engineering Computations 879 887 Janssen W G Kulcu D G Horemans H L Stam H J amp Bussmann J B 2008 Sensitivity of Accelerometry to Assess Balance Control During Sit to Stand Movement EEE Transaction on Neural Systems and Rehabilitation Engineering 16 pp 479 484 Janssen W Bussmann J Horemans H amp Stam H 2005 Analysis and Decomposition of Accelerometric Signals of Trunk and Thigh Obtained during the Sit to Stand Movement Medical amp Biological Engineering amp Computing Jenness L 1988 Patent No 4 750 480 Kelly P B amp Schoendorfer D W 2003 Patent No 6 611 783 161 Kemoun G Thoumie P Boisson D amp Guieu J 2002 Ankle dorsiflexion delay can predict falls in the elderly Journal of Rehabilitation Medicine 34 6 278 283 Kerr K M White J A Barr D A amp Mollan R A 1997 Analysis of the sit stand sit movement cycle in normal subjects Clinical Biomechanics 12 236 245 Koninklijke Philips Electronics 2009 How Philips Lifeline Medical Alarm Works Retrieved September 20 2009 from Philips http www lifelinesys com content how lifeline works index jsp Koninklijke Philips Elec
43. t exist yet The NMEA files follow the same type of system All of the NMEA files are named KinetaMapNMEAXXX csv where XXX is replaced by an integer between 0 and 255 There are two things you should be aware of when accessing the files The first is that the newest files won t necessarily be the ones with the highest number Let s assume that you ve run the KinetaMap 4 times there would be 8 files on the SD card one pair of files for each time the KinetaMap was run assuming your logging both NMEA and data files The files would be numbered 000 to 003 Now if you delete the 000 files then the next time the KinetaMap is run it will create file 000 again The other thing to be aware of is that the file numbers for the data file and the NMEA file might not match each other If you ve turned the NMEA logging off and created several data files then you turn the NMEA logging back on the file numbers for the NMEA files will be several numbers behind the file numbers for the data files Logging Scheme After finishing the startup procedure the KinetaMap will go into the main logging routine unless the default parameters have been changed to enable the configuration menu If the configuration menu has been enabled the device must receive a command via bluetooth to start logging This will be covered in depth in the Configuration Menu section of this user guide 2009 SparkFun Electronics Inc All Rights Reserved Product features speci
44. the COP These forces remain constant R acts at a distance p from the ankle joint Aj and W acts at a distance g from the ankle joint According to the inverted pendulum model the body sways in the A P direction while a person is standing still The body experiences a counterclockwise moment equal to Rp and a clockwise moment equal to Wg and has a mass moment of inertia equal to the product of the moment of inertia of the whole body about the ankle joint I and the angular acceleration of the body a act to create this forward and backward sway Rp Wg Ia where I is the moment of inertia of the whole body about the ankle joint and a the angular acceleration At time 1 the COG is ahead of the COP and Wg gt Rp resulting in a clockwise angular acceleration or the body 21 swaying forward As a result the body will increase the COP so that it lies anterior or in front of the COG This causes Rp gt Wg and a to reverse As a reverses the angular velocity o will start to decrease until at time 3 it reverses This causes the body to sway backwards Again the body needs to adjust to prevent itself from swaying further backward and so it decreases the COP until it lies behind the COG This causes a to reverse until at point 5 the angular velocity completely reverses and causes the body to sway forward This cycle continuously repeats while a person is standing still Winter D 1995 2 1 3 Maintaining Balanc
45. the KinetaMap that it should send the NMEA messages received from the EM 408 module to the bluetooth module This means that if an SPP bluetooth connection has been established 2009 SparkFun Electronics Inc All Rights Reserved Product features specifications system requirements and availability are subject to change without notice 7 KinetaMap GPS 08275 is a trademark of SparkFun Electronics Inc All other trademarks contained herein are the property of their respective owners KinetaMap UG 090401 ome eA ge sparkfun LECTRONICS KinetaMap User Guide 2009 04 01 between another device and the KinetaMap the GPS messages will be sent to the connected device i e blackberry If the config menu is also enabled then the GPS messages will only be transmitted once the config menu has been exited If the config menu and send gps to bt parameters are both disabled the bluetooth module will be turned off This will significantly improve the battery life of the KinetaMap log batt parameter Default 1 The log batt parameter tells the KinetaMap if it should log the battery voltage value or not By default the log batt parameter is enabled If the parameter is changed to 0 the KinetaMap will not save the battery voltage values to the SD card log nmea parameter Default 1 The log nmea parameter tells the KinetaMap if a NMEA file should be created in the current session If this parameter is enabled a NMEA file will
46. the foot and monitor a person s balance control by tracking their COP Although COP has been shown to be one factor that relates to a person s balance control acceleration of a person s COM has shown to be more reliable in indicating whether a person is balanced In addition the sensors need to be placed under the foot an inaccessible location and requiring the use of footwear Gyroscopes have been used in combination with accelerometers but only measure tilt Although tilt can indicate whether a person is balanced or unbalanced it is not a characteristic shown research to be characteristic of balance control Therefore using a gyroscope in the balance control indicator may not be the most accurate means of interpreting a person s balance condition 61 Through research pressure and acceleration were identified as the signals most sensitive to balance control Therefore preliminary designs utilized the measurement of these signals or a combination of these signals to create the balance control indicator 62 3 PROJECT STRATEGY Falls are the leading cause of unintentional death in the elderly population Centers for Disease Control amp Prevention 2009 Although there are many physical risk factors associated with aging that are inevitable and can cause falls balance control can be improved to reduce a person s risk of falling Stevens 2005 Current fall detection technology such as MyHalo Monitoring is not proactive and de
47. to independently assess their balance control and requires daily exercise In order to design a device that assesses balance control we had to understand how humans maintain balance and why the elderly experience an increased amount of falls In addition we identified current methods used to assess and monitor balance control and evaluated the advantages and disadvantages of each method The final balance control indicator aims to strengthen the user s awareness of their balance condition while requiring a rehabilitation activity These features make the device distinct among existing products and the team hopes this will enable the device to reduce elderly falls in the future This report describes the strategic design of the current balance control indicator The report will discuss the background of human balance control advantages and disadvantages of current technology and the gap in the current market Following the report will detail the team s project approach strategy of design testing and analysis and final design and verification 17 2 LITERATURE REVIEW A human s balance control system enables a person to maintain balance while standing during locomotion i e walking running and upon perturbation e g tripping Three sensory systems in the body are used to maintain balance vision the vestibular system and the somatosensory system As the functionality of these systems deteriorates a person s ability to remai
48. trials represented imbalance Therefore the device was designed to calculate the time interval between the positive and negative peaks of the STS acceleration curve and determine if the time corresponded to a balanced or an unbalanced situation In preliminary testing the time between peaks was not shown to be significantly similar between subjects so specifications were designed based on the data obtained from Subject 1 This section details the functions specifications and components used in the final design 5 1 OVERVIEW OF FINAL DESIGN Our final design and wearable balance control indicator is called Duino Balance and is shown in Figure 64 The device was enclosed in a plastic project box and is attached to a belt that can be worn around the user s waist The overall size of the device is 4 inches long by 2 5 inches wide by 2 inches tall and weighs about 2 pounds BE A Wearable Balance Control Indicator 120 The belt is worn around the user s waist with the device located on the right hip as shown in Figure 65 The device is only be worn when performing the STS activity and provides instant feedback to the user about their balance control If the user is balanced during the STS activity then the device does not buzz However if the user unbalanced during the STS activity then the device buzzes for 3 seconds Figure 65 Placement of device on user 5 2 DESIGN COMPONENTS Duino Balance consists of three main components an
49. 0 0 49 0 31 0 41 0 55 0 65 0 26 STDEV 0 16 0 18 0 22 0 19 0 18 0 10 0 06 0 14 T 0 35 0 45 0 44 0 597718 0 44 0 18 0 09 0 55 APPENDIX E Center of Pressure Data from Preliminary Testing The following shows a representative balanced and unbalanced sit to stand trial for subjects 1 4 labeled by a black and green circle respectively Subject 1 Yof FP in Xof FP in Subject 2 Yof FP in Xof FP in Subject 3 Y of FP in Subject 4 Yof FP in X of FP in X of FP in 64 E 2 APPENDIX F Acceleration Plots from Preliminary Testing The following show a representative balanced black and unbalanced green sit to stand trial for subjects 1 8 Subject 1 Subject 1 Balanced 2 1 5 1 NENNEN shoo di 00123445678 8 91011 Al 1314151616171819202021222324242526 Acceleration m s Subject 1 Unbalanced UAM M wr 1r mV MU VM acceleration m s 12 14 16 18 20 22 24 Time s Subject 2 Subject 2 Balanced acceleration m s Time s Subject 2 Unbalanced 1 5 SY 7 a AM M m PD IEP lh EMEN LP GOLA 1 acceleration m s 9 10 11 12 13 14 15 16 17 Time s F 2 Subject 3 Subject 3 Balanced a ee MM i o f acceleration m s 9 10 11 12 13 14 15 16 17 18 Time s Subject 3 Unbalanced hi Baii E EE WY Se MS ASV War ay
50. 3 entries available at any given time because an additional entry is available at the output filter of the device Rav 0 Page 18 of 18 A 11 APPENDIX B KinetaMap User Manual 303 sm KinetaMap User Guide 2009 04 01 Overview KinetaMap is a combination of technologies that allow data logging and transmission of GPS location and accelerometer data Kinesiology is a growing field of research where human movement and motion can be analyzed for various applications KinetaMap has the ability to capture things like pedestrian gait vehicle braking or package handling The logs give you GPS location and the raw acceleration readings it s up to you to decipher what they mean KinetaMap comes with basic firmware that currently supports acceleration and GPS logging Flip the power switch and GPS will be logged once per second and accelerometer readings X Y and Z axis will be logged at up to 100 Hz We designed KinetaMap to be as flexible and hackable as possible It has the LPC2148 USB bootloader built in which allows updates to the firmware easily and quickly over USB KinetaMap contains 1GB of flash that will last for weeks of continuous logging It attaches to a computer over USB and shows up on any operating system as a flash drive Quickly move files onto or off the KinetaMap without having to remove the flash card All log files are comma separated text files for easy parsing with Exc
51. 4 4 PRELIMINARY EXPERIMENTS ccescceeeceeceeeeeeeceeeaeeeaeeeaaecaaecaaeceaeceaeceaeceaeeeeeseneseaeeeaeesaeeeaaeeaaes 98 4 4 1 Materials and methodist rennene aa ae a aai aaa nnns entren 99 4 4 2 Results of Preliminary Te Sting ccccsscccccccssssssssscecececsssesseaeceescesseaaeseeeesessseseaeaeeeeeees 103 4 4 3 Conclusions of Preliminary Results ccccccccssscecessececeeseeeeceseeeeececseeeececseeaecesseaaeesesenaeens 107 4 5 Conceptual desiBgri aeuo te t eti iem te e eti et teu div det atis 108 FINAL DESIGN onic E 120 5 1 OVERVIEW OF FINAL DESIGN sseesseeeeeeeeeenee enne entren ens etre sn rn sn trs e trennen nnne nennen 120 5 2 DESIGN COMPONENTS ties tettn tee deae esi etos reta Eneas eee dene Gots den eine aei AEn 121 5 3 BUILDING THE DEVICE naiaiaee teer e eh t tie the eoe een e O en de e ERES 126 5 4 PROGRAMMING THE DEVICE essent nnne nnne nhsn tentent nnne nnne inns inns etse 132 DESIGN VERIFICATION cic ctp e cae eR ERE ARR NEXR E Fe Ee RENE IHE P Ren dong cenit ne enn 137 DISCUSSION sirni M 142 DESIGN VALIDATION inet nne neue eoe eR o e e a n e o XR UTR ERR EE RR eee RE RR a PN ERR e E a aia 148 8 1 PRELIMINARY TESTING WITH ELDERLY SUBJECTS ccesceeseceseceeeeeeeeeeeeeeaeesaeeeaaecaaecaeeeaeeeaeees 148 8 1 1 Materials and Methods eescseeeceesseceeseceeaeeceseeeesaeceea
52. 5 int a sqrt sq x sq y sq z magnitude of acceleration if a gt StartMinA amp amp TestStarted TestStarted true TimeStart T if T TimeStart gt TestLength amp amp TestFinished TestFinished true if TAD TAA lt TE RingBuzzer true BuzzerOff T RingDuration H 3 if RingBuzzer if T gt BuzzerOff RingBuzzer false RingBuzzer if TestStarted amp amp TestFinished if a gt AA AA a TAA T else if a lt AD AD a TAD T void setup pinMode 4 OUTPUT set a pin for buzzer output void loop f RingBuzzer 4 2048 RingDuration ring buzzer on pin 4 at 2048Hz for RingDuration Functions Writes val to address register on device void writeTo int device byte address byte val Wire beginTransmission device start transmission to device Wire send address send register address Wire send val send value to write Wire endTransmission end transmission reads num bytes starting from address register on device in to buff array void readFrom int device byte address int num byte buff Wire beginTransmission device start transmission to device Wire send address sends address to read from Wire endTransmission end transmission Wire beginTransmission device start transmission to device Wire requestFrom device num request 6 bytes from device int i 0 while
53. 7 8 Table 11 Results of shoulder width trials Subject of Trials Average Average Standard Coefficient of p value Comfort Time s Deviation Variance Level 1 5 1 35 5 0 85 0 20 0 23 lt 0 01 2 10 5 0 71 0 26 0 37 0 01 3 10 5 0 71 0 14 0 19 0 02 4 10 5 0 59 0 28 0 48 0 02 5 7 5 0 55 0 16 0 29 0 10 6 5 5 0 75 0 05 0 07 lt 0 01 7 5 4 8 0 74 0 10 0 23 0 12 8 5 5 0 35 0 06 0 17 0 23 106 Table 12 Results of tandem trials Subject of Average doe Standard Coefficient Trials Comfort Time s Deviation of p yatue Level 1 5 Variance 1 11 3 4 0 45 0 16 0 35 lt 0 01 2 10 4 0 40 0 18 0 45 0 01 3 10 3 8 0 49 0 22 0 44 0 02 4 10 3 8 0 31 0 19 0 60 0 02 5 7 3 5 0 41 0 18 0 44 0 10 6 5 3 0 55 0 10 0 18 lt 0 01 7 5 3 6 0 65 0 06 0 09 0 12 8 5 3 8 0 26 0 14 0 55 0 23 Table 13 p values comparing average shoulder width to average tandem time between peaks Subject of Trials p value 1 11 lt 0 01 2 10 0 01 3 10 0 02 4 10 0 02 5 7 0 10 6 5 lt 0 01 7 5 0 12 8 5 0 23 4 4 3 Conclusions of Preliminary Results Based on preliminary results the team verified that the time between the positive and negative peaks of the STS acceleration curve was longer in shoulder width trials than in tandem trials Through client feedback and COP data the team also showe
54. 78 amp _versi on 1 amp _urlVersion 0 Micro mercury switches Optical sensors Embedded in a coat b http www sciencedirect com scie nce _ob ArticleURL amp _udi B6V424 NMC8781 amp _user 74021 amp _rdoc 1 amp _fmt amp _orig search amp _sort d amp _d ocanchor amp view c amp _searchStrid 1 140042312 amp _rerunOrigin google amp _acct C000005878 amp _version 1 amp _ Pros Automatically detects a fall Alerts a monitoring station May prevent long lie injuries Cons Does not detect a fall before it occurs Only detects when user is Horizontal Can only detect falls in the sagittal plane 45 Sensor for Device Fall Detection Location Pros amp Cons CSEM wrist fall detector 2 MEMS 3 axis accelerometers http www csem ch docs Sh ow aspx id 9383 Wrist p www csem ch docs ow aspx id 9383 Pros Automatically detects a fall Alerts a monitoring station Discrete and doesn t draw attention to user Call button Able to wear to bed Cons Does not detect a fall before it occurs False alarms Difficult to distinguish fall from wrist acceleration data due to normal movement of the forearm during ADLs Z Star fall detector 3 axis accelerometer http www csem ch Trunk Pros Automatically detects a fall Small size Alarm when fall occurs Alerts a monitoring station Detects orientation of user Cons Does not detect a
55. Data 1 FIFO control FIFO status ADXL345 0x32 to Register 0x37 DATAX0 DATAX1 DATAYO DATAY1 DATAZO DATAZ1 Read Only These six bytes Register 0x32 to Register 0x37 are eight bits each and hold the output data for each axis Register 0x32 and Register 0x33 bold the output data for the x axis Register 0x34 and Register 1x35 hold the output data for the y axis and Register 0x36 and Register 1x37 hold the output data for the z axis The output data is twos complement with DATAx as the least significant byte and DATAx1 as the most significant byte where x represent X Y or Z The DATA FORMAT register Address 0x31 controls the format of the data It is recommended that a multiple byte read of all registers be performed to prevent a change in data between reads of sequential registers Register 0x38 FIFO CTL Read Write D7 6 Ds D4 03 D2 D bo HFO MODE Trigger Samples FIFO MODE Bits These bits set the FIFO mode as described in Table 19 Table 19 FIFO Modes Function FIFO Is bypassed FIFO collects up to 32 values and then stops collecting data collecting new data only when FIFO Is not full FIFO holds the last 32 data values When FIFO Is full the oldest data Is overwritten with newer data When triggered by the trigger bit FIFO holds the last data samples before the trigger event and then continues to collect data until full New data Is collected onty when FIFO Is not ful
56. Dunne R Bergman A Rogers L amp Rivara F 1993 Elderly Persons Attitudes Towards Footwear A Factor in Preventing Falls Public Health Reports in Brief 106 2 245 248 Dym C L amp Little P 2004 Engineering Design A Project Based Introduction Hoboken John amp Wiley Sons Inc Dym C L amp Little P 2003 Engineering design a project based introduction 2 Illustrated Ed Wiley Encyclopedia M W M 2010 Dorsiflexion Definition Retrieved April 10 2010 from meriam webster com www meriam webster com medical dorsiflexion Faludi R 2007 April 23 Rob Faludi Retrieved February 16 2010 from Buzzer Example Code http www faludi com 2007 04 23 buzzer arduino example code Fleming J amp Brayne C 2008 Inability to get up after falling subsequent time on the floor and summoning help prospective cohort study in people over 90 British Medical Journal 337 344 Fung Y C 1993 Biomechanics Mechanical Properties of Living Tissues 2nd Ed New York Springer Verlag Gielo Perczak K 2009 Major Qualifying Project Guide for Writing Report Worcester Worcester Polytechnic Institute Department of Biomedical Engineering Grabiner M D Pavol M J amp Owings T M 2002 Can fall related hip fractures be prevented by characterizing the biomechanical mechanisms of failed recovery Endocrine 17 1 15 20 160 Gray L 2000 Vestibular System Structure a
57. During all trails observers need to note whether the device buzzes And after each trial the physical therapist and subject should rate the comfort level as done in preliminary testing Using data from the KinetaMap plot the magnitude of the acceleration for each trial and calculate the time between the positive and negative peaks Label the trial unbalanced if the calculated time was less than or equal to T and it should have buzzed Label the trial balanced if the calculated time was greater than T and it should not have buzzed For each trial label what actually happened during the test buzz or no buzz For balanced trials calculate the percentage 153 of times the device did not buzz and for unbalanced trials calculate the percentage of times it did buzz Results should be at least 90 accurate for balanced trials meaning that the device did not incorrectly buzz for more than 10 of the trials Results should be at least 80 accurate for unbalanced trials meaning that the device did not buzz for more than 20 of the trials when it should have buzzed If the device achieves the accuracy rates then the device was successful at identifying a balanced and unbalanced STS activity and thus notifying the user of their balance control 154 9 CONCLUSIONS AND FUTURE RECOMMENDATIONS By creating a device that can proactively monitor balance control in the elderly population and utilize a repetitive daily activity we have impacted the abi
58. Each log in the data file consists of a timestamp the accelerometer readings the battery voltage and the GPS data GPS data will only be received when valid GPS messages are received from the GPS module If there is not valid fix or if the message was corrupted the GPS data will not be logged If one of these fields has been disabled in the configuration file or if there is no new data for the field the field will be blank for the given log Every time data is saved to the data file the tri color LED will blink the LED will blink Blue if data has been saved but there wasn t any valid GPS data saved If valid GPS data is saved the LED will blink Green The LED should blink at roughly 1 Hz All of the data in the logging sequence is saved to the data file and the NMEA file as long as the parameters are enabled properly Only one set of files is used for the entire session A session is only ended if power is no longer provided to the KinetaMap or if a USB cable is plugged into the device from a powered port LED Behavior Below is a table which explains the behavior of the tri color LED on the KinetaMap while using the default firmware Function LED Status Charging RED Loading System Parameters GREEN Initialization Waiting for BLUE Bluetooth Connection Logging Data No Blinking BLUE Valid GPS Fix 1Hz Logging Data Valid Blinking GPS Fix GREEN 1Hz Retrieving Data from the Log Files on the SD Card The
59. Lauren Roberts Fairlawn Rehabilitation Hospital The Tuite Family Louise Mayer Liz Dailey Brian Gaffey WPI Biomedical Engineering Department ABSTRACT Each year one in three elderly fall Studies show that many factors contribute to an elderly person s risk of falling but if the factors causing imbalance are improved a person s risk of falling may be reduced A device that detects and alerts the user of an off balance situation before the fall occurs could identify a specific need for improved balance control This paper describes the design testing and verification of a prototype wearable device that is worn on the right hip during the sit to stand activity STS to detect and notify the user of an unbalanced STS By signaling an off balance situation during STS our device notifies the user of poor balance control and identifies the need for balance control improvement TABLE OF FIGURES Figure 1 Location of the COM COG and BOS Shumway Cook amp Woollacott 2007 19 Figure 2 The body modeled as an inverted pendulum adapted from Winter Patla Prince Ishac amp Gielo Perczak 1998 t Seat t oraret tos tes oval Minhas dee Ate era dee e tesa eve iets 21 Figure 3 Sinusoidal Behavior of COP A Working Model Simulation Gielo Percak 2010 23 Figure 4 a Anterior Posterior Stabilization Strategies and b COP acceleration vs Position Shumway Cook amp Woollacott 2007 tere
60. MQP In our study we are investigating how the hip moves while sitting down and rising from a chair This data will be collected in order to identify a movement pattern that is indicative of an unbalanced situation This information will be used to develop a device that will detect when a user is off balance In our experiment you will be asked to wear a belt with an attached accelerometer device and sit and rise from a chair with your feet on a force platform a device similar to a bathroom scale You will be asked to rise from a chair several times There is a minimal risk of falling during the experiment but we will be holding the chair and standing next to you in order to catch you if you do lose balance The benefit for participating in our study is you will become more aware of balance control difficulty and gain awareness about your body and mobility This will also help with our project and lead to the design of a balance control indicator device for the elderly community If you are interesting in participating in our study and helping with our MQP please reply to this email and we can schedule a time for testing Thank you Amanda Martori Liz Tuite and Kevin Goggins Fall Detection MQP Team K5
61. NG Figure 9 PC Device Addressing Rav 0 Page 10 of 10 Table 11 PC Timing T 25 C Vs 2 5 V Vooww L8 V SCL Glock frequency SCL cycle time becas SCL nigh time tow SOL low time bao ssa Start repeated start condition hold time by car data setup time tana data hold time thu sm Setup time for repeated start tsu ss Stop condition setup time tas Dus free time between a stop condition and a start condition ts rise time of both SCL and SDA when receiving ta rise time of both SCL and SDA when receiving or transmitting tr fall time of SDA when receiving tz fall time of both SCL and SDA when transmitting tr fall time of both SCL and SDA when transmitting or recelveing Capacitive load for each bus line 456 25 06 13 06 350 0 06 06 13 grecgmeczccsg g o P Limits based on characterization results with fe 400 kHz and a 3 mA sink current not production tested All values referred to the Ve and the V levels given In Tabia 10 1 ts ts the data hold time that Is measured from the falling edge of SCL R applies to data in transmission and acknowledge times A ransmitting device must internally provide an output hold time of at least 300 ns for the SDA signal with nespect to Wee of the SCL signat to bridge the undefined region of the falling AL rum rr gre must be met only if the device does not stretch the low period ti of the SCL signal The rax murm value for t bs a functi
62. ORS OF FALEINQ 5 rer ha ret o Te rane Y eoe Y tek en Fo RE ne Rea eee ea o ne nh SR IER ERES 26 2 3 REDUCING THE NEGATIVE EFFECTS OF FALLS sess nen nennt nente nnn nnnis 31 2 3 1 Improving Balance Control eeeseessesesesesee nennen aiaa ii a a i sanas eser na naa 31 2 3 2 Fall Detection Patents ione rene tt ten reete tue en des aka taa ai ee iR ues 33 2 3 3 Current Fall Detection Technology ccccsssccccecessessnsececeeecessesseaeseeeescesseseaaeseeeeseeesesenaeess 41 2 3 4 Limitations of Current Technology esses eren nennen nennen nnns sns nnne 46 2 4 METHODS FOR ASSESSING BALANCE CONTROL essseseeeseeenee enne enne nnne trennen 47 2 4 1 Vibrotactile Based Patent eeeeesseceeccecesceeeneeeeesaeceeacessneeesaeeeeaaeceeaaessaeeeeaaeeeeaaeseeaeeen 47 2 4 2 Current Devices for Assessing Balance Control essere 52 2 5 PROACTIVELY MONITORING BALANCE CONTROL sees enne nennen tnter tentent 55 2 5 1 Accelerofmetry ios eee E E tese Tre eben eb ever Darei deua 55 2 5 2 GYFOSCOPES eee M C 58 2 5 3 Strain Gauges and Pressure Transducers esses nennen nennen ennt nnns 59 2 5 4 Motion Analysis sneen ete eei tette tree atc best ta tdeo ects a egere as 60 2 5 5 Advantages and Disadvantages of Current Balance Monitoring Methods 61 3 PROJEGESTRATEQGY sei eps inf te SER dates teras ete be bias ee eae a e Reig
63. Project Number KGP 0109 DESIGN OF A WEARABLE BALANCE CONTROL INDICATOR A Major Qualifying Project Report Submitted to the Faculty Of the WORCESTER POLYTECHNIC INSTITUTE In partial fulfillment of the requirements for the Degree of Bachelor of Science by Amanda Martori Elizabeth Tuite Kevin Goggins April 29 2010 Approved Professor Krystyna Gielo Perczak Major Advisor Professor Yitzhak Mendelson Co Advisor Table of Contents AUTHORSHIP Zim E 5 pe d ei Bsbrcis dE 8 ABST RAG pem H 9 TAB E OFFIGURES idet tt tet metere aor doo ated ets 10 TABLE OF TABLES eee here RE P Ee OO IND RIVE Y ERE edt sp eval S RES vtr eas 13 EXECUTIVESUMPMBBARY 5 ecran tener eae e eae to ee e geh eget pe ern ue a en ue qe been inten 14 T INTRODUCTION idet t teet ertet teretes eile ter ets 16 2 LITERATURE REVIEW eite o ere tne e aiea RE an Ea ea deu ede ae Eae ee en eb resta eu ee i nune us 18 2 1 UNDERSTANDING BALANCE CONTROL ccccceeeceeseeseeeeeeeaeceaeceaeceaeceseceaeeeeeeseeeseaeeeaeeeaeesaaeeaaes 18 2 1 1 Basics of Balance ie tete tun tate i ttu rit eer edet ttt da 18 2 1 2 Human Balance Control and the Inverted Pendulum Model esses 20 2 1 3 Maintaining Balance during Perturbation eese eese enne nnne 22 2 1 4 Example of Balance Control Balancing a Pencil esee 25 2 2 RISK FACT
