Home

fulltext - DiVA Portal

Contents

1. 28 2 6 6 Special Parameters uude t HO e ER e rre 29 2 6 7 Universal Gait Par melers eiiis cotra e oe PUT edu 29 2 628 dEea t res EXIFOCIOT s d et eet ec sette ce vetere te ous oet 29 2 6 9 Techniques Used for Feature Extraction eese eene enne 29 2 7 BINDINGS ertet tert re rper Pp e pe RR RE 30 3 HARDWARE CUOI NE e ERE Vene e oaa E REESE ERE en RE 32 3 1 SHIMMER PLATFORM ete ds 32 3 1 1 SHIMMER Base board ekeren tates se e e aea A TE EENE ES ereb Sy 33 IL Daughter board 4 sog Sun ete enge ute pedum eee 34 MicroSD Card 2GB x sane een antt E E Dee tet e eec ts 34 Jd 4 Teads edd e HR ERR HE 34 ceu RD e E ede 35 SLG USB hg ss HR e RII HE 35 BUT QGane Multi Charger sine ehe ned HP Re He REP 35 3 1 8 Straps and bands isses e eerte teg 35 S0 YBiueiOOtli sse tete inte Ium dei 35 3 2 ungbpjNe EH 36 MES RESUME 37 4 1 BSL430 BSL STANDS FOR BOOTSTRAP LOADER 212 2 4 nennen ennt 37 4 2 SHIMMER CONNECT pU EUREN E E EE NEAR EUN ERE Eds 38 4 3
2. EET EET eee TUE eee PUE 10 1 7 PROBLEM STATEMENT OR ee EE PETER UR 10 1 8 RBSEARCHOQUESTIONS 5 idee e ed EL 11 1 9 EVALUATION OF GAIT 515 11 1 10 SHORT DESCRIPTION OF 5 11 1 11 MOTIVATION etcetera dede 12 1212 METHODOLEOGY ERR ee E E 12 1 13 RESULTS 13 1 14 TECHNICAL 13 1 15 ADVANTAGE AND FUTURE 4 404440 n nen nennen nnn nn nnn 13 1 16 OWVERALDE rettet dee dost retur eddie dete o eet dede eee tet dede 14 1 17 ORGANIZATION OF CHAPTERS c esineen ee e n n n nn enhn nnn n n e una a nn 14 1 18 MAJOR ACHIEVEMENTS eene n n e nnn nnn n n n a e an n n a n usn n p 14 1 19 RELATED WORK t eese teste rece ede tei en 14 2 PARKINSON S DISEASE ic ciscscsscssecccvscsscvsvodecssovssesscsnescesssscsccedecdeesdenstecscsdeedeseevesessesesdsossexcties 17 2 1 17 2 2 ETIOLOGY OF PARKINSON S DISEASE eceeeeeeeeen n
3. Message Gait is normal End Comments Std Standard Deviation Threshold Value 200 calculated during various measurements CSV File m 25 values because when Sampling rate 100 means getting 100 values sec whereas a patient falls 1 4 of a second i e 25 File m 3D Array i e int temp 25 Best Feature Algorithm is flexible It means that user can set Sampling rate and Threshold value 74 Appendix F Code Offline Gait Analysis function void offlineFallManag fileName Fs Threshold1 Threshold2 siren siren fs wavread SIREN 1 WAV SIREN 2 WAV siren2 siren_fs wavread SIREN 2 WAY filedata csvread circlegait csv 1 1 arr_counter 1 Fs 100 processed_buff_size round Fs 4 for loop1 1 processed_buff_size length filedata processed_buff_size x filedata loop1 loop1 processed_buff_size 1 1 filedata loop1 loop1 processed_buff_size 1 2 z filedata loop1 loop1 processed_buff_size 1 3 std_x std x std_y std y std_z std z std arr arr counter 1 std x std arr arr counter 2 std y std arr arr counter 3 std 2 arr counter arr_counter 1 Thresholdl 250 Threshold2 270 if std x gt Threshold1 std y gt Threshold1 II std z gt Threshold1 display Alarm ON wavplay siren siren fs break end if std x gt Threshold2 Il std y gt Threshold2 Il std z gt Threshold2 display Alarm ON
4. 39 Figure 12 Real time EyesWeb Patch Display eese 40 Figure 13 Real time EyesWeb Patch Save into 41 Figure 14 Data Acquisition to Track 50 Figure 16 Data Acquisition through 5 53 Figure 17 Two Accelerometers on Waist eese neret nnne nennen enne 53 Figure 18 Real Time Algorithm Simulation eese 62 List of Graphs Plots Plot 1 Consciously slow gait during turning becomes more stable sess 58 Plot 2 SitChair2Chair Gait Ploting in MATLAB esee 58 Plot 3 FastGait in EyesWeb showing higher acceleration 59 Plot 4 CircleGait in EyesWeb showing higher acceleration esse 59 List of Tables Table 1 Description of SHIMMER enne 33 Table 2 Step count Number of steps taken by the Subject sse 56 Table 3 Experiment 01 shows 6 different types of gaits with higher std deviations after each 25 samples x3 ee See e a ei e 57 Table 4 Experiment 02 shows 6 different types of gaits with higher std deviations after each 25samples 5 eei emi esti e E erbe ert petet 60 List of Appendices Appendix A AGB Test ode dca su eeu duit exe SY ERU e renes 70 Appendix B Pseudo Code RR HERO RO EORR ERN ed
5. 1000 50 4 0 0 50 100 150 200 250 300 350 400 450 500 Figure 17 Real Time Algorithm Simulation No hurdle in the path no wall no turning aside line of sight is recommendable The figure 18 is a simulation diagram of the real time gait analysis algorithm for tracking movement of patients For example if Threshold 1 400 and Threshold 2 600 are set it means that an unbalanced movement with standard deviation 400 600 will be received and the corresponding warning emergency alarms will be activated In this way practitioners clinicians and health providers can employ this method to actively understand and respond to the movements of the patients The proposed system may generate a high rate of false alarms in such condition push button can be designed to cancel false alarm as suggested by different authors and one such is 4 The comparison between our proposed system and the results from other systems is bit difficult One reason is that our study is based on the experimentation and calculation of one lady with PD Secondly 4GB test is designed the first time to fulfill our own specific requirements This test can be helpful for similar method of taking measurements and open to the public along with the MATLAB code to design and modify the system as required 62 So a practitioner or clinician may understand our method and he she may need our test and results for his
6. 43 5 1 X Existing Scales Tests for Gait Assessment Followings are some of standard scales and tests for measuring movability of persons who are able to walk likely to fall checking cognition level and related physical examinations 5 1 1 Modified Fall Efficacy Scale is a 14 questions one page form to know about daily routine life of a PD patient Unlike Fall Efficacy Scale MFE includes more questions related to outdoor activity range from 0 10 not confident to completely confident Some questions are related to know about the habit of meals walking patterns sit stand up on chair and bed dressing bathing light shopping n public transport 5 1 2 Berg Balance Scale BBS BBS is a 14 items scale of 15 feet walkway scoring 0 4 time 15 20 minutes to measure the balance of the patient while performing different tasks A recent study conducted in Finland shows that a change of 8 points in the BBS is required for a genuine change in the balance between two assessments Some of the items include sit stand with and without arm support and turning 360 degrees etc 5 1 3 Unified Parkinson s Disease Rating Scale UPDRS UPDRS is the most commonly used rating scale for clinical assessment based on an interview and clinical examination It consists of 5 parts or sections such as Section I Assessment of Mentation behavior and mood Section Assessment of Activities of daily life ADLs including s
7. But after implementing the real time algorithm in Matlab thresholdl 2 are changed with little difference So it is up to the user to set the threshold values after calculating standard deviation at different maximum points Algorithm suggests setting both threshold values as under Threshold gt 240 6 1 Warning Alarm Threshold gt 320 6 2 Emergency Alarm Both above 6 1 and 6 2 can be changed depending upon the movement of the patients Algorithm is flexible yet sensitive and can be validated through 4 i e can be adjustable to 61 take minimum allowed threshold value ranging 200 300 Some important parameters for the algorithm are given below Max Min threshold values 0 2000 Capture duration 120 seconds approx Saving file with name ShimmerData dat start is basically similar to push button to be used in the future device Maximum Distance range good distance at 12 15 steps away Maximum Distance range in general 25 steps Simulation of algorithm is shown in the Figure 18 below _ EX gt ee File x Threshold 1 400 Threshold 2 600 4 4 Freerun d 4000 T T T T T T T T T Accel X Accel Y Accel Z 3500 3000 I 2500 m P Pe Me ten P Pe SESS Pod s anh cates PSI SIN nmm P P mA AA e S e S Vintner mS AP SR IS Sr NIS I eye es 1500
8. ii Carbidopa or Lodosyn is another medicine used for peripheral metabolism of Levodopa and it allows crossing the blood brain barrier So Carbidopa is used instead of Benserazide in the US while both have same function in treating Parkinson s disease Entacapone has almost the same effects iii Other Dopamine agonists include bromocriptine pergolide pramipexole ropinirole piribedil cabergoline apomorphine bromocriptine and lisuride Agonist is a medicine chemical that can combine with a receptor on a cell to produce a physiologic response Detail on Appendix D 2 4 2 Surgical Treatment 22 Surgical treatment is prescribed by physicians when effect of medications is greatly reduced or no effect at all For treating rigidity there is a surgical method called Pallitomy It is also to note that Thalamotomy is used for treating the symptoms of tremor There is another surgical technique with the name Deep Brain Stimulation which will also be discussed later on down in the report in chapter 6 Deep Brain Stimulation During this surgery special electrodes are inserted into the targeted place in the brain This complex surgical procedure is performed using MRI and neurophysiologic to make sure that electrodes are implanted in the right place After that a special device called an impulse generator is used in order to provide electrical impulse to the part of the brain involved in motor function A controller to ON OFF and battery with