64. S amp Baldus H 2010 A body fixed sensor based analysis of power during sit to stand movements Gait and Posture 31 272 2778 165 Appendix A ADXL 345 Data Sheet ANALOG DEVICES 3 Axis 2 g 4 g 8 g 16 g Digital Accelerometer FEATURES Ultralow power as low as 40 pA In measurement mode and 0 1 pA in standby mode at Vs 2 5 V typical Power consumption scales automatically with bandwidth User selectable resolution Fixed 10 bit resolution Full resolution where resolution Increases with g range up to 13 bit resolution at 216 g maintaining 4 mg L SB scale factor In all g ranges Embedded patent pending FIFO technology minimizes host processor load Tap double tap detection Activity inactivity monitoring Free fall detection Supply voltage range 2 0 V to 3 6V VO voltage range 1 7 V to Vs SPI 3 and 4 wire and PC digital Interfaces Flexible Interrupt modes mappable to either Interrupt pin Small and thin 3 mm x 5 mm x 1 mm LGA package APPLICATIONS Handsets Medical Instrumentation Gaming and pointing devices Industrial Instrumentation Personal navigation devices Hard disk drive HDD protection Fitness equipment GENERAL DESCRIPTION The ADXL345 is a small thin low power 3 axis accelerometer with high resolution 13 bit measurement at up to 216 g Digital output data is formatted as 16 bit twos complement and is acces sible through either a SP 3 or 4 wire or PC digital interface The ADXL345
65. a physical therapist from Fairlawn Rehabilitation Hospital Questions are bolded and responses are not bolded Would the device be useful in a clinical setting Home setting If so how If not why yes probably in a clinical rehab setting we could use it in addition to balance scales that we use during a PT eval ie Berg or Tinetti scale which all can help identify those who are a high fall risk it may be too costly for home insurances may not buy into it C 8 What type of patient do you think this would be used for patients who are in a fall risk catagory Parkinson s MS post stroke general medical elderly pts over the age of 80 Would you change anything about this device it seems simple in terms of either off balance with sit to stand or not it is good that it is clear cut How could this device benefit an elderly individual again it could identify high fall risk patients need for a person to get P T for balance training need for an assistive device such as cane or walker C 9 APPENDIX D Data from Preliminary Testing Table 19 Time between maximum and minimum peaks of shoulder width trials SW shoulder width AVG average STDEV standard deviation CV coefficient of variance Subject 1 Subject 2 Subject3 Subject 4 Subject 5 Subject 6 Subject7 Subject 8 Trial Time s Time
66. aeceeaeeceaeeeeaaeeseaeeseeeeesaeeseaaeeeeaeeeaes 148 8 1 2 Determine the Difference Between Balanced and Unbalanced STS 152 8 1 3 Verity the device 5e et set ast ae ti A Acne Ree anda 153 9 CONCLUSIONS AND FUTURE RECOMMENDATIONS esses nennen nnne nennen nnne tnnt 155 GLOSSARY secs eie oe tie e Pede ere e t eu ere prx e E an io OR EARN Ie de diae eda 157 uddsiddsx M 159 Appendix A ADXIE 345 Data Sheet oot e ertet deb o rete e eoe al rne eu v dub Sea ree Ee VENERAT A1 APPENDIX B KinetaMap User Manual esses enne nn terna nnns nn netiis ass ars sane an B1 APPENDIX C Client InterVIews iini oreet tatur cen Fr ca e eoe Ee ORE RE oe e ERE ER o ena ee deo teet tecta C1 APPENDIX D Data from Preliminary Testing usesessseeseeeeseeee nennen enne ni nnns sns nnne D1 APPENDIX E Center of Pressure Data from Preliminary Testing enne E1 APPENDIX F Acceleration Plots from Preliminary Testing eese nennen F1 APPENDIX G Data from Device Verification essen enne enne enne nnne nennen G1 APPENDIX H Code for Device sipia tie ce a a Pene Dn e SIE pne eee E ehe RUD die Nonne Oo n eR ade H1 APPENDIX I Abstract Accepted to Northeast Bioengineering Conference eseeeeeene l1 APPENDIX J Conversions of accekeration data essere nennen nennen nnne nnne J1 APPENDIX K RB APPROVAL
67. alance control Through research and client interviews the team determined design objectives constraints and necessary functions The most important objectives were that the device was easy to use sensitive to balance control accurate and adapted to elderly life According to these objectives the team determined an activity that the device should monitor the signal and sensor used as a monitor the location of the device and the method of analyzing the signal The team chose to monitor the STS activity using an accelerometer placed on the right hip and analyzed the magnitude of the acceleration The team performed preliminary experiments to determine the difference between a balanced and unbalanced STS in terms of acceleration Subjects attached the SparkFun KinetaMap tri axial accelerometer to their right hip and performed 10 STS s with feet shoulder width SW apart balanced and 10 with feet in tandem unbalanced with feet placed on the 14 AMTI force platform Data from the force platform was used to verify a balanced and unbalanced STS and data from the KinetaMap was used to plot the magnitude of acceleration for each trial Time between the positive and negative peak amplitudes of the acceleration plots was calculated and the team found that time between peaks of balanced trials was significantly longer p 0 05 than the time between peaks of the unbalanced trials which was consistent with research Pai amp Patton 1997 Prel
68. amp Jacobsen 2002 Accelerometer based designs have also be applied to patient monitoring systems Patent number 5 515 858 Figure 11 is a monitoring device that resembles a watch and is worn on the wrist The main objective of this device is to monitor physiological conditions such as temperature and pulse but the device also incorporates an accelerometer 4 to detect movements of the hand or wrist in order to monitor the actions of the user The device is able to determine any abnormal acceleration or lack of movement for an extended period of time and automatically transmit an alarm to a surveillance monitor This design also has a call button which the user can press to send an alert to a caregiver and receive assistance in getting up after a fall Myllymake 1996 One benefit of this design is that by monitoring the movements of the user the device could identify if the user loses consciousness after a fall or remains lying on the ground for a period of time and is not moving One drawback of the device is that the device does not specifically detect a fall which is partially addressed with the addition of a call button This issue is partially addressed with the call button but if the user becomes unconscious as a result of a 38 fall and cannot press the button the faller cannot receive help until a period of time goes by and the device recognizes that the user is not moving Figure 11 Patent Number 5 515 858 Wrist held m
69. an unbalanced situation before the fall occurs The vibrations would also enhance the sensitivity of the mechanoreceptors in the ankle which would help convey information regarding the position of the ankle joint more quickly to the central nervous system In turn the vibrating components would increase the ability of the user to recover his or her balance An advantage of the ankle brace design shown in Figure 26 was that it could be worn consistently throughout the day In addition the location would have a limited interference with the user s everyday activities and could also remain out of the public view However it may difficult for 76 the user to attach the wires from the strap to the vibrating components and gyroscope or accelerometer The elderly may have limited grip strength and also decreased vision which would make this task more difficult and hence make the device more difficult to put on and take off In addition assistive devices that aid the elderly in putting on socks already exist in the market indicating that elderly people have trouble putting on and taking off socks Therefore it may also be difficult for an elderly individual to put on the ankle brace One advantage of this design was that the device directly notifies the user of an off balance situation before a fall occurs However small changes of acceleration or position at the ankle joint may be difficult to detect making the device inaccurate ei Stra Wiri
70. ance condition In order to do this operation of the device requires the user to perform the STS rehabilitation activity which is a repetitive daily activity and most sensitive to balance control To strengthen awareness the device is worn on the right hip and monitors acceleration in all planes of movement during the STS activity Finally the device detects an abnormal acceleration indicative of an unbalanced situation and instantly alerts the user of the imbalance Therefore the team answered the four questions as follows 1 What activity should be monitored The STS rehabilitation and daily activity 2 Where should the device be located At the right hip bone 3 What signal and sensor should be used Acceleration will be monitored with a tri axial accelerometer 4 How should the signal be analyzed Analyze time duration of STS looking at the magnitude of the x y and z axis of acceleration Since the most accurate and sensitive means to monitor balance control during STS was by acceleration at the hip the hip mounted accelerometer Preliminary Design 2 was shown to be a feasible conceptual design The team hypothesized that the time duration of the STS would be longer when a person was balanced and shorter during an unbalanced STS To verify the accuracy and repeatability of this method the team performed preliminary testing which is discussed in the following section 4 4 PRELIMINARY EXPERIMENTS The purpose of the device was to detect an u
71. anced condition sees 151 Figure 88 a COP plot for a balanced STS b COP plot for an unbalanced STS 152 12 TABLE OF TABLES Table 1 Existing Fall Detection Technology eene ener 43 Table 2 Devices that Assess Monitor Balance Control seen en ene 54 Table 3 Pairwise Comparison Chart of Marketability Objectives eseseeeeeeen ene 80 Table 4 Pairwise Comparison Chart of Safety Objectives sssssssseeeeeneeenen een eene 81 Table 5 Weighted Objectives Marketability eese een een rennen 82 Table 6 Weighted Objectives Safety sssssessssesseeeeeee eene eene nen nennen rennen rennen 82 Table 7 Numerical Evaluation Matrix Marketability eee enne 83 Table 8 Numerical Evaluation Matrix Safety sees eene 84 Table 9 Numerical Evaluation Matrix Overall Objectives eene 84 Table 10 Subject information sec nni 02e hr rene enr n e ce tee ae e E E RI erp ado 100 Table 11 Results of shoulder width trials eene nem eene 106 Table 12 Results of tandem trials 5 2 rre ea Gebote ede aeta e aie eteastiens 107 Table 13 p values comparing average shoulder width to average tandem time between peaks 107 Table 14 Morphological Chart Electronic functions and means cccccseessssececeeeesssecneaeeeeeeessesenseseeee
72. and Figure 10 are met Single or multiple byte reads writes are supported as shown in Figure 9 With the SDO ALT ADDRESS pin high the 7 b t C address for the device is 0x1D followed by the R W bit This translates to 0x3A for a write and x38 for a read An alternate PC address of 0x53 followed by the R W bit can be chosen by grounding the SDO ALT ADDRESS pin Pin 12 This translates to 0xA6 for a write and OxA7 for a read If other devices are connected to the same PC bus the nominal operating voltage level of these other devices cannot exceed Vooo by more than 0 3 V External pull up resistors Re are necessary for proper FC operation Refer to the UM 10204 PC Bus Specification and User Manual Rev 03 19 June 2007 when selecting pull up resistor values to ensure proper operation Table 10 PC Digital Input Output Voltage Parameter Limit Unit Digital Input Voltage Low Level Input Voltage Ve 025 x Vovo V max High Level Input Voltage Vs 0 75 X Vovo V min Digital Output voltage Low Level Output Voltage Vix v max Limits based on characterization results not production tested 2 Tho Emit gen b onfy for Vi 2 V When Veo 2 V the mit 60 4 V masc THIS START I5 EITHER A RESTART OR A STOP FOLLOWED BY A START NOTIS THE SHADED AREAS REPRESENT WHEN THE DEVICE 55 LISTENING Figure 9 PC Device Addressing Rav 0 Page 10 of 10 ADXL345 FC With CS tied high to Vio the ADXL345 is in PC mo
73. are loaded on the KinetaMap the initialization process takes place The initialization process largely depends on the loaded parameters If gps logging or communication is enabled than the GPS module is initialized with the proper parameters if neither of these options are selected then the GPS module is left off The Bluetooth module is only turned on and initialized if the config menu setting or the send gps to bt parameters are enabled The accelerometer is initialized to the selected range and then the firmware configures the internal timer for readings based on the accel frequency parameter The Real Time Clock used to timestamp each log is initialized to 00 00 00 Once a valid GPS reading has been logged the RTC is updated to the current time indicated in the GPS message the time is only updated to GPS time if a GPS function is enabled though While the KinetaMap is initializing the peripherals the tri color LED on the front of the KinetaMap will be solid Blue Note More information on the parameters and how they affect the initialization can be found in the Configuration File section of this user guide 2009 SparkFun Electronics Inc All Rights Reserved Product features specifications system requirements and availability are subject to change without notice 3 KinetaMap GPS 08275 is a trademark of SparkFun Electronics Inc All other trademarks contained herein are the property of their respective owners KinetaMap UG 090401 om
74. ary Design 3 Ankle Brace Accelerometer Gyroscope LT 4 1 4 Preliminary Design 4 V Tact Belt AM NEEDS ANALYSIS LT 4 2 1 Rank ordering Design Objectives LT FUNCTIONS amp SPECIFICATIONS LT 4 3 1 Choosing the Activity LT 4 3 2 Choosing the Signal and Sensor LT 4 3 3 Choosing the Location LT 4 2 4 Choosing the Method of Analyzing Acceleration Data LT 4 3 5 Final Solution LT PRELIMINARY EXPERIMENTS LT 4 4 1 Materials and method LT 4 4 2 Results of Preliminary Testing LT amp KG 4 4 3 Conclusions of Preliminary Results LT CONCEPTUAL DESIGN LT amp AM FINAL DESIGN AM 5 1 5 2 5 3 5 4 OVERVIEW OF THE FINAL DESIGN AM DESIGN COMPONENTS AM BUILDING THE DEVICE AM PROGRAMMING OF THE DEVICE AM DESIGN VERIFICATION AM DISCUSSION LT AM amp KG DESIGN VALIDATION LT 8 1 PRELIMINARY TESTING WITH ELDERLY SUBJECTS LT 8 1 1 Materials and Methods LT 8 1 2 Determine the Difference Between Balanced and Unbalanced STS LT 8 1 3 Verify the device LT 9 CONCLUSIONS AND FUTURE RECOMMENDATIONS KG GLOSSARY KG APPENDICES LT amp AM ACKNOWLEDGEMENTS The design team would like to thank the following people for their continuous support throughout the course of our Major Qualifying Project Without your advice expertise and volunteered time this project would not have been possible We greatly appreciate all your help Professor Mendelson Professor Gielo Perczak Lisa Wall Barbara Milanese Dave Chris amp Kelly Johnson
75. ased risk of falling Rubenstein 2006 2 3 REDUCING THE NEGATIVE EFFECTS OF FALLS The prevalence and cost of falls in the elderly population has resulted in the development of many rehabilitation methods risk assessment strategies and devices All these options are intended to reduce the negative effects of falls Centers for Disease Control amp Prevention 2009 2 3 1 Improving Balance Control Since many factors affect balance methods of improvement focus on different aspects of balance control In addition to assessment of fall risk exercise has been shown to be the most successful form of intervention for reduced fall risk and improved balance control Stevens 2005 Several exercise methods for improving balance control are physical therapy Tai Chi and the Nintendo Wii Fit balance board and gaming console Physical therapy focuses primarily on increasing muscle strength and flexibility Tai Chi involves controlling the movement of the COM and reducing body sway Mao Hong amp Li 2006 The Wii Fit balance board and gaming console evaluates COP BOS and works on improving reaction time and proprioception Clark Bryant Pua McCrory Bennell amp Hunt 2010 Studies have shown that all three methods are effective for strengthening a person s balance control Stevens 2005 Muscle weakness especially in the lower extremities is one cause of falls in the elderly Studies have shown that muscle weakness negatively affected bala
76. assessment where a person is tested for gait balance and neurological function and reviewing medication allows a physician to individually manage a patient s needs in order to prevent falling For example the physician may refer the patient to a specialist or change a medication However these assessments are recommended for high fall risk patients or those who have already suffered a fall and have gait and balance problems The National Council on the Aging 2005 As result many elderly are not assessed for their fall risk and could be at risk of falling These people may not be assessed for their fall risk before it is too late 65 3 2 OBJECTIVES amp CONSTRAINTS Therefore the team identified two goals of the project to strengthen the user s awareness of their balance control and to proactively force user rehabilitation in order to operate the device Objectives were the criteria that the project and device addressed in order to meet the defined goals The objectives of this project are outlined in Figure 23 66 Accurate Sensitive to balance control Unharmful to user Securely attached Easy to put on take off Wearable Balance Control Indicator Easy to use Easy to operate Manufacturable Easy to repair Portable Comfortable Marketable Durable Aesthetically appealing Adapted to elderly daily life Cost efficient Figure 23 Objectives of the design A constraint was a condition that i
77. at the switch face of the device was against the tabletop and Z axis was perpendicular to gravitational acceleration Plots were created for X Figure 49 Y Figure 50 and Z Figure 51 axis as well as a Magnitude plot Figure 52 to demonstrate the observed change in baseline value at different orientations X Axis Acceleration N lt wn E c 2 pan 2 o o o lt UN m wEHIIITIEIHE gINS TH Id Ig 10 20 30 40 Time sec Figure 49 X axis acceleration during Z axis tilt 110 Acceleration m s 2 Y Axis Acceleration Time sec Figure 50 Y axis acceleration during Z axis tilt N lt v E c 2 9 o ol o lt q Z Axis Acceleration Time sec Figure 51 Z axis acceleration during Z axis tilt 111 Magnitude of Acceleration N lt wn E c 2 i E o o o lt 30 Time sec Figure 52 Magnitude of Acceleration during Z axis Tilt Next x axis rotation tests were performed and the device was positioned with X axis initially positioned 90 degrees askew relative to gravitational acceleration with the same starting position as described for the Z axis test The device was then tipped approximately 45 degrees with the X axis coming more closer to the gravitational acceleration vector pointing downward This position was then held followed by the device being tipped to a position where the X ax
78. b Length 0 79 Width 0 51 Diameter 0 47 Height 0 55 Width 0 2 Length 4 Width 2 5 Height 4 Weight 0 6 Ib 2k Qty 2 1000 Weight 1 Ib Length 4 Width 2 5 Height 2 Weight 2 Ib 131 5 4 PROGRAMMING THE DEVICE As mentioned previously the Arduino Microcontroller is programmed using the Arduino software which is software that utilizes a programming language similar to C C The code was written in an Arduino sketch compiled and uploaded onto the Arduino Microcontroller Board through a USB Since through testing we determined that the time between the maximum positive acceleration peak and minimum negative acceleration peak Figure 76 was significantly different for balanced and unbalanced trials the device was programmed to detect these two peaks and measure the time between them _ time 2 5 4 s Om 3 E os 5 K A y d H 05 g x 1 5 I3 A 2 5 1 2 3 4 5 6 Time s Figure 76 Acceleration curve for the STS activity showing the peaks and time between them 132 If the time between the peaks is greater than the expected time then the user is considered balanced and the device resets However if the time between the peaks is less than the expected time the user is unbalanced and the device buzzes Through testing we determined that the expected time between the peaks for subject 1 was 0 6 seconds and the devic
79. based on the data analysis a Identify wants and needs of stakeholders i e client designers users b Establish and prioritize design objectives c Identify constraints of the design Therefore the team completed four tasks 1 Identify a daily activity for the device to monitor 2 Identify a signal that can be used to monitor balancing control and an accompanying sensor to detect that signal 3 Identify a specific location where the device can be easily worn and the signal can be accurately monitored 69 4 Identify a specific parameter of the signal to analyze and use to distinguish between a balanced and unbalanced condition The following chapter details how the design team completed each of the aforementioned tasks 70 4 ALTERNATIVE DESIGNS The goal of this MQP was to design a wearable device for the elderly that detects an unbalanced situation before a fall occurs The design team identified constraints and objectives by researching and understanding balance control identified advantages and disadvantages of current fall detection and balance control technology and interviewed stakeholders Appendix C Based on the constraints and objectives the team developed three design alternatives A needs analysis was performed and results of the analysis were used to identify necessary functions of the device and develop a conceptual design Preliminary experiments were conducted to determine feasibility of the design This section descr
80. be created and the NMEA GPS messages from the EM 408 module will be saved directly to this file If the parameter is disabled a NMEA file will not be created for the session enable waas parameter Default 1 This parameter will enable or disable the WAAS function on the EM 408 GPS module full scale accel parameter Default 0 The full scale accel parameter sets the range of the LIS302 accelerometer If the parameter is set to 0 the range will be set to 2G mode if the parameter is setto 1 the range will be set to BG mode enable rmc parameter Default 0 This parameter enables or disables RMC messages from the GPS module enable gga parameter Default 1 This parameter enables or disables GGA messages from the GPS module accel frequency Hz parameter Default 10 The accel frequency parameter sets the frequency that acceleration and battery values are read This parameter must be an integer between 1 and 100 100 Hz is the maximum bandwidth of the LIS302 accelerometer NOTE Testing was performed using a 20 Hz read frequency Setting the frequency to higher values should work but may cause adverse affects to the data file logging The Configuration Menu The configuration menu allows the user to manage the SD card perform function checks on the accelerometer and gps module and edit the system parameters over a bluetooth connection using the serial port protocol Instructions on how to connect to the device over bluetooth are
81. both standing and gait Spherical reflective markers are placed on anatomical landmarks on the body and the infrared cameras track the movement of the markers The video footage is analyzed to assess the movement and location of the markers during the activity and can be used to analyze body sway and the movement of the COM Culhane O Connor amp Lyons 2005 While motion analysis can be an 60 effective and beneficial assessment of balance and gait the process is expensive time consuming and can only be performed in a laboratory setting 2 5 5 Advantages and Disadvantages of Current Balance Monitoring Methods The accelerometer pressure transducers and strain gauges and gyroscope are four devices that can be used to measure acceleration pressure and tilt respectively A person s acceleration pressure under the foot and tilt of a person s body has also been shown to relate to balance control Therefore each device was investigated for use in the current balance control indicator The advantages of the accelerometer are that it is most accurate when placed on the trunk which is an accessible location and could make the device easily positioned In addition the accelerometer has shown to be accurate in monitoring balance control during activities such as the sit to stand and has been used to monitor balance control in place of a motion analysis system Pressure transducers and strain gauges could be used to measure pressure under
82. cal psychological and social consequences To avoid such an evolution specific physical therapy must be proposed as soon as possible Morisod amp Coutaz 2007 Pressure Ulcers An area of skin that breaks down when the patient remains in one position for too long without shifting weight 157 Proprioceptive System The sense of the orientation of one s limbs in space Anissimov 2010 Mechanoreceptors A sensory receptor that responds to mechanical pressure or distortion Mediolateral Movement movement along the frontal plane of a body i e lateral sway Gielo Perczak 2009 Microelectromechanical Systems MEMS the integration of mechanical elements sensors actuators and electronics on a common silicon substrate through microfabrication technology micromechanical components are fabricated using compatible micromachining processes that selectively etch away parts of the silicon wafer or add new structural layers to form the mechanical and electromechanical devices Reithel 2010 Somatosensory System A diverse sensory system comprising the receptors and processing centres to produce the sensory modalities such as touch temperature proprioception body position and nociception pain Boulpaep amp Boron 2003 Tai Chi A meditative form of gentle stretching and postural changes in a slow and flowing manner Rose 2005 Vestibular System A sensory system in mammals that determines body position with re
83. cceleration vs time data during Shoulder Width STS with device 138 In addition we used the COP data as another method to confirm whether the trial was balanced The COP data from the shoulder width STS trial is shown in black in Figure 81 The COP was compact and had little variation which was also representative of a balanced trial Y of FP in Xof FP in Figure 81 COP graph for shoulder width black and tandem green STS trial An example of a tandem STS test with the device is shown in Figure 82 The device buzzed after this trial and visual analysis also suggested that the subject was unbalanced Figure 82 Tandem STS with Device 139 In order to confirm whether the device successfully identified an unbalanced trial we first analyzed the KinetaMap data The acceleration vs time graph that we generated for this tandem trial is shown in Figure 83 The maximum and minimum acceleration peaks were identified and circled in blue We determined that the time between these peaks was 0 45 seconds This time was less than the expected time of 0 6 seconds which was representative of an unbalanced trial Tandem Acceleration m s Time s Figure 83 KinetaMap acceleration vs time data during Tandem STS with device Similar to the shoulder width trials we verified whether the trial was unbalanced by analyzing the COP data from the force platform The COP data for the tandem trial is shown in green in
84. ccurs PROACTIVELY FORCE USER REHABILITATION Require rehabilitation activity for operation of device Utilize repetitive daily activity Elderly Disabilities Available Testing Equipment Affordability Figure 28 Outline of goals red functions blue specifications green and constraints grey of the design To determine if the waist mounted accelerometer met the necessary functions and specifications Figure 28 the team posed four questions 1 What activity should be monitored 2 Where should the device be located 3 What signal and sensor should be used 4 How should the signal be analyzed The following sections describe the team s process of answering these questions to further identify specifications of the design These sections also verify that the waist mounted 86 accelerometer would be able to fulfill the specifications making it the best choice for the final design 4 3 1 Choosing the Activity The two goals of the device were to proactively force user rehabilitation and strengthen the user s awareness of their balance condition In order to do this operation of the device should require the user to perform a rehabilitation activity Specifically the rehabilitation activity should be a daily activity so that operation of the device is integrated into the daily life of the user To strengthen awareness the device should detect and notify the user of an unbalanced situation before a fall
85. ct their posture and balance before they actually fall The device would also be able to detect that a fall has occurred Vibration Tactors Velcro Belt MP3 Holder SHIMMER sensors Figure 27 Preliminary Design 4 V Tact Belt 78 4 2 NEEDS ANALYSIS In order to evaluate the design alternatives the team first determined the requirements of the design Designers discussed with all stakeholders the needs of the device or the attributes it must have and wants of the device attributes that a stakeholder would like to have but may not be possible given other constraints Gielo Perczak 2009 Based on the needs and wants of the stakeholders objectives were rank ordered necessary functions were determined and specifications were outlined All design alternatives were then assessed according to how well they met the client s wants and needs and fulfilled the necessary functions of the device This sections describes the specific tools used to rank order the objectives and establish necessary functions and specifications of the device 4 2 1 Rank ordering Design Objectives The team prioritized design objectives by rank ordering them using pairwise comparison charts PCC A PCC is a tool that compares each objective against every one of the other objectives in order to rank order them according to their importance to the final design In the following PCCs a 1 means that the objective in the row was more important than the objec