9. timing 3 5 years are provided with the device Detail on Appendix D 2 4 3 Therapies Speech Therapy In order to improve speech and voice impairment LSVT Lee Silverman Voice Treatment is most commonly and widely used speech therapy This therapy is also proved good for improvement in facial expression swallowing and talking Neuroprotective Therapy Based on the theory that 300 400 thousands dopaminergic neurons can be protected somehow from early death using some potential Neuroprotective agents if identified in time Research is being carried out to get better suitable results Gene Therapy Recent clinical successes have opened chances to replace or correct infected mutated genes in different diseases including PD for a cure Another good option is Fetal cell transplantation There are some other therapies such as speech therapy occupational therapy that are also helpful for independence mobility and improved motor symptoms 2 4 4 Diet amp Nutrition Parkinson s disease can have negative impact on digestive track due to the reduced strength of associated muscles Gastroparesis is one such symptom in which food stays in the stomach for more than two hours So a light balanced nutritional easily digestible small frequent meals walk after each meal is recommended to take without causing any stomach disorder Mediterranean diet can be introduced in the patients with PD 23 lii Excessive and combining t
10. Std m Threshold2 Yes No Yes Std m gt Threshold1 No di B END 5 gt Print Emergency Alarmi Print Warning Alarm Algorithm is flexible It means that user can set Sampling rate and Threshold value 72 Appendix D Parkinson s disease Symptoms Tremor There are many types of tremors and all cannot be associated with PD Like an Essential tremor is wrongly associated with Parkinson s disease because it does not affect during the rest The patients who have a tremor show less frequency in FOG However in some cases tremor is never observed 14 Mostly tremor develops later on when disease makes progress Almost 30 PD has a tremor which causes trembling of limbs jaws and uncontrollable movement of the body 39 Bradykinesia General muscle weakness rigidity and tremor may contribute to but not necessarily Sometimes slowness may occur during initiating movement repeating of motion and rapid necessary fine actions such as writing In short in Bradykinesia small stepping shorter arm swing and stooped forward are general symptoms Rigidity Freezing of gait is more complicated stages associated with rigidity Bradykinesia Resting Tremor and Rigidity are often visible Muscle blocking is visible as tremor and Bradykinesia while this blocking during muscle contraction and relaxation does not allow the smooth movement of elbow knee joint and other body joints So in a result musc
11. s disease include Short period memory loss forgetting and revising things Quick change in emotions such as anger sadness and excitement Hallucinations amp Delusion 21 One of the earliest signs of PD is losing some or all of sense of smell Sometimes it can occur many years before disease is diagnosed Autonomic Nervous System It is an automatic nervous system which is not under control of any person It is the nervous system that controls your autonomic functions such as heart breathing muscles glands amp urination etc Patients with PD may experience some or any of these disorders Other autonomic problems are constipation erectile dysfunction in men reduced sexual interest in women and low blood pressure with dizziness becoming fainting profuse sweating swallowing disorder and drooling excessive saliva 2 4 Treatments There is not an absolute cure for the patient with PD but treatment through some medication surgery therapies and multidisciplinary actions can be given to slow down its progression or may help to control symptoms 2 4 1 Therapeutics Most of the medicines are used to treat the patient with PD by increasing Dopamine level in the brain in order to reduce the progression and or controlling symptoms of the disease i such famous medicine is Levodopa or L DOPA which is converted into Dopamine to treat the motor symptoms somehow L DOPA is extremely useful in Bradykinesia