86. ction force are different the body will sway in the anterior posterior direction Le forward or backward and medial lateral direction i e side to side While standing the body continuously attempts to balance these moments to reduce sway Winter Patla Prince Ishac amp Gielo Perczak 1998 The concepts of moments COG BOS and COP as they relate to balance will be described as they relate to human balanced control and through an example of balancing a pencil 2 1 2 Human Balance Control and the Inverted Pendulum Model Maintaining balance in a human is difficult because the COG is located at a distance 2 3 up from the point of the reaction force or the feet Depending on the position of the COG in relation to the point of the reaction force the body will sway forward or backward The body is able to react to forward and backward sway in order to recover balance and prevent falling As a result the body is continuously swaying forward and backward or in the anterior posterior A P direction This phenomenon and factors that determine how the body sways is referred to as the inverted pendulum model shown in Figure 2 Winter Patla Prince Ishac amp Gielo Perczak 1998 20 Figure 2 The body modeled as an inverted pendulum adapted from Winter Patla Prince Ishac amp Gielo Perczak 1998 In Figure 2 the body s COG labeled W acts downward and an equal and opposite reaction force R acts upward Force R represents
87. d Worcester MA 01609 Re IRB Expedited Review Approval 10 067 Fall Detection Device Dear Dr Gielo Perczak The WPI Institutional Review Committee IRB approves the above referenced research activity having conducted an expedited review according to the Code of Federal Regulations 46 Consistent with CFR 46 116 regarding the general requirements for informed consent we remind you to only use the attached stamped approved consent form and to give a copy of the signed consent form to your subjects You are also required to store the signed consent forms in a secure location and retain them for a period of at least three years following the conclusion of your study You may also convert the completed consent forms into electronic documents pdf format and forward them to the IRB Secretary for electronic storage The period covered by this approval is 8 April 2010 until 7 April 2011 unless terminated sooner in writing by yourself or the WPI IRB Amendments or changes to the research that might alter this specific approval must be submitted to the WPI IRB for review and may require a full IRB application in order for the research to continue Please contact the undersigned if you have any questions about the terms of this approval Sincerely Cooke Qvssou IL Kent Rissmiller WPI IRB Chair K4 Participant Email Dear j We are recruiting subjects to participate in a research study for our Major Qualifying Project
88. d allow the user to perform the STS activity many times before recharging the device There are several ways the battery pack can be charged including using the USB plug on the Arduino to connect a computer a mini USB to connect the battery pack to a computer or a wall adapter plug Huynh 2009 http www liquidware com shop show ATG Arduino to Go Figure 70 Rechargeable battery pack left and battery attached to Arduino right Allum 1999 Arduino 2009 The user charges our balance control indicator using a 9V wall adapter plug This 2 5mm wall adapter plug with a positive center Figure 71 connects to the power connector on the Arduino board The Alternative Current AC from the wall is converted to 9V Direct Current DC that is used to charge the battery pack that powers our device Maker Media 2009 Figure 71 9V Wall Adapter Plug Maker Media 2009 125 A slide switch like the one shown in Figure 72 is used to turn the device on and off Digikey 2009 This type of switch was selected because we want an elderly user to be able to easily turn the device on and off Figure 72 On Off Slide Switch Digikey 2009 Lastly a ProtoShield board Figure 73 was used as a PCB with attached header pins that connect directly to the header pins on the Arduino Sparkfun Electronics 2009 The board was used to attach the accelerometer buzzer two 2k resistors and a 100Q resistor both of which are needed to reduc
89. d that tandem trials represented an unbalanced situation and shoulder width trials represented an unbalanced situation Since the time between peaks was significantly different p lt 0 05 for all subjects who 107 performed 10 tandem and 10 shoulder width STS s the team concluded that a specific time range corresponds to a balanced STS and a specific time range corresponds to an unbalanced STS The preliminary data showed similar results between subjects of the same gender and age however not enough data were collected to show that these similarities were significant Therefore the time ranges were specific to the individual and the team designed specifications according to the data of one particular subject The following section discusses how the team conceptualized specifications and made decisions regarding the final design 4 5 Conceptual design The purpose of the device is to detect an unbalanced STS in terms of acceleration and notify the user when an unbalanced STS occurs Based on preliminary results the time between the positive and negative peaks of these amplitudes was significantly longer in shoulder width trials than in tandem trials Therefore the team concluded that a specific time range corresponded to a balanced STS and a specific time range corresponded to an unbalanced STS These time ranges were specific to the individual Based on this finding and necessary functions of the device the team brainstormed the steps that
90. ddress x31 an electrostatic force is exerted on the mechanical sensor This electrostatic force moves the mechanical sensing element in the same manner as acceleration and it is additive to the acceleration experienced by the device This added electrostatic force results in an output change in the x y and z axes Because the electrostatic force is proportional to V5 the output change varies with Vs The self test feature of the ADXL345 also exhibits a bimodal behavior that depends on which phase of the clock self test is enabled However the limits shown in Table 1 and Table 12 to Table 15 allowable voltage range Use of the self test feature at data rates less than 100 Hz may yield values outside these limits Therefore the part should be placed into a data rate of 100 Hz or greater when using self test Table 12 Self Test Output in LSB for 2 g Full Resolution Rav 0 Page 13 of 13 REGISTER MAP Ox01 to Ox01C ox1D OXIE Ox1F FIFO STATUS Rav 0 Page 14 of 14 Inactivtyy threshold Inactivity time Axis enable control for activity and inactivity detection Free faii threshold Free fali time Axis control for tap doubie tap Source of tap doubie tap Data rate and power mode control Power saving features control Interrupt enable control Interrupt mapping control Source of Interrupts Data format control X Axts Data 0 X Axis Data 1 Y Axis Data 0 Y Axis Data 1 Z Axis Data 0 Z Axis
91. de requiring a simple 2 wire connection as shown in Figure 8 The ADXL345 conforms to the UM10204 PC Bus Specification amd User Manual Rev 03 19 June 2007 available from NXP Semiconductor It supports standard 100 kHz and fast 400 kHz data transfer modes if the timing parameters given in Table 11 and Figure 10 are met Single or multiple byte reads writes are as shown in Figure 9 With the SDO ALT ADDRESS pin high the 7 bit C address for the device is x1D followed by the R W bit This translates to 0x3A for a write and 0x38 for a read An alternate PC address of 0x53 followed by the R W bit can be chosen by grounding the SDO ALT ADDRESS pin Pin 12 This translates to OxA6 for a write and OxA7 for a read If other devices are connected to the same C bus the nominal operating voltage level af these other devices cannot exceed Vooo by more than 0 3 V External pull up resistors Ra are necessary for proper PC operation Refer to the UM 10204 PC Bus Specification and User Manual Rev 03 19 June 2007 when selecting pull up resistor values to ensure proper operation Table 10 IC Digital Input Output Voltage Digital Input Voltage Low Level Input Voltage Ve Limits based on cha results not production tasted Tho Emit given b onfy for Vsus 2 V Waan V 2 V the limit 504 V marc THIS START 15 EITHER A RESTART OR A STOP FOLLOWED BY A START NOTIS TME SHADED AREAS REPRESENT WHEN THE DEVICE 75 LISTENI
92. device design are sensors that can measure orientation based on the principles of angular momentum Omega Engineering Inc 2003 55 Body of interest Damper Spring Proof mass qx i n Hi 1 i Figure 20 MEMS accelerometer and muscle behavior model Fung 1993 There are a number of functions that an accelerometer can have one of which is tilt This is due to the accelerometer s inherent ability to detect gravitational acceleration on each sensing axis relative to its rotational position A MEMs accelerometer can be modeled after a cantilever beam with a proof mass in a gas damped chamber between two capacitive plates As gravity acts on the beam it displaces from its neutral position toward the lower capacitive plate This would be read the same as if a force was displacing the sensor upwards accelerating at the equivalent local gravity Figure 20 This is a useful application for devices that are intended to make use of this rotational property of the accelerometer However in applications where the accelerometer is intended to measure motion components as well as rotational components of movement or only motion components it can become difficult to discern which portions of the signal are resultant of tilt and which are of motion The key to understanding the effects of different components of motion on the signal can be understood from observing the output on the timescale A rotation of the accelerometer will resul
93. ds of sensors and alarms to watch people get out of bed Do you think this device would be more useful than anything currently used I think it would be as long as they re well aware for the old people I think it would be good for people who are recuperating you know young people like me who have never had to deal with an issue like this and then something happens and they need a device like this just to remind them that oh maybe I need to sit for a while maybe I need to slow down until I get my balance Sometimes people are medicated for pain to when they re recuperating and they re not as steady as they think they are and maybe it would be useful in that respect postoperatively or when they re home recuperating Do you think if they re forgetful about other things they d be forgetful about this I think they d forget to put it on or if they re really confused they d forget what the buzz meant But I think if they had a person watching them who could remember to put it on overnight so that it could buzz in the morning if it was comfortable enough to be worn all the time C 4 Do you think they d remember to turn it on if they could wear it all day and they sat down would they remember to turn it on before they sat up It depends on how alert they are their minds You know myself I wouldn t forget and if I were sick in bed I d be thankful for it you know my sister people don t want to fall
94. e negative effects of falls particularly on the elderly population The Centers for Disease Control and Prevention CDC have administered various fall prevention and education programs aimed at reducing the occurrence of falls in the elderly These programs address a few of the many risk factors associated with elderly falls For example two risk factors that cause falls are decreased balance control and strength in the elderly As a result two studies funded by the CDC and conducted over a three year time span utilized education and exercise programs intended to improve balance and increase strength in the elderly participants thereby reducing falls However neither study produced a significant 16 reduction in falls Centers for Disease Control amp Prevention 2009 On the other hand previous research showed that Tai Chi exercise can improve balance and decrease falls among the elderly but researchers do not know if the general elderly community can adopt this exercise into daily life Rose 2005 According to this research regular exercise that the elderly can readily integrate into their daily life is one way of reducing the risk falling Stevens 2005 The CDC has also identified a need to increase elderly self efficacy and sense of balance control in relation to fall risk in order to prevent elderly falls Centers for Disease Control amp Prevention 2009 Therefore the MQP team set out to design a device that would enable elderly
95. e PERS is that it is user activated so if the user cannot reach the device is unconscious or is in any other way unable to press the button the device is not beneficial Figure 15 Philips LifeLine Personal Emergency Response System Koninklijke Philips Electronics 2010 42 Table 1 Existing Fall Detection Technology Sensor for I Device i Location Pros amp Cons Fall Detection 3 axis Strap is worn on Pros myHalo Fall Monitor http www halomonitoring com accelerometer upper torso near sternum http www halomonitoring com Clip can be worn on belt Automatically detects the fall Alerts call center that user has fallen who can then alert emergency response Detects vital signs Cons Only for home use Multiple components and set up required Chest strap could be difficult for the elderly to put on Does not detect fall before it occurs Brick House Alert Fall Monitoring and Panic Button System http www brickhousealert com 7 Ww M 44 Fall Detector 3 axis accelerometer Clips onto side of hip http www brickhousealert com Pros Automatically detects the fall Alerts central monitoring station who can alert emergency response Waterproof can wear in shower Cons Only for home use Does not detect fall before it Occurs Set up required Philips Lifeline amp other Call Buttons Q A http www lifelinesys com No sensor f
96. e beginning of the code are used One of the defined parameters is 134 StartMinA which is the threshold acceleration for subject 1 this acceleration was determined through testing to be 0 5 m s Another important parameter is fg the expected time between the maximum and minimum peaks which for subject 1 is 0 6 seconds corresponding to the value 0 6 1000 milliseconds Initially 44 time corresponding to the max acceleration tap time corresponding the minimum acceleration A4 maximum acceleration and Ap minimum acceleration are set equal to zero TestStarted TestFinished and RingBuzzer were set to false The test is started and TestStarted is true when A the current acceleration is greater than StartMinA The time the test was started was defined as TimeStart and current time was defined as T The expected test length was defined at TestLength In this section of code f statements were used A moving time window was used to search for the maximum acceleration peak A4 and minimum acceleration peak Ap Therefore throughout the duration of the test if A was greater than the value stored as the maximum acceleration A4 and then A4 was now equal to A Similarly if A was less than the value stored as the minimum acceleration Ap then Ap was now equal to A The values 144 for and tap correspond the time at which A and Ap respectively occurred and these values change as A4 and Ap change TestFinished was set to true when T minus TimeStar
97. e benefit to other is this will help with our project and lead to the design of an early fall detection device for the elderly community Record keeping and confidentiality Records of your participation in this study will be held confidential so far as permitted by law However the study investigators the sponsor ar it s designee and under certain circumstances the Worcester Polytechnic Institute Institutional Review Board WP IRB will be able to inspect and have access to confidential data that identify you by name Any publication or presentation of the data will nat identify you Compensation or treatment in the event of injury in the unlikely event of physical injury resulting from participation in the research you understand that medical treatment may be available from WPI including first aid emergency care and that your insurance carrier may be billed for the cost of such treatment No compensation for medical care can be provided by WRI You further understand that making such medical care available or providing it does nat imply that such injury is the fault of the investigators You do not give up any of your legal rights by signing this statement For mare information about this research or about the rights of research participants or in case of research related injury contact Krystyna Gielo Perczak BME Department WPI 100 Institute Road Worcester MA Tel 508 331 5100 You may also contact the Chair of the WP Instituti
98. e during Perturbation The COP of the inverted pendulum model has also been found to behave sinusoidally as shown in Figure 3 Winter Patla Prince Ishac amp Gielo Perczak 1998 When the body sways forward the COP lies outside the COG and has a positive amplitude of acceleration as shown on the graph When the body sways backward the COP lies behind the COG and has a negative amplitude of acceleration as shown on the graph The amplitude of the COP is largest at first and continually decreases This shows that upon perturbation a person is accelerating the quickest and sways most As the body reacts to maintain balance the body sways less and the acceleration decreases Shumway Cook amp Woollacott 2007 22 wn N o Ss E c 2 P c P 2 o o o lt Time sec Figure 3 Sinusoidal Behavior of COP A Working Model Simulation Gielo Percak 2010 Depending on the acceleration of the perturbation the body uses different methods to maintain balance Balance control and anterior posterior stabilization strategies are utilized to prevent or attempt to prevent a person from falling There are three major anterior posterior stabilization strategies the ankle hip and stepping strategies as shown in Figure 4a The ankle strategy shown by the number 1 in Figure 4a only provides a small range of motion and is therefore used when only a small adjustment is needed to maintain balance The hip strategy number 2 in Figu
99. e eA KinetaMap User Guide 2009 04 01 Parameter Name Default Value config_menu O Off log gps 1 On log accel 1 On send gps to bt O Off log batt 1 On log nmea 1 On enable waas 1 On full scale acce O Off enable rmc O Off enable gga 1 On accel frequency 10 Hz After the peripherals are initialized the device is almost ready to start it s main routine Before it can begin though it needs to create files to log too if logging is enabled There are two files that are created in the default settings the data file and the NMEA file If any type of logging is enabled than a data file is created The data file consists of the time the acceleration values the battery voltage and some basic GPS information Also if NMEA logging is enabled an NMEA file is created The NMEA file saves the unaltered GPS messages from the EM408 module All of the data logged by the KinetaMap will be logged to these two files The KinetaMap will create a new set of files each time the device is turned on The names of the created files follow a specific format The data files are all named KinetaMapXXX csv where XXX is replaced by an integer between 0 and 255 When the KinetaMap creates a file it will start with the name KinetaMap000 csv If this file already exists then the KinetaMap will increment the number following KinetaMap It will do this until if finds a filename that doesn
100. e fall indicators and emergency response systems are a capable means of ensuring that elderly will receive proper attention after a fall has occurred but neither addresses the need for preventative measures that can minimize or eliminate the risk of fall Whereas the balance control indicator device will allow the elderly to strengthen their balance control potentially eliminating the risk of fall and resultant injury the fall detection device will simply be able to detect the fall once it has happened and reduce the risk of long lie injuries and increase the survival rate PERS systems may be less effective as they require that the user be conscious and capable of activating a call signal in order to receive help The primary advantage present in the balance control indicator device is the ability to reduce the risk of fall through preventative exercises with instant feedback 143 The objectives for the device specifically strengthening of the user s awareness of their balance condition and proactive forcing of user rehabilitation were both met in the design of the balance control indicator device The Berg balance test indicates that the STS activity is a valid indication of the user s overall balance control and therefore by creating a device that provides understandable feedback on the balance condition with which a user performs this activity the device has succeeded in assisting the user in becoming aware of their balance control conditio
101. e from this data will be dependent on the sensitivity of the sensing axis i e 2g accelerometer on a 100 duty cycle will be measuring either 2g or 2g 50 duty cycle will be recording either 1g or 1g This technique of pulse width modulation is necessary because of the Boolean nature of digital electronics as a device can either be fully on or fully off and an accelerometer requires intermediary measurements 57 Accelerometers have been used frequently in motion analysis research studies for both walking and balance control Janssen W G Bussmann Horemans amp Stam 2005 According to a study by Janssen et al The results of the study indicate that accelerometer is able to provide a sensitive measure of balance during the sit to stand movement Accelerometery offers the benefit of low cost and portability An important advantage of accelerometery therefore is that it allows measurement outside of a movement laboratory Janssen Kulcu Horemans Stam amp Bussmann 2008 Accelerometer signals have also been shown to contain information on kinematic events that will enable us to define time markers to describe the phasing and duration of the sit to stand movement without the use of a gait laboratory Janssen W G Bussmann Horemans amp Stam 2005 Research has shown that the acceleration of a person s COM can determine how close they are to their stability limit For example a quick acceleration would cause a
102. e platform Preliminary studies were conducted on 5 subjects 1 male and 4 females all ages 20 21 years To determine their weight subjects stood on the force platform with feet shoulder width SW apart for 10 seconds Next the subjects sat in the chair with their feet positioned on the force platform NetForce data collection was started and the KinetaMap was turned on After 10 seconds subjects were asked to rise from the chair at a comfortable speed with their feet SW apart control The was tumed off 10 seconds after the activity was completed This process was repeated for 10 separate trial The process was also repeated for 10 trials where subjects stood up with their feet in tandem T variable During the STS trials subjects were asked to rate their comfort level on a scale from 1 5 where 1 corresponds to completely off balance and 5 indicates balance The NetForce data fles were imported into BioAnalysis software and plots of center of pressure COP were obtained The COP data were used to determine if the subjects were off balance and quantitatively represent balance control Raw acceleration data from the x y and z components of the KinetaMap were saved in Microsoft Excel for further analysis Acceleration was converted into m s and the magnitude of the acceleration profile for each trial was plotted The maximum acceleration a minimum acceleration a time between these points t and threshold acceleration a be
103. e the power from the Arduino to the accelerometer and buzzer Figure 73 ProtoShield Board Sparkfun Electronics 2009 5 3 BUILDING THE DEVICE A schematic diagram of the device is shown in Figure 74 where the ADXL345 accelerometer is shown in red Arduino Duemilanove Microcontroller Board is shown in blue 126 and the CEM1203 buzzer is shown in green Arduino Duemilanove Microcontroller Board CEM 1203 Buzzer ATmega168 Microcontroller Figure 74 Schematic of Device The Arduino Duemilanove Microcontroller supports both SPI and PC communication However the Arduino programming language does not have functions for SPI communication therefore we connected the ADXL345 accelerometer to the Arduino using IC Inter Integrated Circuit PC wiring connections between the Arduino and ADXL345 accelerometer are shown in Figure 74 The PC pins on the Arduino are Analog pins 4 and 5 where analog pin 4 is wired to the Serial Data SDA pin on the accelerometer and analog pin 5 is connected to Serial Clock SCL pin on the accelerometer Arduino 2009 Pull up resistors the 2k resistors shown in Figure 74 were suggested in the data sheet for the accelerometer are needed to reduce the voltage from the Arduino because the ADXL345 cannot handle more than 3 6V Analog Devices 2009 The SDO pin on the ADXL345 was connected to ground on the Arduino The 3V3 and CS pins on the accelerometer were connected to the 3V3 pin on the Ard
104. e user s feet This device can be connected to a computer and a medical professional is able to analyze the data to determine and monitor a user s balance control 52 Trafton 2008 One major disadvantage of the iShoe is that the data needs to be brought to a doctor s office and interpreted by a medical professional Another disadvantage is the device is not proactive and does not provide instant feedback to the user This device is only beneficial when the user is wearing shoes and putting on the device could be difficult for an elderly user The Wii Fit Balance Board and gaming system provides an interactive way for user s to assess their balance control The balance board contains pressure sensors that display the user s COP on a television screen through a video game interface The Wii Fit game has several games for user s to play in order to improve their balance control Nintendo 2009 While this device provides visual feedback to the user and an opportunity to work on their balance control this device does not directly assess the user s risk of falling This device is also targeted for the younger population and the video games are not catered to an elderly user While both of these devices are able to assess and monitor balance control neither device is proactive or directly assesses the user s risk of falling The devices are also not designed to be easy for an elderly person to use them 53 Table 2 Devic
105. e was programmed specifically for this expected time The block diagram of the final design is shown in Figure 77 Starts peak r i n when eee records records signal find max ie acceleration maximum minimum and min acceleration exceeds seen threshold 0 5 m s peak peak min peaks Figure 77 Block Diagram of Final Design The program consisted of four main sections 1 Defining and initializing the C communication between the Arduino and ADXL345 2 Reading and writing the data from the accelerometer 3 Processing the accelerometer data 4 Sounding the buzzer Several of the parameters are defined as constants in the beginning of the program using the define function The define function was used to assign a constant value to a variable so that whenever this variable is used throughout the code the compiler replaced the variable with the defined constant Arduino Reference 2009 The define function was used specifically to define the expected time between peaks the threshold acceleration expected length of time for 133 the entire STS activity and the duration time for the buzzer This feature made the device easily programmable for different users The first section of the program was used to define the address of the accelerometer and initialize the I2C communication between the ADXL345 and the Arduino and was adapted from an example code on the Arduino Forum Arduino Forum 2009 The first line