12. 10K 470M 9 Degree of Freedom daughter board 0 5 50Hz output e Provides gyro magneto sensing Comprehensive d 3m ee e Sensitivity 2 mV sec rehabilitation object tracking sports 9DoF Kinetic sensing ax solutions Magneto range 500 25g pinhole reset 4 training 4 Gyro field range 0 7 4 5Gauss 390 1620 counts mGa GPS Accurate time Embedded GPS receiver daughter board Environment maritime sensing tracking iPS PE e location Pressure Range 300 1 100hPa Temp Range 0 65 2 d objects amp people measurement Timing accuracy lt 100ns Position accuracy lt 3m Measures accurate location amp time Table 1 Description of SHIMMER Devices 3 1 1 SHIMMER Base board SHIMMER can be used as baseboard and can measure 3 axis acceleration independently While all other circuit boards can be connected as daughter boards MSP430 Microcontroller This main board has MSP430 Microcontroller which works as CPU and main parameters such as 8MHz 10 RAM 48KB Flash memory for programming and debugging an 8 channels integrated ADC DAC analog signal to capture motion 16 bit registers 3 light emitting diodes integrated Bluetooth radio module and a slot to hold 2GB MicroSD card The main board is light weight 22g compact 53 32 25 and highly configurable as shown in the Table3 1 It works on low level layers Physical amp MAC and low rate PAN standards on low power Bluetooth Roving
13. experiment is conducted 3 weeks after the first experiment In 2 experiment two readings were taken with a half hour pause We may assume that the patient may perform the same as before Subject underwent 6 7 types of gaits in 4 meter Go amp Back 4GB pattern to closely monitor and analyze the patient s gait The pathway was clearly shown to the patient before taking measurements The patient was instructed to walk at her normal slow fast holding objects and circle gait and go back This 52 protocol is bit similar to 7 One person was checking real time gait through computer and saving files the other person was recording the gait for validation later on While another lady is there to assist the patient in the walkway We changed the walkway from L shaped to 4 meter Go amp Back 4GB fashion to give more challenges to induce patient to pose unbalanced gait perticularly during turning Details of measurements of all gaits along with comments are shown in both gait tables The 6 types of gait for the first experiment are as under b d g a Figure 15 Data Acquisition through WSN SlowGait NormalGait FastGait GlassBoth Sit_Stand_GlassBoth Pick_GlassBoth The 7 types of gaits for the second experiment putting two accelerometers on either sides of waist are as under a b d e f g NormalGait FastGait BothGlass FastBothGlass Figure 16 Two Accelerometers on Waist SitChair2C
14. no 2 pp R1 R20 Apr 2004 T Tamura T Yoshimura M Sekine M Uchida and O Tanaka A Wearable Airbag to Prevent Fall Injuries IEEE Transactions on Information Technology in Biomedicine vol 13 no 6 pp 910 914 Nov 2009 H Onodera T Yamaguchi H Yamanouchi K Nagamori M Yano Y Hirata and K Hokkirigawa Analysis of the slip related falls and fall prevention with an intelligent shoe system in 2010 3rd IEEE RAS and EMBS International Conference on Biomedical Robotics and Biomechatronics BioRob 2010 pp 616 620 K H Low J W Tani T Chandra and P Wang Initial home based foot mat design x00026 analysis of bio gait characteristics to prevent fall in elderly people in 2009 IEEE International Conference on Robotics and Biomimetics ROBIO 2009 pp 759 764 D T H Lai R K Begg and M Palaniswami Computational Intelligence in Gait Research A Perspective on Current Applications and Future Challenges IEEE Transactions on Information Technology in Biomedicine vol 13 no 5 pp 687 702 Sep 2009 C C Yang Y L Hsu K S Shih J M Lu and L Chan Real time gait cycle parameters recognition using a wearable motion detector in 2011 International Conference on System Science and Engineering ICSSE 2011 pp 498 502 J Cancela M Pastorino M T Arredondo M Pansera L Pastor Sanz F Villagra M A Pastor and A P Gonzalez Gait assessment in Parkinson s disease pati