106. ed acceleration and detected and notified the user of an unbalanced situation abnormal acceleration before a 95 fall occurred Therefore the team identified the most accurate repeatable method of analyzing acceleration of the COM that clearly illustrated the difference between a balanced and unbalanced STS Figure 38 Sensitive to balance control A Repeatable Figure 38 Factors method of analysis needs to fulfill Through research the team found that the signal of the COM was directly related to its acceleration in the A P and M L directions Winter D 1995 In particular the magnitude and frequency of this signal and reaction time of the subject were the most important in classifying the balance control Winter D 1995 The team identified 5 potential methods for analyzing frequency magnitude and time duration of acceleration Fast Fourier Transform FFT acceleration on the X axis acceleration on the Y axis acceleration on the Z axis and route sum of squares of the X Y Z axis FFT is used to enhance or remove periodic noise in a signal and yields the power of the signal as a function of the frequency Although FFT is an accurate means of assessing balance control research showed that analysis of a short record can result in erroneously high means and median frequencies Winter D 1995 Since the STS activity is short only lasting approximately 3 seconds we eliminated the FFT method of analysis 96 Resea
107. eee te ra perenne ete dont Fen ea eeu t etn ee veo e rade ee edenda e rns Doe hen K1 AUTHORSHIP PAGE ABSTRACT LT EXECUTIVE SUMMARY 1 INTRODUCTION LT 2 LITERATURE REVIEW 2 1 UNDERSTANDING BALANCE CONTROL LT 2 2 2 3 2 4 2 5 2 1 1 Basics of Balance LT 2 1 2 Human Balance Control and the Inverted Pendulum Model LT 2 1 3 Maintaining Balance during Perturbation LT amp AM 2 1 4 Example of Balance Control Balancing a Pencil LT RISK FACTORS OF FALLING AM REDUCING THE NEGATIVE EFFECTS OF FALLS AM 2 3 1 Improving Balance Control AM 2 3 2 Fall Detection Patents AM 2 3 3 Current Fall Detection Technology AM 2 3 4 Limitations of Current Technology AM METHODS FOR ASSESSING BALANCE CONTROL AM 2 4 1 Vibrotactile Based Patents AM 2 4 2 Current Devices for Assessing Balance Control AM amp LT PROACTIVELY MONITORING BALANCE CONTROL AM 2 5 1 Accelerometry KG AM LT 2 5 20 Gyroscopes KG amp LT 2 5 3 Strain Gauges and Pressure Transducers KG amp LT 2 5 4 Motion Analysis AM 2 5 5 Advantages and Disadvantages of Current Balance Monitoring Methods LT 3 PROJECT STRATEGY LT 3 1 3 2 CLARIFYING THE ORIGINAL PROBLEM STATEMENT LT OBJECTIVES amp CONSTRAINTS LT 3 3 3 4 REVISED PROBLEM STATEMENT LT PROJECT APPROACH LT ALTERNATIVE DESIGNS LT 4 1 4 2 4 3 4 4 4 5 PRELIMINARY DESIGNS LT 4 1 1 Preliminary Design 1 Shoe Insole KG 4 1 2 Preliminary Design 2 Waist mounted Accelerometer KG 4 3 Prelimin
108. eir weight was collected and saved in NetForce software The subject then sat in the chair and NetForce data collection was started The subject then turned on the KinetaMap and waited for the KinetaMap s LED to start blinking blue signifying that the accelerometer had started collecting data Ten seconds after the LED started blinking blue the subject stood up from the seated position Ten seconds after the subject reached a balanced standing position the subject turned the KinetaMap off to stop data collection This was repeated 10 times with the subject s feet positioned shoulder width apart while performing the STS Figure 43 and 10 times with the subject s feet positioned in tandem Figure 44 Subjects who conducted preliminary testing in the home setting performed a minimum of 5 tandem and shoulder width STS trials due to time 101 constraints of the subjects schedules After each STS trial the subject rated their comfort on a scale from 1 5 1 being completely unbalanced 5 being completely balanced Figure 43 Shoulder width foot position representing a balanced condition Figure 44 Tandem foot position representing an unbalanced condition The KinetaMap collected data at 20Hz and logged each trial in a Microsoft Excel document in terms of time and the X Y and Z components of acceleration The X Y and Z components of acceleration were converted into m s using the ADXL2345 Tri Axial accelerometer data sheet A
109. el Octave MatLab or your favorite data analysis package KinetaMap includes a Bluetooth module that supports SPP serial port profile This allows a computer to wirelessly connect to and download the log files With a range of 100m a computer is capable of querying a KinetaMap located within a vehicle outside A computer may also wirelessly modify the configuration and settings internally saved within KinetaMap KinetaMap has also been tested with Bluetooth and Google Maps enabled phones KinetaMap allows precise location and map directions on your phone rr Specifications Dimensions 3 7x2 5x1 1 28x63x94mm e Weight 3 60z 103g e Battery 3 7V 1100mAH LiPo Battery Current Draw 140mA Bluetooth Active 110mA Bluetooth Inactive 4mA Sleep Mode 2009 Sparktun Electronics Inc All Rights Reserved Prodi ct feat res specifications system requirerrerts and availability are subject to change without notice 4 KinetaMap CI 082 5 is a trademark ot SparkFun Electronics Inc All otrer trademarks contained Ferein are tre property ot their respective owr ers m KinetaMap UG 090401 AO C ge sparkfun LECTRONICS KinetaMap User Guide 2009 04 01 the side of the device is the miniUSB connector Plugging a USB cable into a powered KinetaMap from a computer will open a Mass Storage Device The Mass Storage Device allows access to the log files and co
110. en Office Calc The data file created by KinetaMap will always have the following format regardless of what parameters are enabled or disabled UTC X Y Z Batt Fix Lat Lat Dir Long Long Dir Altitude Type 0 2 0 16 5014 O0 1 0 17 5014 1 2 0 16 5014 1 2 0 16 5020 170937 44 1 40039N 10512 6W 1485 GGA 170937 2 0 16 5020 170937 2 0 16 5014 Each log consists of one row of data if there are no values to be logged for a specific field that field is left blank in the log The UTC field is the timestamp The timestamp is initially set to 00 00 00 HH MM SS however if GPS functions are enabled this value is updated to the current time once a valid GPS message is received If GPS functions are not enabled the UTC will measure the duration of the KinetaMap session The X Y and Z acceleration fields are the acceleration values retrieved from the LIS302 triple axis accelerometer The acceleration values can fall anywhere in the range of 127 to 127 depending on the current acceleration The LIS302 accelerometer has two modes 2G mode and 8G mode To find the G force of the accelerometer the values in these fields must be scaled with respect to the selected mode Instructions on how to scale these values can be found in the data sheet of the LIS302 The Batt Volt field represents the current battery voltage This reading is represented in mV The ba
111. en 77 Figure 27 Preliminary Design 4 V Tact Belt eene ren rennen 78 10 Figure 28 Outline of goals red functions blue specifications green and constraints grey of the ap 86 Figure 29 Factors needed in choosing an activity essen en ener 88 Figure 30 Berg Balance Test American Academy of Health and Fitness 2010 sss 89 Figure 31 Choosing the Device Activity essesseeseeseeeeeeeeen nennen nennen eene rennen enne 90 Figure 32 Factors signal and sensor need to fulfill esee eene enne 91 Figure 33 Signals to monitor balance control pressure acceleration tilt sse 92 Figure 34 Choosing the Device Signal and Sensor eene 93 Figure 35 Factors location of device needs to fulfill eseeeeeeeen emen 94 Figure 36 Possible locations of device http www ehow com how 4629008 draw person standime html Tanneriin i tete i oderit titt chueta decade dea tuve eq Pe eee Evo coast qaa a TER 95 Figure 37 Choosing the Device Location drieen deit ii iaiia eieaa naiita ener rennen rennen 95 Figure 38 Factors method of analysis needs to fulfill enne 96 Figure 39 Choosing the Method of Analysis eese eene 97 Figure 40 KinetaMap Triple Axis Accelerometer and Data Logger SparkFun Electronics 2009 100 Figure 41 Chair and force platform experimental set up eese
112. ent contains a triple axis accelerometer that automatically detects a fall after the fall has occurred Upon detection of a fall the MyHalo system connects to the MyHalo Operating Center who then calls a caretaker The MyHalo System also sends messages to caretakers via e mail text message or a personal web page The sensing component is placed no more than 2 inches below the sternum as shown in Figure 14 The device can also be clipped on the pants of the user however in this position other vital signs e g blood pressure heart rate cannot be obtained Halo Monitoring 2009 Disadvantages of the MyHalo Monitoring System are that it does not detect a fall before the fall occurs it is only for use inside the home or immediate area of the home e g yard 41 garage and that wrapping the strap around the chest to put the device on and take it off may be difficult Figure 14 Correct positioning of MyHalo chest strap and device adapted from Halo Monitoring 2009 Philips Lifeline Personal Emergency Response System PERS is another method of fall protection This device is contained within an apparatus such as a watch or necklace The device contains a button and when the buttons is pressed the system contacts an operator who calls a caretaker or sends emergency response Figure 15 The button is pressed by the user when in need of assistance and cannot reach help Koninklijke Philips Electronics 2009 The main disadvantage of th
113. es that Assess Monitor Balance Control iShoe 2008 i shoe 0716 html Pressure sensors Insert for a shoe http web mit edu newsoffi ce 2008 i shoe 0716 html Pros Able to detect abnormal pressure patterns Bluetooth enabled to transfer data to a doctor Monitors balance control Cons Does not detect a fall before it occurs User must be wearing a shoe to use the device Data must be interpreted by a medical professional Wii Fit Balance Board http www nintendo com wii what accessories balanceboa Center of Balance COB Results Oh Your center of balance is slightly to the right http wiifit com body test body control Pressure sensors Board that users stand on ME m http wiifit com body test body control Pros Videogame to improve balance control User awareness of balance control Measure COP Cheaper than a force platform Cons Games are not catered to the elderly Many steps for operating the device Games are not related to repetitive daily activities Although there are currently devices on the market to alert emergency response detect a fall after it occurs and assess balance control there is a need for a device that proactively monitors balance control provides instant feedback and notifies the user of their risk of falling 54 2 5 PROACTIVELY MONITORING BALANCE CONTROL An important step in preven
114. fications system requirements and availability are subject to change without notice 4 KinetaMap GPS 08275 is a trademark of SparkFun Electronics Inc All other trademarks contained herein are the property of their respective owners KinetaMap UG 090401 ome eA B 3 spar fun LECT NICS KinetaMap User Guide 2009 04 01 The main logging routine will continue running until either the KinetaMap is turned off the battery dies or a USB cable is plugged into the device from a computer or any powered port The logging routine is drastically affected by the system parameters however it should be relatively easy to understand how the system parameters affect the logging scheme The routine will be continually checking for incoming messages from the GPS module The GPS module will be sending either or both GGA and RMC messages depending on the enable_gga and enable_rmc parameters GPS messages are received at 1 Hz intervals By default only the GGA messages are enabled The GPS messages are loaded into two places the coordinates along with the altitude if it s a GGA message are saved to the data file The entire message is saved directly to the NMEA file The routine will also read the accelerometer and battery voltage at the rate defined in the accel_frequency parameter If the frequency is greater than 1 Hz there will be more accelerometer readings than GPS messages This will be reflected in the log file by blank spots in the log
115. from a chair to assess balance control 4 The device can detect imbalance in an elderly person and indicate a need for physical therapy intervention to improve balance control One device on the market the iShoe detects abnormal pressure patterns via a shoe insole 5 Data collected by the PiShoe can be interpreted by a medical professional to evaluate the user s balance control However this device does not directly notify the user of their balancing ability This paper describes the design of a prototype balance control device for the elderly that gives the user instant feedback regarding their balance control The device will enable the user to seek medical attention in order to improve their balance control and decrease their risk of falling IL MATERIALS AND METHODS The first part of the project involved data collection using Figaro Belt Attachmant of KinetaMap and Experimental Set Up on the right side of the subject externally adjacent to the iliac crest Figure 1 Balance control data were also collected using an AMTI AccuSway force platform AMTI NetForce and BioAnalysis software were used to record and analyze the balance control data respectively The force platform was placed directly on the floor and a wooden platform at the same height was placed adjacent to the force platform A chair was placed on top of the wooden MM o 1 so that when the subjects sat down their feet were positioned comfortably on the forc
116. ge either capacitive resistive inductive or piezoelectric is attached to a wheatstone bridge in order to produce a small amplitude electrical voltage signal it is considered a pressure transducer Pressure transducers are available in many different forms with many different applications ranging from small IC component mounted pressure transducers to large pressure transducers useful in industrial and automotive settings The analog voltage output of pressure transducers makes them easily useable via basic signal processing Aston 1990 Research has shown that variations in the position of a person s COP along their feet can represent how well the person is balanced For example if a person is unbalanced their COP may be positioned closer to their toes while if a person is balanced their COP may be positioned close to the middle of their foot Pai amp Patton 1997 Current technology such as the iShoe has utilized the concept of measuring pressure along the feet as a means of assessing balance control Trafton 2008 Since pressure transducers and strain gauges are two methods of measuring pressure the team investigated these devices for use in the current balance control indicator 2 5 4 Motion Analysis Motion analysis laboratories are used to assess balance control and gait These laboratories have expensive infrared motion capture cameras and force platforms The force platform measurements are used to assess COP and BOS during
117. han balanced trials Figure 88 Y of FP in A m b e Figure 88 a COP plot for a balanced STS b COP plot for an unbalanced STS 152 After verifying that all trials are properly categorized as balanced or unbalanced the team needs to evaluate the acceleration plots Each acceleration plot should have a positive and negative amplitude For each plot calculate the time between the maximum peak and minimum peak Calculate the average time between peaks for balanced trials STS gt 3 and unbalanced trials rated lt 3 and the corresponding standard deviation Calculate the time range for a balanced STS and unbalanced STS by adding and subtracting the standard deviation from the corresponding mean For example if the mean is 5 and the standard deviation is 1 the range will be 4 6 Choose T by identifying the greatest time within the unbalanced time range that is not included in the balanced time range For example if the balanced time range is 5 5 7 3 and the unbalanced time range is 3 5 6 Te would be 5 Insert the new T into the code as explained in Section 5 4 8 1 3 Verify the device To verify the design attach both the KinetaMap and the new device to the belt such that when the belt is attached to subject 1 the new device is positioned at the right hip and the KinetaMap is positioned at the left hip With the belt attached conduct the same testing procedure as was performed in preliminary testing
118. he box because the box was too long 3 5 2 Battery Pack lt lt 2 5 Figure 75 Final Product of Device The list and size of all the components in the Duino Balance device is included in Table 17 The cost of each component and the total cost of the device is also listed The total cost to produce our device was 143 69 The myHalo Monitoring and BrickHouse fall detection devices cost about 200 for the device plus additional monthly fees for the monitoring service Halo Monitoring 2009 The Wii gaming console and Wii Fit Balance Board costs approximate 129 300 Therefore the hardware of our device would not only be cheaper than the current monitoring systems but also significantly cheaper in the long term because it would not require monthly fees for monitoring 130 Table 17 Cost of Device ELCHE CNN RN Arduino Duemilanove Microcontroller Board High Capacity Lithium Battery Backpack Arduino ProtoShield Board ADXL345 3 Axis Accelerometer Board CEM 1203 Buzzer Slide Switch 9V Wall Adapter Plug Plastic Project Box 3 Resistors Laptop Shoulder Strap Velcro Total 29 95 47 35 16 95 27 95 1 95 1 55 6 50 3 69 0 30 5 00 2 50 143 69 Length 2 95 Width 2 1 Height 0 6 Weight 0 07 Ib Length 2 73 Width 2 1 Height 0 498 Weight 0 22 Ib Length 2 95 Width 2 1 Height 0 6 Weight 0 1 I
119. he only signal and vibration pattern that triggers the device to detect a fall The fall alert can be sent through a wireless communication as a message to a cell phone or pager The detection range of this device is around 15 feet which is large enough to cover most rooms in a home or assisted living facility Alwan Felder Kell amp Dalal 2004 The obvious limitation of this device is its lack of portability it cannot be used for active elderly clients that intend to leave one room and go to another room or go outside to the supermarket etc This device despite its limitations was able to perform with 0 false alarms and 100 true positives Alwan Felder Kell amp Dalal 2004 35 e e eee my Rate of Travel etector Figure 9 Patent Application Number 2006 0195050 and Floor Vibration Prototpe Alwan Felder Kell amp Dalal 2004 2 3 2 3 Accelerometer Based Patents The most common method of detecting a fall is through the use of an accelerometer Devices that use an accelerometer to detect a fall vary by threshold acceleration values the algorithm used to confirm the fall and location of the sensor or sensors on the body Patent number 6 433 690 Figure 10 is a wearable accelerometer based device that aims to monitor the user detect a fall and automatically alert a caregiver or call station that a fall has 36 occurred The remote monitoring device 30 is worn on the
120. heir quality of life is a global one and the creation of a low cost device that could be accessible to lower income societies globally is an important step in the direction of improvement of the lives of elderly from many different situations Manufacturability and environmental consciousness were factors that were considered during the production of the device and because of this a device has been designed that is very manufacturable and additionally is RoHS compliant 147 8 DESIGN VALIDATION Through preliminary testing the team designed a device to be used during the STS activity that detects and directly notifies the user of an unbalanced STS The final device was tested for subject 1 college age female and buzzed during an unbalanced STS situation and did not buzz during a balanced STS situation Operation of the device required the user to perform the STS both a rehabilitation and daily activity Thus the device demonstrated its ability to strengthen the user s awareness of their balance control and force user rehabilitation In addition the design team verified that the design met client objectives by interviewing physical therapist and an elderly user Both reported that the design will be straightforward easy to use and not cumbersome In addition both clients reported that it could be used in a clinical setting such as a nursing home physical therapy session and also as additional tool to assess a patient s fall risk A
121. ht on a bare floor she has a fractured tibia She was feeding the cat in the middle of the night she was in her bare feet because she has some neuropathy of her feet she was feeding the cat she took a terrible fall on some water she spilled She had had her left knee replaced last year and she was careful to protect that so she broke her tibia right where it connects to the knee She had a lot of issues and MRIs So she has a special needs boy so she needs to hurry around in the morning and get him off to school and she s always hurrying When we went on vacation with her she used to get up and blast around and make the beds that s Rose Do you think before this fall she would have used this device No I think she wouldn t have because she d say I don t do that you know because we all think different I mean my sister would think that way but not Rose because we all think different Do you think that maybe after her surgery on her knee she was unstable She was slower it was painful walking but she was steady Do you think if a Dr told her to use this she would have No because she was steady C 6 But if she walks slower than she s probably not steady she probably had balance problems before No she had pain I actually shouldn t say that you know she probably did have some balance problems Yes through our research we found that walking slowly is a sign of balance problems Rose does have s
122. ibes the process used and strategic decisions that we made in developing functions and specifications of a feasible conceptual design 4 1 Preliminary designs Through extensive literature review and patent search the team identified advantages disadvantages and methods of signal detection of current fall and balance control technology Advantages of current devices include sense unbalanced situation before a fall occurs device is comfortable and catered to the elderly population Disadvantages of current technology include detects a fall after it has occurred not catered to the elderly population requires professional intervention In addition the team found that current devices monitor pressure acceleration or tilt to detect imbalance or a fall Based on these findings the team developed the four design alternatives described in this section 4 1 1 Preliminary Design 1 Shoe Insole The first preliminary design Figure 24 utilized an array of micro strain gauges embedded in a shoe or shoe insole The strain gages would detect a threshold pressure indicative of an 71 unbalanced situation Upon detecting this threshold pressure the insole would vibrate to notify the user that an off balance situation had occurred Advantages of this design were that pressure is a common and accurate means of assessing balance control In addition a force platform would be needed to identify the threshold pressure and WPI has possession of this equ
123. ill Bogert A J Pavol M J amp Grabiner M D 2002 Response time is more important than walking speed for the ability of older adults to avoid a fall after a trip Journal of Biomechanics 35 2 199 205 Buckman R F 2006 Patent No 7 150 046 Centers for Disease Control amp Prevention 2009 January 19 Cost of Falls Among Older Adults Retrieved September 25 2009 from CDC http www cdc gov HomeandRecreationalSafety Falls fallcost html Centers for Disease Control amp Prevention 2009 September 24 Focus on Preventing Falls Retrieved September 25 2009 from CDC http www cdc gov Features OlderAmericans 159 Chern Sheng Lin H C L C S 2007 Wearable device for real time monitoring of human falls Science Direct Chong R K Chastan N Welter M L amp Do M C 2009 Age related changes in teh center of mass velocity control during walking Neuroscience Letters 458 1 23 27 Clark R A Bryant A L Pua Y McCrory P Bennell K amp Hunt M 2010 Validity and reliability of the Nintendo Wii Balance Board for assessment of standing balance Gait and Posture 31 307 310 Culhane K M O Connor M amp Lyons G M 2005 Accelerometers in rehabilitation medicine for older adults Age and Ageing 34 556 560 Digikey 2009 Digikey Retrieved December 13 2009 from Tyco Alcoswitch http media digikey com photos Tyco 20Alcoswitch 20Photos SSJ1204 jpg
124. iminary data of subject 1 was used to calculate the time range during which a balanced STS occurred and time range during which an unbalanced STS occurred The latter end of the unbalanced time range was chosen as the expected time Te separating a balanced from unbalanced STS This minimized false errors but included the maximum amount of unbalanced situations The device Duino Balance was built with a tri axis accelerometer Arduino Duemilanove Microcontroller Board and a CEM1203 buzzer a rechargeable battery pack and slide switch for powering the device as well as a protoboard for connecting and attaching all the components Duino Balance is enclosed in a plastic project box and attached to a belt to be worn around the user s waist The device was programmed to detect the minimum and maximum peaks of the STS and measure the time Tm between these peaks If Tm Te the device was programmed to buzz If Tm gt Te the device was told to reset Device verification was conducted using the same tests used during preliminary testing During SW tests the device buzzed once when it should not have buzzed 90 accuracy and during tandem trials the device reset twice when it should have buzzed 80 accuracy Therefore the design was verified by having greater than 75 accuracy The team also validated the design by interviewing clients who reported the device was straightforward easy to use and not cumbersome and could be used in a clinical set
125. ions are not claimed to accurately predict a fall Therefore the device could deploy an air bag unexpectedly Another drawback to this device is that the undergarment could be uncomfortable and hard for an elderly person to put on and take off However if the device were optimized to detect only falls it could be useful in limiting hip injuries 33 Figure 8 Patent 7 150 048 Active protective garment Buckman 2006 2 3 2 2 Ambient Sensor Patents Ambient devices differ from wearable fall detection devices in that they detect a fall by measuring indirect factors such as pressure or vibrations in the floor An ambient floor vibration sensor Figure 9 was developed and patented by the Medical Automation Research Center at the University of Virginia This device is positioned on the floor and has a transducer at the bottom of the unit that comes in direct contact with the floor and is used to measure the vibrations patterns in the floor to detect human falls The device uses parameters such as frequency amplitude duration and succession The study at the University of Virginia showed that 34 different activities like walking and running have vastly different vibration patterns and these patterns are also very different from the vibration patterns that occur during and after falling An anthropomorphic dummy was used to determine the signal and vibration pattern of a human fall that was then programmed into the device as t
126. ipment Therefore no extra costs would be necessary for testing However the main disadvantage of this design was that the user would have to be wearing some type of footwear in order to use the device This was unfavorable because our client interviews Appendix C revealed that elderly have trouble putting on and taking off footwear Therefore this design would make the device difficult to manage 72 Coce dhekechon ines wf QC coder d x Z Naoto NE 5 X ffi Chain m aes Aesuce ined ucers Connmunice bec device CU clelechors Citeste CoMmuniates w Merocemhller anidedel i Svo bee M ae ed wj QV aecolocomekes A God c t ZEN e A ICE eoe petis pex c ee of Cook ortek meos nig I E d Cox Chance m COG liledito oe techn Cea k Figure 24 Preliminary design 1 Shoe Insole 4 1 2 Preliminary Design 2 Waist mounted Accelerometer The second preliminary design Figure 25 utilized a waist mounted triple axis accelerometer to monitor the acceleration of the user The accelerometer would detect a threshold acceleration indicative of an unbalanced situation Upon detecting this threshold acceleration the device 73 would buzz to notify the user that an off balance situation had occurred The electrical components of the device would be housed in a plastic box A clip on the back of the box would be used to attach the device to a belt or pants The
127. is pointed approximately parallel to the gravitational acceleration vector downward X Figure 53 Y Figure 54 and Z Figure 55 axis were again plotted in addition to a magnitude plot Figure 56 Pulsatile noise artifacts were due to the imperfections of the movement of the device human hand 112 X Axis Acceleration Acceleration m s 2 Time sec Figure 53 X axis acceleration during X axis tilt Y Axis Acceleration N lt D E c 2 i 2 o o o lt Tme sec Figure 54 Y axis acceleration during X axis tilt 113 Z Axis Acceleration eee Time sec Acceleration m s 2 Figure 55 Z axis acceleration during X axis tilt Magnitude of Acceleration Acceleration m s 2 20 30 Time sec Figure 56 Magnitude of Acceleration during X Axis Tilt 114 Lastly y axis rotation tests were performed and the device was positioned with the positive Y axis pointing directly upward in parallel with the gravitational acceleration vector The device was then rotated toward the USB port on the top of the device thusly also creating a change in the orientation of the Z vector relative to gravity first at approximately 45 degrees followed by a perpendicular orientation Observed below are the X Figure 57 Y Figure 58 Z Figure 59 and magnitude Figure 60 changes for this procedure Note that during any change in orientation of the device more
128. is well suited for mobile device applications It measures the static acceleration of gravity in tilt sensing appli cations as well as dynamic acceleration resulting from motion or shock Its high resolution 4 mg LSB enables measurement of inclination changes less than 1 0 Several special sensing functions are provided Activity and inactivity sensing detect the presence or lack of motion and if the acceleration on any axis exceeds a user set level Tap sensing detects single and double taps Free fall sensing detects if the device is falling These functions can be mapped to one of two interrupt output pins An integrated patent pending 32 level first in first out FIFO buffer can be used to store data to minimize host processor intervention Low power modes enable intelligent motion based power management with threshold sensing and active acceleration measurement at extremely low power dissipation The ADXL345 is supplied in a small thin 3 mm x 5 mm x 1 mm 14 lead plastic package Rev 0 Inhorrmton sashad by forts use norfor Ne rem rar Hm an m am mr TE um E UTE EUH s OT pege for disclaimers Devices is bebeved to be accurate and relable However no One Technology Way P O Box 9106 Norwood MA 02062 9106 U S A Tat 781 329 4700 Fax 781 461 3113 www analog com 2009 Analog Devices Inc All rights reserved SPECIFICATIONS Ta 25 C Vs 2 5 V Vou oo 1 8 V acceleration 0 g Cs 1 pF tantalum
129. l investigate how the hip moves while sitting down and rising from a chair This data will be collected in order to identity a movement pattern that is indicative of an unbalanced situation This information will be used to develop a device that will detect the point when a user is aff balance Procedures to be followed The subject will wear a belt with an attached accelerometer device You will stand on a force platform in front of a chair and sit down Then you will be seated in a chair with your hands on your hips and with your feet on the force platform Then you will be asked to rest your feet on the force platform with one foot directly behind the other so that the heel of one foot is touching the toes of the other Next you will be asked to rise from the chair at a comfortable speed and you will be able to rest for 2 minutes between trials You will also be asked to rise from the chair as fast as possible with 2 minutes of rest between trials You will be provided a break after half an hour Your participation will last for approximately 1 hour K1 Risks to study participants There is a minimal risk of falling The student investigators will be holding the chair and standing next to the platform in order to catch the subject if they da lose their balance and fall Benefits to research participants and others The benefits are you will become more aware about balance control difficulty and gain awareness about your body and mobility Th
130. le such as the Achilles tendon within pockets of the brace The accelerometer or gyroscope acts to monitor the acceleration or change of position of the ankle joint respectively A threshold acceleration or change in position defined through testing would occur before the individual was going to fall Upon reaching this threshold the sensor would activate the vibrating components in the brace The vibrations would signal to the user that he she was at risk of falling The ankle brace preliminary design also consisted of a removable strap that attaches via Velcro around the top of the ankle brace The strap houses the battery of the device and wiring components The wires would plug or snap into the accelerometer or gyroscope and vibrating components This preliminary design was most closely related to the device described in the Patent Number US 2004 0173220 A1 consists of a wearable system e g a shoe or sock that contains actuators that create vibration feedback in the ankle or foot Harry Collins Prplata amp Kleshinkski 2004 The vibration feedback acts to increase the sensitivity of the mechanoreceptors in the foot or ankle and enhance the sensory function of those with decreased sensory performance e g the elderly Although the patented device could enhance balance control in an elderly individual the device does not detect and notify the user when a fall is about to occur Therefore the preliminary design differs in that it would detect
131. lity of the elderly to become aware of their own balance control and helped them to prevent falls and fall injuries By developing a device that is low cost and can be used with no professional intervention we have enabled the elderly to frequently observe their own balance control from their home without the requirement of transportation to a clinical setting the without incurring the large expense of clinical visits and post fall hospitalization By developing a device that is simple to use with no outward complexity beyond a simple toggle switch and status indicator we have developed a device that will be unintimidating to the elderly and promote frequent use Therefore a device of this nature bridges the void between existing balance control indicators and in home fall monitoring devices for the elderly Combining simplicity of operation present in the fall indicators and the proactive nature of the balance control indicators the device creates a simplistic means for proactive balance control monitoring This addresses issues with devices not catered to the elderly population e g Wii Fit and those requiring professional intervention e g the iShoe and those which do not proactively monitor balance control or force user rehabilitation e g the Phillips Lifeline pendant and myHalo fall monitoring system This system using principles of monitoring acceleration of COM and how it relates to the stability limits of the user implements a simple means
132. ll the subject to stand from the seated position Ten seconds after the subject reaches a balanced standing position tell them to turn the KinetaMap off in order to stop data collection Repeat this procedure 10 times with the subject s feet positioned shoulder width apart Figure 86 giving the subject at least 3 minutes rest between trials and at least 2 hours rest every 5 trials If the subject is capable of standing with their feet in tandem have the subject repeat the same test except rising from the chair with their feet in tandem Figure 87 The test should be repeated 10 times with at least 5 minutes rest between trials and at least 2 hours rest every 5 trials If the subject is unable to stand or perform the STS with feet in tandem the STS should be repeated 20 times with feet shoulder width apart with 3 minutes rest between trials and at least 2 hours rest every 5 trials After each STS trial ask the subject to rate their comfort 150 on a scale from 1 5 1 being completely unbalanced and 5 being completely balanced The physical therapist should also rate the patient s balance control on the same scale after each trial Figure 86 Shoulder width foot position representing a balanced condition Figure 87 Tandem foot position representing an unbalanced condition The KinetaMap collects data at 20Hz and logs each trial in a Microsoft Excel document in terms of time and the X Y and Z components of acceleration The X
133. low which only baseline noise occurred were identified for each subject Figure 3 The mean time t was calculated for tandem and SW trials of each subject and compared using a two tailed Student s T Test Twenty additional SW and T trials were conducted for subject 1 A time t was established for subject 1 such that t t corresponded to an unbalanced STS Preliminary results were used to develop an algorithm to detect an unbalanced STS The algorithm was designed to detect a a and to calculate time t Ift lt t the algorithm activates a buzzer The device includes an ADXL345 tri axial accelerometer that is used to sense the acceleration pattems when the user sits down and mses fom a chair The algorithm was programmed on an Arduino microcontroller development kit and interfaced with the accelerometer and buzzer to alert the user of an off balance situation sarm Figure 2 COP Tendans STS Balanced goen amp Unbalanced hiak a3 z d 15 H B UG tnn E Oo Du ween iL t Ws oe ae LA AL Fie x 24 M eee i i 1 hoss LI io ir ia 16 i 20 Firme 3j Or a Oo oe Figure 3 Acceleration signal of balanced SW STS Amdam Time x C 7 u Figure 4 Acceleration signal of unbalanred T STS _ oe A 37 m HI RESULTS Preliminary testing of 5 subjects resulted in 50 trials rated asa 3 34 trials rated as a 4 and 14 trials rated as a F All subjects reported feeling m
134. ly Acceleration data were collected from each subject Force platform data were not collected from subjects 5 6 7 8 because these tests were performed in a home setting and the force platform was not available in this location Force platform data were collected from all other subjects 99 Figure 40 KinetaMap Triple Axis Accelerometer and Data Logger SparkFun Electronics 2009 Table 10 Subject information 1 2 3 4 5 6 7 8 Gender F F F M F M M F Age 21 20 21 21 53 23 63 50 Height 5 8 5 5 5 7 5 1 5 5 11 5 10 5 T Mass kg 67 70 63 63 67 82 75 59 The force platform and a regular chair were set up as shown in Figure 41 The force platform was placed directly on the ground A wooden platform the same height as the force platform was placed adjacent to the force platform The chair was placed on top of the wooden platform and in front of the force platform so that when the subject sat in the chair her feet rested comfortably on the force platform The KinetaMap was attached with Velcro to an adjustable belt The belt was positioned so that the KinetaMap was mounted on the right side of the subject externally adjacent to the iliac crest Figure 42 100 Figure 41 Chair and force platform experimental set up Figure 42 Attachment of the KinetaMap device The subject first stood still with feet shoulder width apart on the force platform Th
135. m Ten shoulder width STS tests were performed followed by ten tandem tests We used the KinetaMap and force platform data to confirm whether the STS activity was balanced or unbalanced The device was programmed specifically for subject 1 with an expected time between the maximum and minimum acceleration peaks of 0 6 seconds During the STS activity if the time between the peaks was greater than 0 6 seconds then the trial was considered balanced and the device should reset without a buzz But if the time between the peaks was less than or equal to 0 6 seconds the device should buzz to indicate that the trial was unbalanced An example of the progression of one of the shoulder width STS trials is shown in Figure 79 During this trial the subject appeared to remain balanced throughout the activity and the device did not buzz 137 Figure 79 Shoulder Width STS with Device In order to confirm whether the trial was balanced we analyzed the KinetaMap data and generated the acceleration vs time plot shown in Figure 80 We identified the maximum and minimum acceleration peaks shown by the blue circles in the figure and measured the time between the two peaks The time between the peaks was 0 95 seconds This time was greater than the expected time of 0 6 seconds which was representative of a balanced trial Shoulder width 38 E z 05 2 2 05 o o lt 1 5 1 3 5 7 9 Time s Figure 80 KinetaMap a
136. m talking on the phone and walking and I walked right off the stage Fell down about two steps and I tore my rotator cuff very painful They put the ice on my knees because that s where the blood was I said no I ll get up myself you know And how did it effect you after I ve suffered with it for a few years Because I went for PT and got pretty good strength and pretty good range of motion I went back a second time for PT and I neglected doing exercises I know I m going to do them all my life but now I m seriously considering having it repaired but now I m 82 years old and I wonder will I need rehab and how am I going to manage to get dressed and take care of myself so that s a problem C 1 What s the biggest change you ve had in your daily life after the fall One of the things I loved was once a month I d go to mustard seed and volunteer to cook for like 200 people with my church and I liked being there and I liked being with the crowd and cooking and doing the serving I can t rotate that arm I cannot like big bowls I can t stir the things They asked me to just do the dishes they wanted me there and I said I can t even do that I can t rotate to dry or wash them And that was one thing now it s cooking and carrying bundles there s a lot of things I can t talk with my hands like I like to I didn t know how much I did that I have to sometimes pick up a mug of beer with two hands and at the end of the day can you bring
137. most sensitive to monitoring balance control because they are closest to the body s COM Since the device has to be comfortable and easy for the user to put on and take off the team eliminated the sternum location because a device placed here would require the user to lift their arms In addition the team found through research that the most sensitive location on the trunk for monitoring balance control was the hip Ward Evenson Vaughn Rodgers amp Troiano 2005 The lower back was a difficult place to reach and locate While sitting the device would also be more prone to being bumped if placed on the lower back 94 Lower Back Figure 36 Possible locations of device http www ehow com how 4629008 draw person standing html Therefore the team eliminated lower back and chose to monitor acceleration during the STS at the hip as shown in Figure 37 The hip bone is a universal anatomical marker making the device user friendly and since the hip is a reliable location for measuring acceleration it ensures high accuracy Trunk L ower Sternum Back Figure 37 Choosing the Device Location 4 3 4 Choosing the Method of Analyzing Acceleration Data The two goals of the project were to proactively force user rehabilitation and strengthen the user s awareness of their balance condition In order to force rehabilitation the device required the user to perform the STS activity To strengthen awareness the device monitor
138. mount of time during which their COM is outside the BOS However these gait changes often put them at a greater risk of falling Woolley Czaja amp Drury 1997 Kinematic and kinetic differences at the trunk hip knee and ankle are also present in the elderly and lead to an increased risk of falling The elderly have a decreased range of motion in 29 ankle knee and hip Elderly people have reduced ankle range of motion larger ankle plantar flexion at heel contact reduced ankle power generation and delayed dorsiflexion Kemoun et al found that delay in ankle dorsiflexion during gait is one factor that could prevent falls Kemoun Thoumie Boisson amp Guieu 2002 When walking down stairs elderly people also have a substantial decrease in ankle range of motion Woolley Czaja amp Drury 1997 Tripping over an obstacle or falling down stairs can be also be caused by reduced knee flexion Schillings Mulder amp Duysens 2005 Range of motion at the hip is another factor that increases the risk of falling in the elderly Hip extension in the elderly is reduced during walking Increased anterior pelvic tilt and hip extension moment during the swing phase also occurs in the elderly Kemoun Thoumie Boisson amp Guieu 2002 During falls the elderly also experience increased trunk flexion and trunk velocity Grabiner et al found that older adults had trunk flexion angles after a trip that were double the trunk flexion angle of the
139. mposes a restriction or limitation on the design Dym amp Little 2003 Constraints of the design included elderly disabilities available testing equipment and affordability Because of the limited gait analysis capacity we had in the 67 laboratory we needed to confirm balance condition by making use of only a single force platform and a tri axial accelerometer We also needed to factor in affordability and user acceptance In order for a device to reach the widest population of elderly clients it needed to be affordable and uncomplicated in its operation We needed to account for varying elderly disabilities making sure that the feedback provided by the device as well as the controls to operate the device were minimalist and simple We needed a means of alerting the user that would be able to be understood and sensed by a wide margin of elderly subjects and directions that were simple for clients to understand 3 3 REVISED PROBLEM STATEMENT According to the goals objectives and constraints of the project the design team revised the original client statement as follows Design balance control indicator which can be used particularly by the elderly to improve awareness of their balance control The first part of a project will involve data collection and analysis of acceleration of COP signals during both a balanced and unbalanced STS activity An accelerometer data logger attached to the subject and an AMTI force platform will be u
140. n Whereas the user may not have been aware that he or she was in need of assistance in order to become balanced before the application of the device a sounding buzzer will notify the user that he or she is not so balanced as he may have perceived In the daily performance of the STS activity and by attempting to reduce imbalance so as to not trigger the device to sound the user will be forced to actively participate in rehabilitation Without taking any extraordinary measure as using the device can easily be incorporated into the daily routine of the elderly the elderly user has taken a step towards the recognition and improvement of his or her balance condition The limitations to the data used to design the device would pose a quite passable challenge in verifying the functionality of the device for direct use in the elderly population Foremost the subjects used to gather information used to set device thresholds and values for programming were healthy young subjects mostly under the age of 30 The values that would be obtained in elderly testing may vary which would necessitate the reprogramming of the device As with any real world device there is the possibility for error in the results given by the device as well Through preliminary testing thresholds were collected that represented the data obtained from most of the trials collected There are of course trials which may have been balanced but resulted in a peak to peak time duratio
141. n bus The duration of this state should to during power up Se fonctions are available but the device will not craste a conflict on the POWER SAVINGS Power Modes The ADXL345 automatically modulates its power consumption in proportion to its output data rate as outlined in Table 6 1f additional power savings is desired a lower power mode is available In this mode the internal sampling rate is reduced allowing for power savings in the 12 5 Hz to 400 Hz data rate range but at the expense of slightly greater noise To enter lower power mode set the LOW POWER bat Bit 4 in the BW RATE register Address 0x2C The current consumption in low power mode is shown in Table 7 for cases where there is an advantage for using low power mode The current consumption values shown in Table 6 and Table 7 are for a Vs of 2 5 V Current scales linearly with Vs Table 6 Current Consumption vs Data Rate Ta 25 C Vs 2 5 V Vnnio 1 8 V Table 7 Current Consumption vs Data Rate Low Power Mode Ta 25 C Vs 2 5 V Voom 1 8 V Rev O Page 6 of 6 ADXL345 PC With CS tied high to Vorvo the ADXL345 is in PC mode requiring a simple 2 wire connection as shown in Figure 8 The ADXL345 conforms to the UM10204 PC Bus Specification and User Manual Rev 03 19 June 2007 available from NXP Semiconductor t supports standard 100 kHz and fast 400 kHz data transfer modes if the timing parameters given in Table 11
142. n 63 3 1 CLARIFYING THE ORIGINAL PROBLEM STATEMENT eese enne rennen entrent 63 3 2 OBJECTIVES SC CONS I RAINTS iit Tt ote got retra 66 3 3 REVISED PROBLEM STATEMENT ameiita E T ei TRE EOR EEA AA i a 68 3 4 PROJECT APPROACH eer re eite ete e a Ee RD ett a e et a rae ee aa 69 ALTERNATIVE DESIGNS 4 cero eroe den boe ener ee ete tin Ferd m RAE aa eer neri yep aaiae iaaa idare NiET 71 4 1 Preliminary designs tte rice e ren te nete d edes 71 4 1 1 Preliminary Design 1 Shoe Insole esses enne nnne nnn 71 4 1 2 Preliminary Design 2 Waist mounted Accelerometer eese 73 4 1 3 Preliminary Design 3 Ankle Brace Accelerometer Gyroscope eee 75 4 1 4 Preliminary Design 4 V Tact Belt enne nnne nnn nnne nna 78 4 2 NEEDS Bl 79 4 2 1 Rank ordering Design Objectives cesses nennen nennen nnne nnne nnne 79 4 3 FUNCTIONS AND SPECIFICATIONS esses enne nns nnee testes tentent nnne nnns nnns 85 4 3 1 Choosing the Activity sebisid a Ea ee rne er eode nd ae on e eden dena a 87 4 3 2 Choosing the Signal and Sensor aione iii eene nnne enne nennen nnns 90 4 3 3 Choosing the Location ne net tre Re ede a roe YR EE PRX VER CER Pa dr eX VE Red 93 4 3 4 Choosing the Method of Analyzing Acceleration Data ccccsssccccceeessessnseceeeessessessaaeees 95 4 3 5 Final Solution e eter teer e P D a re n b E e eines 98
143. n Electronics Inc All other trademarks contained herein are the property of their respective owners KinetaMap UG 090401 ome eA B 5 spar fun LECT NICS KinetaMap User Guide 2009 04 01 access the configuration file plug a USB cable from a computer into the KinetaMap and turn the KinetaMap on When the Mass Storage Device drive opens on the computer you will see the configuration file in the drive The configuration file is named kineta_config txt and if you open the file in a text editor you will see eleven parameters You can also view the file from the config menu config_menu 0 log_gps 1 log_accel 1 send_gps_to_bt 0 log_batt 1 log_nmea 1 enable_waas 1 full_scale_accel 0 enable_rmc 0 enable_gga 1 accel_frequency hz 10 The parameters in the configuration file must be kept in this order with a carriage return separating each parameter All of the parameters except the accel_frequency parameter only have two options 1 or 0 If the parameter is set to 1 than the function is enabled if the parameter is set to 0 than the function is disabled Don t forget you can also change the parameters from the Config Menu However if you want to enable the config menu you must first do this by directly editing the config file on the SD card As a side note if the configuration file has somehow been corrupted don t worry about it The firmware will detect a corrupted configuration file delete it and create a new one wi
144. n balanced decreases However studies showed that consistent exercise especially Tai Chi can improve balance control and decrease the risk of falling This chapter discusses the details of balance control how it is assessed why elderly are susceptible to falls and how balance control can be improved The final sections of this chapter discuss advantages and disadvantages of current technology aimed at reducing the negative effects of falls and subsequently identify important objectives that will guide the design of the present balance control indicator 2 4 UNDERSTANDING BALANCE CONTROL To analyze how each sensory system contributes to balance control researchers assume the human body behaves like an inverted pendulum Winter Patla Prince Ishac amp Gielo Perczak 1998 the body is always swaying Certain factors regarding the body s movement can dictate how well a person is balanced This section describes how the body maintains balance during locomotion and while standing how each sensory system contributes to maintaining balance In addition this chapter discusses the parameters associated with balance control and ways that these parameters are measured in order to assess a person s balancing ability 2 1 1 Basics of Balance Balance is how the body moves relative to the gravitational force vector in order to maintain posture and prevent falling To understand how balance is maintained researchers study the relation between
145. n of less than 0 6 seconds and likewise for unbalanced 144 trials Though there is this limitation we do feel that in most instances our device will be able to accurately distinguish balanced from unbalanced situations with minimal opportunity for error The clear and present economic benefits of the widespread implementation of a device of this type are abundantly apparent Minimization of healthcare dollars spent treating victims post fall could reduce overall national spending With an aging population and a need for reduction in healthcare dollars spent the utility of preventative maintenance shines A low cost in home device that can reach a widespread population and assist them in regaining their balance and mobility could reduce overconsumption of natural resources used to produce other devices and facilities that are needed after falls have occurred specifically increased rehabilitation facilities and products as well as mobility devices and intensive care resources The device has the capacity to reach a global market in that it is low cost and that it is needed globally in any culture where there are elderly people In areas where there is a reduced availability of healthcare a low cost device that could prevent the need for expensive treatment would be a welcomed alternative to having lower income brackets not receive the healthcare that they need Likewise in developed areas with modernized healthcare a reduction of expenditures
146. nality Rev O Pago 4 of 4 PIN CONFIGURATION AND FUNCTION DESCRIPTIONS Table 4 Pin Function Descriptions Pin No Digital interface Supply Voltage Must be connected to ground Reserved This pin must be connected to V or left open Must be connected to ground Must be connected to ground Supply Voltage Chip Select interrupt 1 Output interrupt 2 Output Not Internally Connected Reserved This pin must be connected to ground or left open Serlal Data Output Atternate PC Address Select Serlal Data PCy Serlal Data Input SPI 4 Wire Serlal Data input and Output SPI 3 Wire Sertal Communications Clock on C U P wn Voo o GND Reserved GND GND Vs Cs INTI INT2 NC Reserved T i Rev O Page Sof 5 ADXL345 THEORY OF OPERATION The ADXL345 is a complete 3 axis acceleration measurement system with a selectable measurement range of 2 g t4 g tB g or 116 g It measures both dynamic acceleration resulting from motion or shock and static acceleration such as gravity which allows the device to be used as a tilt sensor The sensor is a polysilicon surface micromachined structure built on top of a silicon wafer Polysilicon springs suspend the structure over the surface of the wafer and provide a resistance against acceleration forces Deflection of the structure is measured using differential capacitors that consist of independent fixed plates and plates attached to the moving ma
147. nbalanced STS in terms of acceleration and notify the user when an unbalanced STS occurs In order for the device to do this the team needs to 98 develop an algorithm or a set of well defined instructions to completing a task In this project the task was to notify the user of unbalanced situation Therefore the team needed to conduct preliminary testing to identify a significant difference between the acceleration patterns of a balanced and unbalance STS In particular the team needed to verify that the time duration of the STS would be longer when a person was balanced and shorter during an unbalanced STS This sections details the methods used to verify the hypothesis results and conclusions drawn from preliminary testing 4 4 44 Materials and method Preliminary tests were conducted to identify significant differences between the acceleration patterns of a balanced STS situation and unbalanced STS situation Nine healthy subjects participated in preliminary testing One subject male age 11 was eliminated from testing because he did not perform the STS properly Subject information can be found in Table 10 The SparkFun KinetaMap SparkFun Electronics 2009 data logger Figure 40 containing an ADXL345 tri axial accelerometer Sparkfun Electronics 2009 was used for collecting acceleration data The AMTI AccuS way force platform and AMTI NetForce and BioAnalysis software were used to record and analyze balance control data respective
148. nce control and postural stability in the elderly because they were not able to generate enough muscle force in response to 31 a perturbation Horlings van Engelen Allum amp Bloem 2008 As a result many studies evaluated the effects of physical therapy and strength training on improved balance and concluded that physical therapy is a successful form of intervention Rose 2005 It is also difficult for the elderly to control movements of their COM and BOS during forward and backward sway which can cause falls Tai Chi shown in Figure 7 is a type of exercise that involves slow forward and backward motions and has been used in many research studies for the elderly population Mao Hong amp Li 2006 The movements during Tai Chi help the elderly concentrate on slow movements weight shifting flexibility foot positioning and proprioception Studies have concluded that after participating in Tai Chi the elderly have increased awareness of their body sway and limb movements reduced fear of falling and considerably improved balance control Rose 2005 Figure 7 Elderly participating in Tai Chi Rose 2005 The Nintendo Wii Balance Board and Wii Fit gaming console has several games and activities that specifically target balance control and COP movements A study by Clark et al investigated the validity of the Wii Balance Board for assessing balance control The study 32 concluded that the Balance Board is comparable to
149. nd Function Retrieved April 04 2010 from Neuroscience Online http nba uth tmc edu neuroscience s2 1i10 1 html Gross M Stevenson P Charette S Pyka G amp Marcus R 1998 Effect of muscle strength and movement speed on the biomechanics of rising from a chair in healthy elderly and young women Gait and Posture 5 175 185 Halo Monitoring 2009 Halo Monitoring Retrieved September 20 2009 from myHalo http www halomonitoring com halo LinkClick aspx fileticket afZzmD8X BOc 3d amp tabid 1 10 Harry J D Collins J J Prplata A A amp Kleshinkski S J 2004 Patent No 10 793 729 Horlings C G van Engelen B G Allum J H amp Bloem B R 2008 A weak balance the contribution of muscle strength to postural instability and falls Clinical Practice Neurology 4 9 504 515 Hurt H H 1965 Aerodynamics for Naval Aviators Naval Air Systems Command United States Navy 16 21 Huynh J 2009 Liquidware Open Source Electronics Retrieved December 4 2009 from Lithium Backpack http www liquidware com shop show BP Lithium BackPack Internet Stroke Center 2010 Stroke Scales and Balance Assessment Tools Retrieved April 18 2010 from The Internet Stroke Center http www strokecenter org trials scales berg html Janssen W G Bussmann J B Horemans H L amp Stam H J 2005 Analysis and decomposition of accelerometric signals of trunk and thigh obtained during the sit to
150. nd the ground when standing Certain angles of the body have been shown to correspond to unbalanced situations In addition gyroscopes have been used in combination with accelerometers in fall detection devices Therefore the team investigated gyroscopes for use in the design of a balance control indicator 2 5 3 Strain Gauges and Pressure Transducers Strain gauges are electrical resistors configured to measure strain by means of changing resistance conductance values when the device is deformed A Typical strain gauge will change in its resistance value when compressive or tensile force is applied This changing value of resistance can easily be used in a simple circuit to produce a varying voltage or current change as necessary There are many different types of strain gauges ranging from simple foil strain gauges through piezoelectric sensors designed from semiconductive material with capacitive and fiber optic strain gauges in between Different types of strain gauges tend to have markedly different gauge factors or sensitivity to strain from one another For instance piezoelectric sensors are much more sensitive to deformation than foil gauges These devices are also much more sensitive to temperature deviations like most resistive circuits Omega Engineering Inc 2003 59 Pressure transducers convert pressure signals into an analog electrical signal They sometimes implement and are very similar to strain gauges When a strain gau
151. ndly and since the hip is a reliable location for measuring acceleration it ensures high accuracy Once fumed on the device operates on it own requiring no user intervention The life Balance control parameters vary among users Therefore collected and replaced in the algorithm In the future the device can be improved to assess balance canto during various tests Dwe to its ability to instantly by physicians physical therapists or an elderly individual to ind cut if there is a need for balance control improvement V CONCLUSION We have designed a prototype wearable device that can detect an abnormal acceleration pattern during a STS activity and alerts the user of an off balance situation Our device will allow an elderly person to monitor their balance control while rising fom a chair By signaling an off balance situation during STS our device will be able to notify the user of poor balance cantral ACENOWLEDGEMENTS We would like to thank our Lab Manager Lisa Wall Adriana REFERENCES I F Tideikraar 2007 Falls in Older People Prevention amp zu ogo y L Tean Effect of mcla Sca Enc 2007 News RasereiFdrery 1 2010 Som iShow hutpclurucar ahos net new hind riumg from a chair in healthy siderhy and voumg women ndr ane Poanse 8 175 185 E 1 3151 F amp E 5 Gross M Stevenson pei deer Mam 3 APPENDIX J Conversions of Acceleration Data Raw data from the accelerometer was out
152. nfiguration file as well as provides the ability to load custom firmware Using the KinetaMap Default Firmware While SparkFun encourages it s customers to create and modify custom code to load onto the KinetaMap we ve also created a new and robust default firmware version that should be able to meet the needs of many applications This section of the User Guide will describe the new firmware version 1 2 and give an overview of the flow of the firmware Quick Start If you don t really care how the firmware works and you just want to get going well this section is for you To get the device working out of the box all you need to do is flip the switch to ON By default all of the bluetooth functions are disabled The device will begin logging to two files a data file which contains the accelerometer readings battery voltage readings and GGA GPS data and a NMEA data which logs the unedited GGA GPS messages The accelerometer and battery readings will be logged at 10Hz and the GPS messages will be logged no faster than 1 Hz Invalid GPS messages will not be logged The accelerometer will be set to 2G mode WAAS will be enabled on the GPS module Data will be logged to each file until either the device is turned off the device runs out of battery power or a USB cable is plugged in from a computer To retrieve the data files from the device plug a USB cable into the device from a computer with the power to the KinetaMap turned on This
153. ng p Battery Elastic brace Vibrating Component Vibrating Component gt N Gyroscope accelerometer Figure 26 Preliminary Design 3 Ankle Brace 77 4 1 4 Preliminary Design 4 V Tact Belt The V Tact Belt design Figure 27 combined the concept of a vibrotactile navigation device with the SHIMMER platform accelerometer based sensor Lorincz Chen Patel amp Welsh 2008 The design incorporated small vibration tactors similar to the ones used in a cell phone or pager into the inside surface a belt SHIMMER sensors with the gyroscope board connection will be placed inside MP3 player holders and attached to the outside surface of the belt The device is worn at the waist because it is close to the user s COM and is attached using Velcro so the user can put it on and take it off easily The SHIMMER sensors will be programmed and used to detect when the user is off balance and at risk of falling When the SHIMMER sensors determine the user is off balance a signal will be transmitted to the vibration tactors on the side which the user is unbalanced The vibration tactors will then vibrate and alert the user that they are off balance in a particular direction and the vibration should help them readjust their posture back to a balanced stance or gait This device can be considered an early fall detection device because it detects and alerts the user when they are off balance and at risk of falling and allows them to potentially corre
154. ns to program and reprogram the device based on the changing needs of the user Such device personal computer communication would allow for greater flexibility of the device in that it could be reprogrammed within a clinician s office and used during physical examinations for multiple elderly clients 156 GLOSSARY Base of Support BOS The area of the body that is in contact with the support surface Gielo Perczak 2009 Bluetooth An open wireless protocol for exchanging data over short distances using short length radio waves from fixed and mobile devices Center of Mass COM In a uniform gravity field to represent the unique point in an object or system which can be used to describe the system s response to external forces and torques Nave 2010 Center of Pressure COP The point on a body where the total sum of the aerodynamic pressure field acts causing a force and no moment about that point Hurt 1965 Dorsiflexion Flexion of the foot in an upward direction Encyclopedia 2010 Inverted Pendulum Model A pendulum with a mass above its pivot point Personal Emergency Response Systems PERS An electronic device designed to let the user summon help in an emergency Federal Trade Commission 2002 Post fall anxiety Syndrome Post fall syndrome is commonly observed in geriatric medicine affecting near one out of five fallers Left untreated this condition can lead to a regressive syndrome with physi
155. o software which is open source software based on the C C programming language Code is stored on the Arduino using internal 16KB of Flash memory Since this is open source software there are a lot of resource materials example codes tutorials and other reference materials online The board is 2 1 wide by 2 95 long by 0 6 tall and is shown in Figure 67 relative to a 6 inch A to B USB plug Arduino 2009 122 http www liquidware com shop show ARD Arduino Duemilanove Figure 67 Size of Arduino and its size relative to a 6 inch A to B USB plug Arduino 2009 The most important component of our device is the ADXL345 accelerometer Figure 68 that is used to sense accelerations patterns during the STS activity One of the reasons we chose this specific accelerometer because it is the same accelerometer in the SparkFun KinetaMap that we used in preliminary testing Also according to the ADXL345 Data Sheet the ADXL345 is well suited for mobile device applications It measures the static acceleration of gravity in tilt sensing applications as well as the dynamic acceleration resulting from motion Analog Devices 2009 The ADXL345 is a low power tri axial accelerometer that can be set to different sensitivities but for our application is set to 2g This accelerometer is attached to a breakout board allowing for a simple connection to the Arduino microcontroller board Sparkfun Electronics 2009 The accelerometer is connec
156. oShield Board Sparkfun Electronics 2009 sse 126 Figure 74 Schematic of Device uideri dei retenir nr ie ride kho rk iege ern eheu Regi oet e eria eire oe dao 127 Figure 75 Final Product of Devices iis c iscccssetstessecsidacedscesscbscsaiecsescadecedevsiactssesssotesaddeisevesseatbestlectsocniaatieeetas 129 Figure 76 Acceleration curve for the STS activity showing the peaks and time between them 132 Figure 77 Block Diagram of Final Design eese nnnm enne nennen 133 Figure 78 Block diagram of device functions and specifications for subject 1 esses 136 Figure 79 Shoulder Width STS with Device sees een nennen eene 138 Figure 80 KinetaMap acceleration vs time data during Shoulder Width STS with device 138 Figure 81 COP graph for shoulder width black and tandem green STS trial esssss 139 Figure 82 Tandem S TS with Device to eere t rere ce e I T Gite te E Re etae ERE PR egets 139 Figure 83 KinetaMap acceleration vs time data during Tandem STS with device 140 Figure 84 Chair and force platform experimental set up essen 149 Figure 85 Attachment of the KinetaMap device sees eene nennen 150 Figure 86 Shoulder width foot position representing a balanced condition esses 151 Figure 87 Tandem foot position representing an unbal
157. occurs Therefore the activity should be a daily activity that 1s repetitive in the life of the elderly sensitive to balance control and a feasible rehabilitation technique as shown in Figure 29 Gross Stevenson Charette Pyka amp Marcus 1998 87 Sensitive to balance control Feasible rehabilitation technique Repetitive in Daily Life Figure 29 Factors needed in choosing an activity To ensure that the chosen activity would be sensitive to balance control the team looked at the fourteen activities on the Berg Balance Test BBT shown in Figure 30 The team identified through research and an interview with a physical therapist at Fairlawn Rehabilitation Hospital that the BBT was the most common exam used to assess balance control and that three activities on the BBT were the most sensitive to balance control standing on one foot standing in tandem one foot in front of the other and the sit to stand STS Lauren Roberts Fairlawn Rehabilitation Hospital 88 ITEM DESCRIPTION SCORE 0 4 Sitting to standing Standing unsupported Sitting unsupported Standing to sitting Transfers Standing with eyes closed Standing with feet together Reaching forward with outstretched arm Retrieving object from floor Turning to look behind Turning 60 degrees Placing alternate foot on stool Standing with one foot in front Standing on one foot Total Figure 30 Berg Balance Test American Academy of Health and Fi
158. of things you know And my friend you know she has M S she s going to start some PT because she has a lot of weaknesses and she s having trouble walking PT is wonderful for the elderly and for young people with problems It s hard to keep it up when you get home though doing the same thing over and over It s one thing we were thinking you go to physical therapy and then you go home and it s all up to you and if you had this device you could see how you improved and then keep using it after and see if your balance started to get worse do you think that s useful That s a good point I know that s why my shoulder has gotten worse from not doing exercises I could always walk I could do that forever but those exercises just made me crazy I know I need the strength back for the summer because raking and mowing the lawn and all those things I like to do they re going to hurt you know they re going to hurt Do you think doing those things is going to make it worse I do because I was outside raking one day before vacation and I was trying to get done raking before the sanders and a couple day later it hurt like a son of a gun and I remembered back and I said oh that must have hurt But now I m careful when I rake I rake with my left hand and the other one just goes with the rake Okay I m done with the interview thanks that was great 2 The following are written responses to an e mail interview The interviewee is
159. ome balance problems but she s cautious and she s not unsteady so I don t think the alarm would have helped her But that had nothing to do with the fall she slipped on water you know I think the alarm would help if she were going to recuperate you know just to slow her down The purpose in Rose s case was that she probably could have had balance problems even before the surgery to do strengthening and things like that Did she have PT after her surgery she must have Yes a while and then she had people come to the house and do her exercises with her too Rose s main issue was that she s always moving too fast and someone needs to slow her down You know I guess that could be useful if you could just use it to help slow people down Do you think this would be helpful if we could incorporate it into a physical and have a person just stand up 5 times and if it buzzed they could use that to do physical therapy to improve their balance Yes you know I think that could work Because I don t think Doctors put enough time into physical therapy in older people you know they really don t The older you get it s part of the aging process that s what everyone says you know everything is part of the aging process I think Physical therapy should almost be something that the doctors do at a certain age or when C 7 they see certain symptoms like instability or weakness Physical therapists can pick up on a lot
160. ome use especially for the elderly because there are too many transducers to apply and proper location is important 49 Figure 17 Patent Number 6 984 208 Method and apparatus for sensing body movement Zheng 2006 In addition to monitoring body sway and movement vibrotactile devices can be used to correct posture Patent number 4 750 480 Figure 18 is a belt that is worn around the waist to detect when the user is slouching and activate a vibration signal to alert the user to correct their posture The pad 12 is worn on the front of the abdomen and is used to determine whether the abdominal muscles are tightened or relaxed A switch on the inside of the pad is pressed when the user relaxes their abdominal muscles and this triggers the vibration component to activate and continue vibrating until the user tightens their muscles and releases the switch Jenness 1988 A benefit of the design is the device shows that the user is able to respond to the vibration feedback and adjust their posture accordingly One drawback of the device is the pad is very 50 large and could be uncomfortable to the wearer Another drawback of the device is it needs to be worn on the outside of the user s clothing This would attract unwanted attention to the user making them self conscious and embarrassed and ultimately not want to use the device Figure 18 Patent Number 4 750 480 Posture correcting devices Jenness 1988 There is also a
161. on of dow time t4 the dock rise time tx and the minimum data setup time fad This value Is Clicubated as temei t De tamni Cols the total capacitance of one bus line In picofarads Figure 10 FC Timing Diagram Fav 0 Page 11 of 11 Trigger Mode In trigger mode FIFO accumulates samples holding the latest 32 samples from measurements of the x y and z axes After a trigger event occurs and an interrupt is sent to the INTI or INT pin determined by the trigger bit in the FIFO CTL register FIFO keeps the last n samples where n is the value specified by the samples bits in the FIFO_CTL register and then operates in FIFO mode collecting new samples only when FIFO is not full A delay of at least 5 ws should be present between the trigger event occurring and the start of reading data from the FIFO to allow the FIFO to discard and retain the necessary samples Additional trigger events cannot be recognized until the trigger mode is reset To reset the trigger mode set the device to bypass mode and then set the device back to trigger mode Note that the FIFO data should be read first because placing the device into bypass mode clears FIFO Retrieving Data from FIFO The FIFO data is read through the DATAX DATAY and DATAZ registers Address 0x32 to Address x37 When the FIFO is in FIFO stream or trigger mode reads to the DATAX DATAY and DATAZ registers read data stored in the FIFO Each time data is read from the FIFO
162. onal Review Board Professor Kent Rissmiller Tel 508 831 5013 Email kjr wpiedu or WP s University Compliance Officer Michael J Curley Tel 508 831 6219 Email mjcurley amp wpi edu Your participation in this research is voluntary Your refusal to participate will nat result in any penalty to you or any loss of benefits to which you may otherwise be entitled You may decide to stop participating in the research at any time without penalty or loss of other benefits The project investigators retain the right to cancel or postpone the experimental procedures at any time they see fit Data obtained in this experiment will become the property of the investigators and WPI If you withdraw from the study data already collected from you will remain in the study K2 By signing below you acknowledge that you have been informed about and consent to be a participant in the study described above Make sure that your questions are answered to your satisfaction before signing You are entitled ta retain a copy of this consent agreement Date Study Participant Signature Approved by WPI IRB From 18 2010 To 4 7 2011 Study Participant Mame Please print Date Signature of Person who explained this study K3 100 Insitute Road Wore MA 01600 2291 LSA 08 831 5000 Fax SCIRE 1 409 wwwu wp edu Worcester Polytechnic Institute IRB 1 IRB 00007374 8 April 2010 File 10 067 Worcester Polytechnic Institute 100 Institute Roa
163. ond to the stimuli quickly One disadvantage of this design is the purpose of the device is not to detect a fall A major drawback of this design is the location of the device on the upper chest would make it very difficult for an elderly user to put on and take off 84 Figure 16 Patent Number 5 919 149 Apparatus and method for determination of body sway Allum 1999 Patent number 6 984 208 Figure 17 is a device used to measure the movement and posture of various body parts by transmitting ultrasound signals into the user s muscles The ultrasound signals are sent to the muscles by ultrasound transducers and recorded by the 48 receivers 12 amp 14 The ultrasound signals are either scattered or reflected by the muscles and this can be used to determine the position angle and stiffness of the muscles as well as detect changes in the muscles during particular movements Zheng 2006 This design identified an additional method ultrasound signals that can be used to measure body movement and detect abnormal body movement One benefit of this device is the ultrasound sensors and system can be combined with EMG sensors accelerometers and gyroscopes to apply the design to other applications A drawback of the design is the design would be beneficial in a clinical setting such as a gait analysis lab or hospital where a trained medical professional could apply the electrodes in the proper location but would not be practical for h
164. onitoring device Myllymake 1996 An accelerometer based device can also be used to detect the orientation of the user Patent number 6 611 783 Figure 12 is a device that uses an accelerometer placed on the back of the user s thigh to detect tilt and variation from a reference angle The device can be used to monitor patients in a hospital determine range of motion assessment in physical therapy or prevention and detect falls The threshold values and reference angles for the accelerometer can be adjusted depending on the application When threshold value or reference angle is exceeded an alarm is activated to alert and provide feedback to the user An alert is also sent by a radio frequency transmitter to a caregiver or monitoring station Kelly amp Schoendorfer 2003 This device is easy to use and does not require any manual activation of the device One drawback of this device is the risk of false alarms if the user exceeds the reference angle when bending over to pick something up off the ground or sitting in a chair 39 ul Figure 12 Figure 10 Patent Number 6 611 783 Attitude indicator and activity monitoring device Kelly amp Schoendorfer 2003 Patent 6 997 882 Figure 13 is a combination of the methods used in the previous patents and monitors the user s activities measures physiological signals and detects the velocity acceleration orientation and position of the user This design utilizes sensors attached
165. oprioception muscle weakness reaction time postural sway gender post fall anxiety syndrome use of 26 medication chronic diseases both neurological and musculoskeletal and mobility A fall can be caused by extrinsic factors or intrinsic factors or a combination of the two Rubenstein 2006 In the elderly population physiological factors have a major impact on why the elderly lose their balance and cannot recover when they fall and why they cannot get up after a fall 27 Muscle weakness Diseases Osteoporosis Musculoskeletal ncreased trun flexion Slower walking pecd Smaller Base of Support Reduced stride length Increased reponse time Deceased postura stabili Balance f Control of COM Factors NE contributing none to a fall Elevated walkway Stairs Fear of Falling Previous falls Alcohol Heart attack Figure 6 Factors that could contribute to a fall Balance control decreases with age Elderly people have an especially hard time maintaining balance while walking because the COM is outside of the BOS for 80 of the gait 28 cycle thus falls often occur while elderly people are walking Woolley Czaja amp Drury 1997 Therefore it is important to assess and evaluate gait in the elderly because any additional problems will increase their already high risk of falling For example elderly people who do not pick up their feet when they walk are more likely to t
166. or fall detection Worn around neck 1 http www lifelinesys com Other buttons are incorporated into a watch Pros May prevent long lie injuries Contacts a caregiver monitoring station or emergency response Device is discrete and doesn t draw attention to user Cons Does not detect a fall Not useful is user is unconscious Requires user activation 43 Device Sensor for Fall Detection Location Pros amp Cons SmartFall Cane T UN Buetooth Pereonal device Body sensors infrastructure networkas Contact pressure sensor http cs ucla edu alireza BodyNets08 pdf M Contact presume sensor ne a 5 T d hssr sensor fec dude gne Dans aco r bai eons aboar 2 Pressure sensors 3 axis accelerometer 3 single axis gyroscopes On cane not worn http cs ucla edu alireza Body Nets08 pdf Pros Alerts caregiver or monitoring service when user has fallen Uses subsequence matching Algorithm No set up required Does not restrict user s mobility Cons Not wearable False alarms Cane can fall and user not Bumps into something Rotation of cane from vertical to horizontal Does not detect fall before it http fallsaver net httn fallsaver net happens FallSaver 3 axis accelerometer Back of thigh Pros Minimal false alarms Tilt switch Easy to use Detects when user starts to rise from a si
167. ore unbalanced during tandem STS and all trials rated lt 4 were in tandem COP plots showed that the unbalanced STS tals had a Breater variation in the anterior posterior direction and a less compact COP than the balanced trial Figure 2 The subject s comfort level im the halanred tria was rated as a 4 and the unbalanced trial was rated as a 3 Acceleration plots Figures 3 and 4 had a positive acceleration and negative acceleration component corresponding to the forward lean that initiates the STS and backward lean that completes the motion respectively The average time t between a and a was all subjects Table 1 Far all SW trials of subject 1 t t 70sec and in all tandem trials t amp 3ser Therefore t of subject 1 Table 1 Results of SW and T trials ed aerator ceo EM deviation sec of variance Subject SW T SW w T m 023 035 0 037 045 0 01 2 071 3 63 sore 502 ors 034 5 059 031 2028 019 048 0 60 0 02 5 055 036 015 017 039 048 005 was equal to 0 50 sec such that t 0 60sec comesponded to an unbalanced STS IV DISCUSSION to detect an unbalanced situation during STS Unlike current devices on the market such as the iShoe the device cam assist elderly patients to assess their balance control independently The device also utilizes a universal anatomical marker the hip bone for its positioning This makes the device user frie
168. ppendix A The raw data were multiplied by 18mg digit and divided by 9 8 m s See Appendix J These data were used to plot the X Y and Z components of acceleration and magnitude of the acceleration in Microsoft Excel An offset of about 9 8 m s was observed in each plot due to gravity Therefore the average of the first 5 seconds of 102 acceleration data were subtracted from the entire data set in order to zero each plot The team then quantitatively compared the X Y Z and root sum squares plots of the unbalanced and balanced trials to identify differences The NetForce data files were imported into BioAnalysis software and plots of COP were obtained The COP data for the tandem trials were quantitatively compared to the COP of the shoulder width trials to determine if the subjects were more off balance and quantitatively represent their balance control The COP data were also compared to the subject s comfort level 4 4 2 Results of Preliminary Testing Results of preliminary testing showed that the STS activity produces an acceleration curve as shown in Figure 45 and contained a positive and negative amplitude of acceleration The positive amplitude of acceleration corresponded to when the subject flexed the hips to sway forward during the STS activity The negative amplitude of acceleration corresponded to when the subject extended the lower limbs to sway backward during the STS activity The peaks of these two amplitudes red and blue
169. ppendix C However the device was not tested and verified on an elderly user Therefore this section describes how the study can be continued to prove the device s viability for elderly users 8 1 PRELIMINARY TESTING WITH ELDERLY SUBJECTS The designers need to prove that the current device is able to accurately detect and notify an elderly user of an unbalanced STS and that it does not buzz during a balanced STS This section outlines the preliminary testing that needs to be conducted in order to customize the device for an elderly user 8 1 1 Materials and Methods The team needs to gather all materials needed for the study This includes at least 10 healthy elderly subjects age 70 years or older The health of the subjects should be evaluated by a physical therapist Subjects should score a 41 or above on the Berg Balance Test meaning they 148 can walk independently without an assistive device or supervision Internet Stroke Center 2010 In addition the SparkFun KinetaMap data logger will be used for collecting acceleration data Set the parameters of the KinetaMap according to the parameters outlined in Appendix B The AMTI AccuSway force platform and AMTI NetForce and BioAnalysis software will be used to record and analyze balance control data respectively A physical therapist should be present during all tests to rate the subject s STS performance according to how well they maintained their balance The force platform and a
170. provided in the Connecting over Bluetooth section of this user guide By default the config menu parameter in the configuration file is disabled In order to enable configuration menu you must edit the config txt file by plugging in a USB cable to the KinetaMap with the power turned on Then edit the config txt file to enable the config menu parameter More information on editing the configuration file can be found in the Configuration File section of this user guide 2009 SparkFun Electronics Inc All Rights Reserved Product features specifications system requirements and availability are subject to change without notice 8 KinetaMap GPS 08275 is a trademark of SparkFun Electronics Inc All other trademarks contained herein are the property of their respective owners KinetaMap UG 090401 ome eA B 7 APPENDIX C Client Interviews 1 The following is an interview conducted by team member Liz Tuite The interviewee is an 82 year old female Interview questions are bolded and interviewee comments are not bolded Yes I have I hate to tell you but I have Twice two bad falls Can you explain Yes I can once I was very careless Do you want a long story or short story Just as long as it needs to be we can cut it out Well I was up at Hampton Beach had a nice seafood festival I was going to meet Joe up there my son I was calling him on my cellphone and I couldn t hear him so I took a walk to get away from the music I
171. put into a comma separated variable file type with columns for acceleration in the X Y and Z axis This file type was opened in Microsoft Excel in order to process and analyze the data First magnitude was calculated via a root sum of squares method This can be seen below MQYacceleration E x y z Next Because the data from the accelerometer was outputted on a scale of 128 to 128 rather than it was necessary to convert the measured value to a more traditional acceleration measurement In order to do this the team implemented the function Accelerationm Magacceleration 18 1000 9 8 In order to calculate the acceleration in z Then in order to eliminate the gravitational offset value which was generally within a 15 range of 9 815 the team calculated the average of all the values preceding the STS motion e g the pre motion baseline and then subtracted this value throughout the range of the values Finally because the accelerometer was collecting data at 20Hz we calculated that each datapoint was collected at a 0 05 second interval and associated each data point with its respective time in seconds The data was then plotted in magnitude of acceleration in x versus time in seconds The curve was then analyzed for minimum value maximum value and time duration between minimum and maximum values using built in excel functionality For each subject the three parameters collected for each plot were statis
172. rch also shows that the signal of our COM is directly related to its acceleration in both the A P and M L directions Therefore we eliminated the methods of analyzing acceleration on the X Y or Z axis independently and decided to use the root sum of squares method in order to analyze acceleration of the COM in all planes of movement Using the root sum of squares method of analysis the team could look at one of two parameters of the graph amplitudes or time variations Research showed that as acceleration of the COM increased a person became closer to their stability limit meaning they became more likely to lose their balance and fall Pai amp Patton 1997 Studies also showed that the duration of the STS differs depending on the speed of the movement According to this research if a person accelerated quickly during the STS they became more unbalanced and the duration of the STS got shorter On the other hand if the person accelerated slowly during the STS they can maintain balance more easily and the duration of the STS got longer Therefore the team decided to analyze the time duration of the STS activity Figure 39 Method of Analysis Fast Fourier Route Sum of Transform Squares Time between peaks Amplitude of Peaks Figure 39 Choosing the Method of Analysis 97 4 3 5 Final Solution The two goals of the device were to proactively force user rehabilitation and strengthen the user s awareness of their bal
173. re 4a provides a larger range of motion and is used when the person is at their stability limit Both the ankle and hip strategy can also be combined together in order to maintain balance The third strategy is the stepping strategy number 3 in Figure 4a which is used when the person is about to fall and cannot maintain their balance without taking a step Pai amp Patton 1997 23 COP Acceleration b COP position P Midfoot a Figure 4 a Anterior Posterior Stabilization Strategies and b COP acceleration vs Position Shumway Cook amp Woollacott 2007 These three balance control strategies directly relate to the acceleration and position of the person s COP as shown in Figure 4b The graph shows the COP acceleration in m s on the y axis and the COP position on the x axis relative to the midfoot or the arch and the toe Curve 1 corresponds to the ankle strategy and is a balanced situation because the COP is maintained within the foot Curve 1 has a small amplitude and acceleration Curve 2 corresponds to the hip strategy and when the person is at their stability limit The COP is just outside the toe region and has a higher amplitude and acceleration than situation 1 The third curve is the unbalanced situation that requires the stepping strategy to maintain balance In the off balance situation the COP is far outside the foot region This situation also has the highest acceleration and amplitude From this graph yo
174. re are two ways to access the logged data which resides on the SD card The easiest way to retrieve the data is to plug a USB cable into the KinetaMap with the power on from a computer Doing this will create a Mass Storage Device drive on the computer The files on the SD card will appear in this drive These files can just be copied onto your hard drive or viewed directly from the SD card You can also modify the configuration file config txt from here 2009 SparkFun Electronics Inc All Rights Reserved Product features specifications system requirements and availability are subject to change without notice 5 KinetaMap GPS 08275 is a trademark of SparkFun Electronics Inc All other trademarks contained herein are the property of their respective owners KinetaMap UG 090401 ome eA B 4 KinetaMap User Guide 2009 04 01 You can also view the contents of the SD card from the configuration menu This method will be covered in detail in the Configuration Menu section of this user guide If you ve let the KinetaMap run a few times with the default firmware and default parameters then two types of file should be on the SD card a data file KinetaMapXXX csv and a NMEA file KinetaMapNMEAXXX csv There will be one pair of files for each KinetaMap session A CSV file is a Comma Separated Variable file These files can be opened in any text editor but are specially created for easy use with spreadsheet programs like Excel and Op
175. re eee Cr eene eee cha ner ne o e eene sh ehe eee ven give D dE de en ge verre 24 Figure 5 Balancing a pencil on the tip of the finger to illustrate balance control and a forward sway and by backward SWay ette eret reete ene a n ee eee o nice Ne eene Dre eere dre ee eee pr den 26 Figure 6 Factors that could contribute to a fall eese enne enne 28 Figure 7 Elderly participating in Tai Chi Rose 2005 eee en enne 32 Figure 8 Patent 7 150 048 Active protective garment Buckman 2006 see 34 Figure 9 Patent Application Number 2006 0195050 and Floor Vibration Prototpe Alwan Felder Kell amp Dalal 2004 2 0 3 te Dee rtt tre ie I pepe Eae icit Poe aee og pa eee Boa s CH RR ree Pea deen dace E 36 Figure 10 Patent Number 6 433 690 Elderly fall monitoring method and device Petelenz Peterson amp JETeeloioIRPAL PAP CE 38 Figure 11 Patent Number 5 515 858 Wrist held monitoring device Myllymake 1996 39 Figure 12 Figure 10 Patent Number 6 611 783 Attitude indicator and activity monitoring device Kelly amp Schoendorfer 2003 7 eue a derniere etate enean con anre Rena des coma aute EEEN areae ana Lee acu ASAE 40 Figure 13 Patent Number 6 997 882 Subject monitoring device and method Parker Fabeny Larson amp Monaco 2006 CES 41 Figure 14 Correct positioning of MyHalo chest strap and device adapted from Halo Monitoring 2009
176. repeated use of costly and environmentally damaging alkaline batteries present in some electronics though the need for proper disposal of the device and rechargeable battery pack would need to be addressed In the prototype all components in the design were RoHS compliant or compliant in regard to restriction of hazardous substances such as lead cadmium and mercury In this sense the device is no more damaging to the ecological system than any other small consumer electronic device such as a phone or mp3 player As can be seen above the balance control indicator device addresses a need for a low cost in home balance indicator for use by the elderly to analyze their balance and alert them of a lack of balanced state The device was required by team designed objectives to increase user awareness of his or her balance control and proactively force user rehabilitation both of which the designed device was capable in its ability to do The need for a low cost preventative device that can reduce or eliminate the occurrence of costly hospitalizations and needless diminishment of the quality of life of the elderly population due to falls is strongly presented in the staggering healthcare costs associated with elderly post fall hospitalization and the fact that these costs will 146 be rising exponentially due to a growing number of elderly within only the next ten years The issue of an increased elderly population and thusly the need to maintain t
177. rip over obstacles Other factors like walking patterns stride length response time and ankle and knee flexion also affect balance control and contribute to why elderly people fall Voermans Snijders Schoon amp Bloem 2007 Between the ages of twenty and sixty there is a 25 increase in response time which increases the likelihood that an elderly person is going to fall because they are unable to react as quickly to obstacles or changes in their COM Sturnieks St George amp Lord 2008 Response time is an important factor when determining whether a person will be able to recover from a trip or if they will fall The walking patterns of the elderly also greatly increase their risk of falling Elderly people have a smaller stride length when they walk and as a result have a slower walking velocity Winter D 1995 Walking velocity is particularly important when an elderly person encounters an obstacle or trips because the speed of forward rotation of the body related to the person s walking velocity Bogert Pavol amp Grabiner 2002 Elderly people also spend a longer time in the double support phase which is the phase of the gait cycle where both feet are on the ground Chong Chastan Welter amp Do 2009 Elderly people have a larger toe out angle reduced toe pressure and a higher horizontal heel velocity during heel contact These gait differences occur in the elderly because they are trying to maintain balance and limit the a
178. rticle CA6614434 html Trafton A 2008 July 16 Balance problems Step into the iShoe Retrieved April 5 2010 from MIT News http web mit edu newsoffice 2008 i shoe 07 16 html Uwe Maurer A R 2006 eWatch A Wearable Sensor and Notification Platform Pittsburgh PA Carnegie Mellon University Voermans N C Snijders A H Schoon Y amp Bloem B R 2007 Why old people fall and how to stop them Practical Neurology 7 3 158 171 164 Ward D S Evenson K R Vaughn A Rodgers A B amp Troiano R P 2005 Accelerometer Use in Physical Activity Best Practices and Research Recommendations Med Sci Sports Exerc 37 11 S582 S588 Ward D S Evenson K R Vaughn A Rodgers A B amp Troiano R P 2005 Accelerometer Use in Physical Activity Best Practices and Research Recommendations Medicine amp Science in Sport amp Exercise S582 S588 Winter D A Patla A E Prince F Ishac M amp Gielo Perczak K 1998 Stiffness Control of Balance in Quiet Standing Journal of Neurophysiology 80 1211 1221 Winter D 1995 A B C of Balance during Standing and Walking Waterloo Canadian Cataloguing in Publication Data Woolley S M Czaja S J amp Drury C G 1997 An assessment of falls in elderly men and women Journal of Gerontology Biological Sciences 52 2 M80 M87 Zheng Y 2006 Patent No 6 954 208 Zijlstra W Bisseling R W Schlumbohm
179. s grocery store and I see a spill P1 say young man you d better get that picked up right away because someone could fall and get hurt Do you know of anyone who you think could benefit from the device Sure I do personally my sister She complains of being lightheaded and unsteady in the morning when she gets up I ve told her you know watch your posture and things I think that s something she could benefit from before she gets out of bed in the morning you know see how she could use it cautiously getting out of bed because she s so unsteady C 3 So Why do you think it would help her do you think she would use it Oh I think she would you know it s not troublesome it s very easy to use I think it would be very good in nursing home use you know the people who could use it to make themselves aware You know people want to stay well so it would be nice for them to know if they re unsteady or have a chance of falling so maybe they sit until their head clears you know maybe get some fluids in them change their medicines or tell the doctor s what times this happens so they can adjust their medicine so they don t have drops in blood pressure or something Do you worked at St Pat s right How long did you work there for Oh I worked there for I don t want to say too long I really enjoyed the old people No I was there for at least 5 years and they ve improved a lot of things since then They have all kin
180. sed to collect the data Based on the data acquired a lightweight and compact form factor balance control indicator will be designed and tested that will be worn on a convenient location for the client and utilize a daily activity to strengthen the user s awareness of their balance condition by notifying the user of an off balance STS situation before a fall has occurred and proactively force user rehabilitation by requiring the STS activity in order for the device to function The device will need to be easy to use affordable reliable sensitive to changes in balance control and require no professional intervention to interpret its results The device should be placed in a 68 location that is easy to take off and put on and should utilize a repetitive daily activity so the device can easily be incorporated into the daily life of the user 3 4 PROJECT APPROACH To establish an approach the team identified specific subtasks under each step derived from the problem statement The subtasks were as follows 1 Use a wearable sensor to collect data during daily movement and situations when individuals can lose balance a Chose an activity during which a person can lose their balance b Chose a sensor that can be used to collect data during the activity c Chose a location for the sensor 2 Analyze the data a Identify characteristic parameters of the collected data b Choose specific parameter to analyze 3 Design a balance control indicator
181. sens 109 Table 15 Change in magnitude of acceleration due to tilt on X Y and Z axis sess 117 Table 16 Calculation of Subject 1 s time range essere enne nre 118 Table T 72 Cost OE DeyIee 2 ee o eee ta sea oe ated Ce he ee E Pee eae the YR REC ue Ee eee oo reckon 131 Table 18 Results of Device Testing sorier e t err cutest ren eran te repe inea ao Tarai iat 141 13 EXECUTIVE SUMMARY Every year one in three adults age 65 years and Tideiksaar 2002 costing the United States 20 to 30 billion health care dollars annually Centers for Disease Control amp Prevention 2009 Physicians often help high fall risk patients improve balance however many elderly may not have their balance control assessed before a fall occurs As a result there is a need for a device that detects and directly notifies the user of their balance control condition before a fall occurs Current technology is able to assess balance control however the devices are not catered to the elderly daily life or directly notify the user of their balance control Therefore the team designed a wearable device catered to the elderly population that is used only during the sit to stand STS rehabilitation activity The device requires the user to perform a rehabilitation activity and detects and directly notifies the user of an unbalanced STS thereby strengthening the user s awareness of their balance condition and forcing rehabilitation to improve b
182. shown in Figure 34 acceleration was chosen as the signal to monitor during the STS and a triple axis accelerometer was used as the sensor because it has high accuracy in monitoring balance control in accessible locations Gross Stevenson Charette Pyka amp Marcus 1998 92 Acceleration Pressure Tilt MEMS Accelerometer Figure 34 Choosing the Device Signal and Sensor 4 3 3 Choosing the Location The two goals of the project were to proactively force user rehabilitation and strengthen the user s awareness of their balance condition In order to force rehabilitation the device required the user to perform the STS activity To strengthen awareness the device monitored acceleration and detected and notified the user of an unbalanced situation abnormal acceleration before a fall Therefore the team identified the location to monitor acceleration during the STS that is the most accurate and sensitive to balance control and maintains comfort and ease of use of the device Figure 35 93 Sensitive to change in balance control Easy to put on amp take High off the Accuracy device Comfortable Figure 35 Factors location of device needs to fulfill The team identified that the trunk was the most sensitive location for monitoring acceleration of the body Gross Stevenson Charette Pyka amp Marcus 1998 Specifically the sternum hip and lower back shown in Figure 36 were the three locations
183. spect to gravity and orientation with respect to self generated movements It allows for the transmission of information that allows for compensatory movement and adjustment in body positioning Gray 2000 158 REFERENCES Allum J H 1999 Patent No 5 919 149 Alwan M Felder R A Kell S W amp Dalal S 2004 Patent No 10 550 157 American Academy of Health and Fitness 2010 Berg Balance Scale Retrieved April 10 2010 from American Academy of Health and Fitness http www aahf info pdf Berg_Balance_Scale pdf Analog Devices 2009 June ADXL345 Data Sheet Retrieved from Analog Devices Inertial Sensors http www analog com static imported files data_sheets ADXL345 pdf Anissimov M 2010 April 06 What is Proprioception Retrieved 2010 from wiseGEEK www wisegeek com what is proprioception htm Arduino 2009 July 14 Retrieved November 29 2009 from Arduino Duemilanove http arduino cc en Main ArduinoBoardDuemilanove Arduino 2010 April 1 Retrieved April 5 2010 from Language Reference http www arduino cc en Reference HomePage Arduino Forum 2009 October 10 Retrieved March 15 2010 from Hardware Interfacing http www arduino cc cgi bin yabb2 Y aBB pl num 1255602368 Arduino Reference 2009 January 9 Retrieved March 15 2010 from Define http arduino cc en Reference Define Aston R 1990 Principles of Biomedical Instrumentation and Measurementq New York New York Merr
184. ss Acceleration deflects the beam and unbalances the differential capacitor resulting in a sensor output whose amplitude is proportional to acceleration Phase sensitive demodulation is used to determine the magnitude and polarity of the acceleration POWER SEQUENCING Power can be applied to Vs or Vio vo in any sequence without damaging the ADXL345 All possible power on modes are summarized in Table 5 The interface voltage level is set with the interface supply voltage Voom which must be present to ensure that the ADXL345 does not create a conflict on the communication bus For single supply operation Vania can be the same as the main supply Vs In a dual supply however Voovo can differ from V to accommodate the desired interface voltage as long as Vs is greater than Vow After Vs is applied the device enters standby mode where power consumption is minimized and the device waits for Vio vo to be applied and for the command to enter measurement mode to be bit in the POWER_CTL register Address 0x2D In addition any register can be written to or read from to configure the part while the device is in standby mode It is recommended to configure the device in standby mode and then to enable measurement mode Clearing the measure bit returns the device to the standby mode Table 5 Power Sequencing The devicais complataly of but there pnl The device is on in mode but communication is and wil create a conflict on the communicatio
185. st me And now waking up in the morning I can t use that arm to get my blanket off me It s been really been debilitating and it makes me angry You know my doctor s tell me how healthy I am my cholesterol is fine my everything is fine no liver problems I m just so healthy I say yeah now if I could just learn to stand on my own two feet I d be fine C 2 Did you ever feel before your falls that you were off balance or you had balance issues No I don t think I was off balance because I ve always been a walker You know I ve walked a lot so I ve always been strong I would have been fine if I hadn t been on the phone and it was noisy too so I had to go away from the music So do you think in your case the device I was describing would have helped at all In my case Not really I think as you use better sense How many accidents do you think occur from people using poor judgement I always think about it now I think I ll always be careful where I walk where I step Before either of your falls did you know that falls caused so many problems in the elderly I did but frankly I don t always know I m as old as I am But I work nursing homes so I ve seen I am aware of slippery floors I ve always been because I ve known people who ve gotten hurt that way I ve always been aware of scatter rugs I never have scatter rugs because you can get hurt that way I m kind of aware of hazards even if l m in someone
186. t in a DC or step like component where the baseline of the signal will change and the sensing axis will reflect between zero and one g at rest The offset from the gravitational acceleration vector will be critical in determining the observed output where the output will be reflected by cos 9 81 m s gravity will be the angle which the sensor is offset from 56 directly measuring the reactive force of gravitational acceleration Because of this the accelerometer s output will be a nonlinear sinusoidal response to changes in orientation on each sensing axis It is important to note that an accelerometer will only sense changes in acceleration on the axis that it has defined as measurement axis and will not observe any changes if there is acceleration on an axis that has no sensor associated with it Cantilever Beam with Proof Mass Gas damped Chamber Gravitational Force acting on sensing axis Capacitive Plate Figure 21 The Effect of Gravitation Acceleration on an MEMs Accelerometer Digital Accelerometers output a serial stream of data rather than a voltage on each of the independent measuring axis They utilize pulse width modulation to determine the specific acceleration magnitude recorded This means that there is a square wave with a certain frequency and a varying duty cycle The duty cycle will be proportionate to the severity of acceleration acting on the sensing axis and obtaining an acceleration measur
187. t stated that a sound would be the best way to notify the user Therefore a buzzer was chosen as the best means of notifying the user Since the team chose to use the Arduino the device required a Lithium battery pack as its power source When determining the best means of turning the device on and off the team considered which type of switch would be least likely to get bumped and turn the device on or off unintentionally Therefore a slide switch was chosen The ADXL345 accelerometer was chosen to monitor acceleration because it is the accelerometer in the KinetaMap which was used in testing 109 In order to understand the gravitational effect on the accelerometer and the offset values we obtained as well as the change in offset before and after the STS motion tabletop testing was performed to understand how acceleration offset values changed as the device is tilted The first test performed was the z axis rotation test and was used to determine if changes in the orientation relative to gravitational acceleration will yield a different static offset in the device On each axis device was positioned with Z axis parallel to gravitation acceleration with positive end of the axis point upward on a table top Device was tipped forward so that the Z axis was offset by approximately 45 degrees relative to gravitational acceleration vector device was held in this position for some duration of time Then device was further tipped forward such th
188. t was greater than the TestLength When TestFinished is true if tan minus t44 is less than tg then RingBuzzer is true When TestFinished is true and tap minus t44 is greater than or equal to fg then RingBuzzer is false The final section of the program was the code for RingBuzzer or to sound the buzzer The RingBuzzer code was modified from an example code by Rob Faludi Faludi 2007 The function pinMode 4 OUTPUT was used to define digital pin 4 on the Arduino as the output pin for the buzzer The buzzer was set to buzz at 2048 Hz for 3 seconds or 3000 milliseconds 135 All four sections were integrated together to form the Duino Balance code which was used to analyze a STS activity to determine whether the user is balanced or unbalanced and sound a buzzer for 3 seconds if the user is unbalanced Figure 78 illustrates how the current device works for subject 1 user is off balance 0 60 sec Figure 78 Block diagram of device functions and specifications for subject 1 136 6 DESIGN VERIFICATION In order to verify that our device operated properly we performed STS tests using the same method that was used during preliminary testing except that our device was placed on the right hip and the KinetaMap on the left hip The subject sat in a chair with their feet on a force platform turned on both devices and stood up from the chair with their feet shoulder width apart The same test was then performed with their feet in tande
189. tects a fall after it has occurred Proactive balance control technology is not geared towards the elderly population e g Wii Fit or requires professional intervention e g iShoe The Wii Fit is not catered to the elderly population because the games require running jumping and fast movements which could put an elderly user at risk for injury Clark Bryant Pua McCrory Bennell amp Hunt 2010 The Wii Fit is also a video game interface which could be too complicated for an elderly person to set up and operate Therefore we identified the need for a device that indicates a user s balance control and risk of falling and is catered to the elderly population This chapter details the strategic design process used to determine client needs and wants objectives and constraints of the device The final sections discuss the methods that the team created to reach objectives of the design 3 1 CLARIFYING THE ORIGINAL PROBLEM STATEMENT The team was originally given the following problem statement by Professor Gielo Perczak and Professor Mendelson Design an early balance control device which can be used particularly by the elderly The first part of a project will involve data collection and analysis of signals during daily movement in particular situations when individuals can potentially lose their balance A sensor attached to the subject will be used to collect the data Based on the data acquired an early balance control indicator will be
190. ted to the analog pins of the Arduino using an I2C configuration This allows the Arduino to process the data collected by the accelerometer 123 http www sparkfun com commerce product_info php products_id 9156 Figure 68 ADXL345 Tri axis Accelerometer Sparkfun Electronics 2009 A buzzer is also included in our device to alert the user of an off balance situation The buzzer produces a loud sound at 2 04 kHz and is shown in Figure 69 where its size is compared relative to a U S quarter Sparkfun Electronics 2009 The Arduino processes the data from the sit to stand activity and determines whether or not to sound the buzzer http www sparkfun com commerce product_info php products_id 7950 Figure 69 CEM 1203 Buzzer Sparkfun Electronics 2009 Since the Arduino microcontroller board requires 7 12V of power another component of our device is a lithium battery pack because otherwise the device would require connection to a computer constant connection to a computer This component allows the device to be both wireless and portable The high capacity lithium battery attaches to the Arduino Duemilanove with screws through the holes in the Arduino board and fits below the Arduino board Figure 124 70 The PCB board that the battery is attached to is the same size as the Arduino board and the height of the battery is approximately 0 4 inches The battery can provide power to the device for up to 29 hours this woul
191. th default parameters Also if you want to revert to the default parameters just delete the configuration file and turn the KinetaMap on without a configuration file on the SD card a new one will automatically be created config menu parameter Default 0 The config menu parameter enables disables the configuration menu If the configuration menu is enabled the KinetaMap will not begin logging until a command is received from the bluetooth connection so be careful enabling this parameter The configuration menu will allow you to manipulate the SD card edit the system parameters and perform function checks on the GPS module and LIS302 accelerometer over a bluetooth connection If the config menu and send gps to bt parameters are both disabled the bluetooth module will be turned off This will significantly improve the battery life of the KinetaMap log gps parameter Default 1 The log gps parameter tells the KinetaMap if it should log GPS data or not By default the log gps parameter is enabled If the parameter is changed to 0 the KinetaMap will not save the GPS coordinates to the SD card log accel parameter Default 1 The log accel parameter tells the KinetaMap if it should log the acceleration values or not By default the log accel parameter is enabled If the parameter is changed to 0 the KinetaMap will not save the acceleration values to the SD card send gps to bt parameter Default 0 The send gps to bt parameter tells
192. than one axis will be effected X Axis Acceleration N lt wn E c 2 d i 9 o o o lt Time sec Figure 57 X axis acceleration during Y axis tilt 115 Y Axis Acceleration N lt v E c 2 i 2 o o o lt barman es time sec Figure 58 Y axis acceleration during Y axis tilt Z Axis Acceleration 6 4 e l a fe T E ENEBE ee ee pa 5 4 r ONERE TN 12 14 Time sec Figure 59 Z axis acceleration during Y axis tilt 116 Magnitude of Acceleration 30 time sec Figure 60 Magnitude of Acceleration during Y axis Tilt Table 15 Change in magnitude of acceleration due to tilt on X Y and Z axis Axis of rotation Maximum change in magnitude value x 0 9 m s Y 0 7 m s Z 1 4 m s Note that this greater observed change in Z axis data due to gravitational acceleration is consistent with the higher Og bias sensitivity as noted on the ADXL 345 datasheet as shown below in Figure 61 Sensitivity Change Due to Temperature 0 g BIAS LEVEL Each axis 0 g Output for Xour Your 0 g Output for Zour 0 g Offset vs Temperature for x y Axes 0 g Offset vs Temperature for z Axis Figure 61 ADXL34S Data Sheet 0g Bias Level 117 After the team identified the components of the device the design team brainstormed and researched how the components would interface to notify the user of an
193. the body s center of gravity COG the center of mass COM the center of 18 pressure COP and the base of support BOS The COM is a point on the body equal to the sum of the body s mass as shown in Figure 1 The COG is the vertical projection of the COM shown in Figure 1 The area of the feet in contact with the ground and the area between them when standing is referred to as the BOS The COP is located at the point of the vertical ground reaction force and represents the weighted average of the pressure on the surface area in contact with the ground Winter Patla Prince Ishac amp Gielo Perczak 1998 See Figure 1 Therefore when standing on one foot the COP is located within the surface area of that foot and when standing on both feet the COP is located somewhere between the two feet The COP is totally independent of the COM The ground reaction force exists between the ground and the surface in contact with the ground e g the person s feet and is an equal and opposite reaction to the force of the body weight Winter Patla Prince Ishac amp Gielo Perczak 1998 Figure 1 Location of the COM COG and BOS Shumway Cook amp Woollacott 2007 19 The weight of the body and the ground reaction force exert a moment about the same point of action The moment of each force is equal to the product of the force and the perpendicular distance from the force to the point When the moments due to body weight and ground rea
194. the user s balance and rewards them for strengthening their balance control This device differs markedly from the balance control indicator in that it does not integrate into the daily lives of the elderly utilizing a complex graphical user interface 142 that will be unfamiliar and difficult for many elderly subjects to comprehend Another device with bears some similarity to the balance control indicator is the iShoe pressure mapping insole which has not been released to market but remains in a patent pending status This device is worn underfoot and capable of measuring COP data and gathering balance control information from the user However this device differs from the balance control indicator in that the results are required to be interpreted by a professional and are far more complex than the simple balanced or not balanced logic implemented by the balance control indicator Because of the complexity of this device and the need for professional interpretation it would not be useful in a home situation to integrate into a user s daily routine Because of the simplicity of interpreting the balance control indicator s results as well as the ease of integration into the daily activities of the elderly it has clear advantages over the similar balance control indicating devices that currently exist on the market or in product development The device differs from wearable fall indicators and PERS systems in its core functionality Th
195. tically analyzed using an unpaired t test and the significance in time duration between balanced and unbalanced trials was assessed In order to assess the sensitivity of device attachment angle a similar sit to stand test was conducted and a similar means for data processing was implemented The minimum maximum and time duration values were assessed for similarity between the offset trials and the control 0 Offset for significant difference APPENDIX K IRB APPROVAL Informed Consent for Testing The Univenity of Science and Technology And Life Informed Consent Agreement for Participation in a Research Study Investigators Project Advisors Krystyna Gielo Perczak amp Yitzhak Mendelson WPI Students Kevin Goggins Amanda Martori and Liz Tuite Contact information BME Department WPI 100 Institute Road Worcester MA 01609 Tel 508 631 5100 Email falldetection wpiedu Title of Research Study Acquisition of Acceleration and Center of Pressure Data while Rising from a Chair Introduction You are being asked to participate in a research study Before you agree however you must be fully informed about the purpose of the study the procedures ta be followed and any benefits risks or discomfort that you may experience as a result of your participation This form presents information about the study so that you may make a fully informed decision regarding your participation Purpose of the study In this experiment we wil
196. ting See Appendix C 15 1 INTRODUCTION Every year one in three adults age 65 years and older fall and ten thousand elderly die each year as a result of falls Tideiksaar 2002 Of those who fall 20 30 suffer injuries that impair their ability to live healthy independent lives Centers for Disease Control amp Prevention 2009 These injuries include moderate injuries such as bruises and arm fractures and severe injuries such as hip fractures and head trauma In fact most fractures among the elderly and traumatic brain injuries are caused by falls Centers for Disease Control amp Prevention 2009 In addition falls in the elderly may cause feelings of increasing frailty fear and stress ultimately leading to anxiety during activities of daily living ADLs e g getting out of bed Of those who fall 50 avoid performing ADLs because they fear additional falls Tideiksaar 2002 Furthermore falls cost the United States 20 to 30 billion health care dollars each year Services 2007 and this amount is expected to increase with the increasing elderly population By 2030 80 million people will be elderly an approximate 43 increase since the year 2000 and by 2020 total indirect and direct medical costs of falls may reach 54 9 billion dollars Centers for Disease Control amp Prevention 2009 The financial burden and effects of falls on the quality of life of the elderly cause the need for a cost effective solution that minimizes th
197. ting falls is identifying of a person s risk of falling as soon as possible Zijlstra Bisseling Schlumbohm amp Baldus 2010 This can be achieved by assessing and proactively monitoring balance control One way to proactively monitor balance control is through the use of sensors placed on the body Sensors such as accelerometers gyroscopes pressure transducers or strain gauges are portable can be used in a person s home and are not very expensive Janssen Kulcu Horemans Stam amp Bussmann 2008 2 5 1 Accelerometry An accelerometer measures acceleration relative to freefall and is used in many of the fall detection applications Accelerometers can measure acceleration on one axis two axis or three Accelerometers are micro electro mechanical systems MEMs and behave as a series of small dashpots damped with a gas Figure 20 As the spring mass system moves due to an external acceleration being applied the electrical impedance of the system changes These changes are outputted digitally or through an analog signal and must be processed to be interpreted in a circuit Some considerations taken in choosing an accelerometer for device design include sensitivity for analog accelerometers this will mean amplitude of the change of the output voltage relative to g s of acceleration maximum measurable acceleration and number of axis on which to measure acceleration Accelerometers can be combined with gyroscopes which in the case of
198. tive in the column A 0 indicates that objective in the column was more important than the objective in the row An x was given when comparing an objective against itself Total points that each objective received were summed in the last column of the table The objective with the highest total was the most important The first table Table 3 evaluated objectives that contributed to the marketability of the design while the second table Table 4 evaluated objectives that contributed to the safety of the design 79 Table 3 Pairwise Comparison Chart of Marketability Objectives Marketability Easy to ro PCE Durable Aesthetically Comfortable put pay to elderly So ad Manufacturable appealing on take operate dail efficient of life Durable X Aesthetically appealing Comfortable 4 Easy to operate Cater to elderly daily life Cost efficient Manufacturable SS pe entes temm ESNNENECSEXESZESIMOSM a ESSERI EE TT E t t SUC a a o a ea a E E Eas eG lie eit FS ERES EN USC REN ERE ERR RN According to the results of the marketability PCC the order of importance of these objectives from most to least important was as follows 1 Manufacturable 2 Easy to put on take off 3 Easy to operate 4 Comfortable 5 Adaptable to elderly daily life 6 Durable 7 Cost efficient 8 Aesthetically appealing 80 Table 4 Pairwise Comparison Chart of Safety Objectives Sensitive Minimal to interaction Balance
199. tness 2010 Standing on one foot and standing in tandem were eliminated because they can be dangerous for the elderly user and are also not repetitive daily activities STS was chosen as the activity because it was sensitive to balance control Gross Stevenson Charette Pyka amp Marcus 1998 and repetitive in daily life Figure 31 89 Activity Berg Balance Test Stand on one foot Figure 31 Choosing the Device Activity 4 3 2 Choosing the Signal and Sensor The two goals of the device were to proactively force user rehabilitation and strengthen the user s awareness of their balance condition In order to force rehabilitation the device required the user to perform the STS activity which is both a daily activity and a rehabilitation technique To strengthen awareness the device detected and notified the user of an unbalanced situation before a fall To do this the team had to identify an appropriate signal that was sensitive to balance control and maintains high accuracy The sensor that monitors this signal must be positioned in an accessible location to maintain ease of use of the device Figure 32 90 Sensitive to balance control Accessible High Location Accuracy Figure 32 Factors signal and sensor need to fulfill The team identified three signals that were currently used to monitor balance control pressure under the foot acceleration of the body and tilt of the upper body at the hip as
200. tronics 2010 Lifeline Standard Medical Alert Service Retrieved April 10 2010 from Phillips http www lifelinesys com content lifeline products how lifeline works Life Alert Emergency Response Inc 2010 Retrieved from Life Alert http www lifealert com index html Lindsay J 2008 7 31 Astronaut technology could prevent elderly falls Retrieved 10 13 2009 from www msnbc msn com http www msnbc msn com id 25949023 Lockhart T E Smith J L amp Woldstad J C 2005 Effects of aging on the biomechanics of slips and falls Human Factors 47 4 708 729 Lord S Sherrington C amp Menz H 2000 Falls in Older People Risk Factors and Strategies for Prevention Cambridge United Kingdom Cambridge University Press Lorincz K Chen B r Patel S amp Welsh M 2008 Mercury A Wearable Sensor Network and Platform for High Fidelity Motion Analysis Retrieved from Havard Sensor Networks Lab http fiji eecs harvard edu Mercury Maker Media 2009 Maker Shed Retrieved December 4 2009 from 9V Wall Adapter Plug for Arduino http www makershed com ProductDetails asp ProductCode MKSF3 Mao D W Hong Y amp Li J X 2006 Characteristics of Foot Movement in Tai Chi Exercise Physical Therapy 86 2 215 222 Mathie M J Coster A C amp Lovel N H 2004 Accelerometry providing an integrated practical method for long term ambulatory monitoring of human movement Physiological
201. ttery voltage is read every time the accelerometer is queried The next 6 fields are for GPS data and the names should explain their function GPS data is only logged when a new message is received if there is no new data than the fields are left blank Also the altitude is only logged if GGA messages are enabled Altitude will not be logged when an RMC message is received The last field in the data file is the Type field the Type field indicates if the GPS message received was a GGA message or an RMC message The NMEA file is much simpler The only data saved to the NMEA file are the GPS messages While the data file contains all of the relevant data it is not easy to use the position data in it s format If you want to find a path you must use a scripting program to parse out the GPS data in order to make it meaningful To make it easier to use the GPS data the KinetaMap has the option of logging data to a NMEA file The file contains only the GPS data received from the GPS module This file can be directly imported to websites like GPSVisualizer http www gpsvisualizer com to view position data The Configuration File The configuration file is used during the powerup sequence to load the system parameters for the KinetaMap To 2009 SparkFun Electronics Inc All Rights Reserved Product features specifications system requirements and availability are subject to change without notice 6 KinetaMap GPS 08275 is a trademark of SparkFu
202. tting position Wireless Cons Does not detect a fall Beeps every time user starts to stand up Going to keep beeping until user sits back down 44 Device Sensor for Fall Detection Location Pros amp Cons Floor Vibration Sensor E http fnarc med virginia edu pdfs library ICTTA fa Piezo transducer On floor not worn http marc med virgini a edu pdfs library ICTT A fall pdf Pros Automatically detects a fall once person is on the ground No false alarms Alerts a caregiver of a fall Cons Does not detect a fall before it occurs Lack of portability Only useful for one room UVirginia TEMPO X Axis in Z Axis in Z Plane Plane http marc med virginia edu pdfs library ICTTA_fa 3 axis accelerometer 2 axis gyroscope Z axis gyroscope Center of trunk and front of thigh http marc med virginia edu pdfs library CTTA_fall pdf Pros Able to recognize different dynamic and static positions Low cost Cons Detects fall after it occurs Difficulty determining whether someone is getting into bed or falling against a wall into a seated position Attached to user with tape Smart Coat http www sci IN om science ob Ar ticleURL amp _udi B6V424NMC8781 amp _user 74021 amp _rdoc 1 amp _fmt amp _orig search amp _sort d amp _doc anchor amp view c amp _searchStrid 1140042312 amp _ rerunOrigin google amp _acct C0000058
203. ttp www sparkfun com commerce product_info php products_id 9156 Sparkfun Electronics 2009 Sparkfun Electronics Retrieved December 4 2009 from Buzzer PC Mount 12mm 2 048kHz http www sparkfun com commerce product_info php products_id 7950 Sparkfun Electronics 2009 Sparkfun Electronics Retrieved December 4 2009 from Arduino ProtoShield Kit http www sparkfun com commerce product_info php products_id 7914 Stevens J A 2005 Falls Among Older Adults Risk Factors and Prevention Strategies Falls Free Promoting a National Falls Prevention Action Plan pp 3 18 Sturnieks D L St George R amp Lord S R 2008 Balance disorders in the elderly Clinical Neurophysiology 38 6 467 478 The National Council on the Aging 2005 Falls Free Promoting a National Falls Prevention Action Plan Retrieved April 2 2010 from Healthy Aging Programs http www healthyagingprograms org resources FallsFree_ReviewPaper_Final pdf Tideiksaar R 2002 Falls in Older People Prevention amp Management Baltimore Health Professionals Press Inc Tinetti M E Lui W L amp Claus E B 1993 Predictors and Prognosis of Inability to Get Up After Falls Among Elderly Persons Journal of American Medical Association 269 1 65 70 Torrence R 2008 December 1 MEMS based inertial sensor is not your grandfather s gyroscope Retrieved October 16 2009 from Electronics Design Strategy News http www edn com a
204. u can see that as the acceleration increases the loss of balance increases and thus the person has less time to react Shumway Cook amp Woollacott 2007 In addition to amplitude of acceleration reaction time is an important component of balance There are two different reaction strategies the reactive control and proactive control 24 strategies The reactive control strategy occurs as a result of a loss of balance or the COP moving outside of the BOS The second strategy is the proactive control strategy which occurs in anticipation to an off balance situation The reaction time involved in a fall is small and thus it is important to utilize both the proactive and reactive strategies The proactive strategy occurs when a person becomes used to a routine or occurrence and is able to alter their movements based on anticipation Shumway Cook amp Woollacott 2007 2 1 4 Example of Balance Control Balancing a Pencil Balancing a pencil on your finger is an example that illustrates the concepts of balance control When balancing a pencil at the tip using one finger the ground reaction force is the force of the finger pushing up on the pencil The COG of the pencil X is located far above the tip Figure 5 close to the middle of the pencil In order to balance a pencil at the tip with one finger the finger needs to push up at exactly the same point as the COG If the point of reaction force R of the finger does not push
205. uino 127 One lead of the CEM 1203 buzzer was connected to digital pin 4 on the Arduino Microcontroller Board and the other was connected to ground A 100Q resistor was connected between pin 4 and the buzzer because the current from the Arduino was too high for the buzzer to handle The proper resistor value was determined using the equations and calculations below Maximum current for Arduino 40mA Coil resistance 42 Q 63 Arduino Voltage 5V Ohm s Law V IR V pim R c oqi9Atzqi m ayy ae ARN I Actual Max 119mA 40mA 79mA 79mA gt 40mA Since the current is greater than the Arduino can handle we added a resistor If we add a 100Q resistor to the circuit R 420 1000 1420 VY _ 903524 35 2mA R 14220 SR With the 100 Q resistor the current was 35 2 mA which is less than the maximum current for the Arduino of 40mA The 100 Q resistor provided enough resistance to activate the buzzer 128 Lastly the rechargeable battery pack was connected to the 5V power pin and ground pin on the Arduino Microcontroller Board The battery pack provides enough power to operate the device without being connected to a computer The final assembly of the device is shown in Figure 75 The battery pack was also connected to a slide switch for turning the device on an off Some foam was also added inside the box on both sides of the device The foam was used to prevent the device from moving within t
206. unbalanced situation First sample code was used to verify that the buzzer could make a sound and that the accelerometer could collect data In order for the device to detect the maximum and minimum peaks the team chose to design the device such that it did not look at the baseline noise This eliminated the possibility of the device identifying a peak in the baseline as a maximum or minimum peak Therefore after a certain baseline was reached the device would start looking for the maximum and minimum peaks The following Figure 62 is the basic block diagram of the designed algorithm start peak grin n detection find and find and turn onthe SEU when record record device threshold threshold maximum minimum acceleration peak is detected acceleration Figure 62 Block diagram of algorithm The buzzer would sound if an unbalanced situation occurred and reset if a balanced situation occurred To specify the difference between subject 1 s unbalanced and balanced STS the time range for a balanced STS and unbalanced STS was calculated by adding and subtracting the standard deviation to and from the corresponding average time Subject 1 s time ranges are shown in Table 16 Table 16 Calculation of Subject 1 s time range Subject 1 Average Standard Time Range Time s Deviation s Tandem 0 45 0 16 0 29 0 61 Shoulder width 0 85 0 20 0 65 1 1 118 The team decided that above an expected time
207. up at exactly the same point as the COG then the weight of the pencil W creates a moment equal to the product of the weight of the pencil and the perpendicular distance d distance from point X to point R in Figure 5 The position of COG in relation to R causes the pencil to sway forward or backward If the COG is located in front of the reaction force a clockwise moment is created Figure 5a causing the pencil to turn forward if the COG is behind the reaction force a counter clockwise moment is created Figure 5b causing the pencil to turn backward Balancing a pencil at the tip using one finger is difficult because the COG and point of the reaction force i e the finger are located far apart 25 Figure 5 Balancing a pencil on the tip of the finger to illustrate balance control and a forward sway and b backward sway 2 2 RISK FACTORS OF FALLING There are many risk factors that are a result of normal ageing that can cause an elderly person to be more susceptible and likely to fall than a younger adult as shown in Figure 6 Risk factors can be classified into two main categories extrinsic factors and intrinsic factors Extrinsic factors are present in the environment or the person s home and include factors such as stairs uneven terrain loose carpet poor lighting and wet bathroom tiles Intrinsic factors are present within an individual and generally relate to physiological factors such as balance vision pr
208. were balanced and tandem trials were unbalanced Figure 47 The unbalanced tandem trials showed a large variation of the COP green which showed that the subject s COM moved across a large area On the other hand the balanced shoulder width trails showed less variation as seen by the more compact black circle This showed that the subject s COM did not move or sway across a larger area and the subject maintained their balance Unbalanced Balanced t LLI n Position X i Figure 47 COP plots of tandem green and shoulder width black trials The average times between the positive and negative peaks are shown for each subject in the bar graph Figure 48 and data are listed in Tables 11 and 12 The average time of shoulder 105 width trials was greater than in tandem trials for all subjects tested and was significantly different in subjects who performed 10 shoulder width and 10 tandem trials P values are listed in Table 13 The average comfort ratings for each subject Tables 11 and 12 were greater for shoulder width trials than tandem trials Average Time s 0 oo 0 o 0 P 0 N of trials p value 0 01 0 01 0 02 0 02 0 10 Mal lt 0 01 0 12 5 0 23 H Shoulder Width E Tandem Figure 48 Average time between peaks RL 1 2 3 4 5 6
209. will open a Mass Storage Device The log files are located on this drive and can be downloaded to your hard drive You may also open the config txt file to edit the configuration parameters Just save your changes on the SD card and don t alter the file name The edited parameters will be loaded the next time you turn the KinetaMap on Don t worry about corrupting the configuration file a corrupted config file will be detected by the firmware and handled Power up Sequence Turn the KinetaMap on by flipping the rocker switch to the position Once the device is powered it will go through the startup sequence First the KinetaMap will load the configuration settings from the config file config txt which is located on the SD card The firmware checks to see if the config file exists If the file is not on the card then a config file with default parameters is loaded If the firmware finds the config file then the file is read and the parameters are loaded However if the firmware reads the config file and detects that invalid settings have been saved or if the file is not in the correct firmware than the existing config file is deleted and a new one with default parameters is created While the KinetaMap is loading the parameters the tri color LED on the front of the KinetaMap will be solid GREEN More information on the config file and it s parameters will be discussed in the Configuration File section of this user guide After the parameters
210. would enable the device to detect an unbalanced situation and directly notify the user These steps included 1 Detect the maximum peak of the positive amplitude 2 Detect the minimum peak of the negative amplitude 3 Calculate the time difference between the maximum and minimum peaks 4 Notify the user if a specified time is calculated 108 Given these specific steps the team researched and brainstormed electronic components that would enable the device to achieve these functions A Morphological chart was developed to organize necessary functions and potential means Table 14 Morphological Chart Electronic functions and means Functions Means 1 2 3 Detect peaks Calculate time Arduino Microcontroller VEX between peaks Notify user Buzzer Vibration LED Power device Rechargeable batteries Throw away Lithium batteries Battery pack Turn on off Button Toggle switch Slide switch Monitor acceleration ADXL345 accelerometer To detect the peaks and calculate the time between the peaks the device needed to contain a microcontroller Two microcontrollers were identified the Arduino Duemilanove and the Vex The Arduino Duemilanove was chosen because it has open source software so there were many resource materials example codes tutorials and other reference materials online In order to notify the user of imbalance the device could buzz vibrate or light up During an interview an elderly clien
211. young subjects Grabiner Pavol amp Owings 2002 Kinetic differences present in the elderly are reduced ankle power generation increased hip extension moment during swing phase Grabiner Pavol amp Owings 2002 reduced toe pressure and slower generation of joint moments Lockhart Smith amp Woldstad 2005 The famous slogan I ve fallen and I can t get up is both a fear and a reality of many elderly people Life Alert Emergency Response Inc 2010 Studies of community dwelling elderly people have shown that approximately 50 of fallers including those who have not suffered any injuries as a result of the fall cannot get up on their own after a fall Tinetti Lui amp Claus 1993 The number of people who cannot get up after a fall increases significantly to about 80 over the age of 90 years old Fleming amp Brayne 2008 The inability to get up after a fall can lead to more serious injuries like dehydration hypothermia pneumonia pressure sores muscle damage and increased fear of falling Lord Sherrington amp Menz 2000 Lying on the 30 ground for extended period of time also leads to post fall anxiety syndrome People with post fall anxiety syndrome alter or avoid daily activities because of the constant fear of another fall When someone avoids daily activities or changes their gait because of a fear of falling their muscles become weaker and atrophied and this leads to an abnormal gait and as a result incre

Download Pdf Manuals

Related Search

Related Contents

IBM Brocade 16Gb FC 1-port HBA  Mapefer  KNA-DV3100    Sanyo DS25390 TV  NAVgraph NAVgraph User Manual  Pipe connections  [0] [Impianto, Inverter]  MD Building Products 78907 Installation Guide : Free Download, Borrow, and Streaming : Internet Archive  Document  

Copyright © All rights reserved. Failed to retrieve file