15. DEMO APPLICATIONS AND SAMPLES eese nennen nennen innen tnnt innen enn 38 4 4 MOBILE SYMBIAN enne eene eee nnne ettet nett testen teen tentent E a 38 4 5 EYESWEB5 EVO eee ted 39 45 eene dw tvutizi EE UNE NIRE EDS 39 45 2 Eyes Web Blocks Se Patches NI ee 39 4 6 M E 41 4 7 ENDINGS ET 42 5 GAIT ANALYSIS TOOLS amp 8 22 404 4 1 2001 80 00 20 40 0000000 0 4 43 5 1 EXISTING SCALES TESTS FOR GAIT 55 55 0 44 5 1 1 Modified Fall Efficacy 44 5 1 2 Berg Balance Scale BBS esses seen eene enne 44 5 1 3 Unified Parkinson s Disease Rating Scale UPDRS esee 44 5 14 Instrumental Activity of Daily Living Scale 2 45 5 1 5 Activities of Daily Living ADL esee nennen trennen 45 5 1 6 Timed Up amp Go Test TUG 45 5 1 7 Mini Mental State Examination MMSE csse ener eene enne 45 5 1 8 Hospital Anxiety amp Depression Scale HADS 2 2 2 45 5 1 9 Prevention Home Assessment Chart eese eese eene eene ener 46 5 1 10 Personal Risk Factors Checklist eese esses esee ener nnne 46 5 1 11 4 Me
16. E E E R A EA 55 6 2 1 Step County dev asd terae eno e Giese aah eet ae Anes 55 pP dE CM 57 6 3 RESULTS amp DISCUSSION iss eni eto tete rH ERROR OUO ERU 60 6 4 CONCLUSIONS ot petto nca OR 64 6 5 FUTURE WORK ota a ERU DIO Oe RD RR MR 64 6 6 SUMMARY sr M TED eed 65 6 7 FINDINGS n a ELTE E Med I 66 6 8 BIBLIOGRAPHY eite Gic era p e ee ee i pde eese 67 List of Figures Figure 1 Research 1 12 Figure 2 Data Acquisition to Track 13 Figure 3 Parkinson s disease World Map with Courtesy 1 17 Figure 4 Primary Symptoms in Parkinson s disease 18 Bigure 5Gart Amal ysis eod tees eto dre de den Eee Re RART 28 Ligure 6 Shimmer X YZ seni esas caps iun these HH ete E REED pues 32 Figure 7 Internal Circuitry Parts of SHIMMER Accelerometer and Package Box 34 Figure 8 Shimmer Baseboard Interconnections and Integrated Devices 36 Figure 9 Shimmer 37 Figure 10 AccelGyro shimmer2r 50Hz 1 56 37 Figure 11 EyesWeb GDE Main
17. EyesWeb programs are developed to create Patches and can be saved as an extension test eywx Each patch consists of a set of Blocks and Functionality of EyesWeb is implemented through different types of blocks These can be seen in Catalog view in Figure 12 13 Different blocks have different functions In EyesWeb GUI Objects contains basic functionality blocks while other main categories of blocks are Audio BioMOBIUS 39 DataStructures FileSystem FlowAndControlStructure Geometeric ImageAndVideo Math Operations Peripherals Strings TimeAndDate TimeSeries Talking about developing EyesWeb Patch following blocks have been used I Bang Generator Bang Generator is a GUI object to trigger an action It works like button Connect Start Stop Disconnect Initialize Start Stop Connect Disconnect Bangs are used to control Shimmer while Initialize Bang is used to generate constant value to Constant Generator such as one Parameter Patch_Start True BioMOBIUS SHIMMER Main block is Shimmer block having 0 input and 16 outputs and 16 important parameters Discussing these parameters Serial Port 11 depending upon Accel port nSample sampling rate such as 100 samples sec Dev Config Accel Accel Range 1 5g EyesWeb Patch Real Time Connect SHIMMER Figure 12 Real time EyesWeb Patch Display Constant Generator Show SyncIn True Input Value 0 for X
18. Integrated Devices END OF CHAPTER 3 2 Findings The detailed description about the SHIMMER platform devices has been presented into one Table 1 36 4 SOFTWARE Some software is required to program and connect to the SHIMMER One such is BSL430 to burn program and second is ShimmerConnectV2 0 to connect with SHIMMER devices There is also need to discuss some features related to EyesWeb and Matlab to make this device compatible to enable certain functionality like ECG EMG GSR etc 4 1 BSL430 BSL stands for bootstrap loader It is the default set of open source program provided by the SHIMMER It contains various samples to capture motion data with different combinations of sampling values and sensing parameters as shown in Figure 9 Mainly two of them are bootstraps and legacy bootstraps Bootstraps contain 4 samples and each sample contains following set of programs Configure Shimmer 1 AccessPointShimmer ihex 9 Sampling Rate 2 BlinkShimmer ihex Toggle LED Sensors to sample 3 BoilerPlateShimmer ihex can be used as default Gyroscope GSR 4 HostTimeLogging shimmer ihex Magnetometer AnExADCO ECG AnEx ADC7 5 JustFATLogging_shimmer ihex Strain Gauge HeartRate 6 SixAxisTransmitter_shimmer ihex Enable 5V regulator _ Enable Voltage monitoring 7 Sleep_shimmer ihex Accel Range GSR Range 1 5g 10kQ 56kQ v Cancel eem Figure 9 Shimmer Configurations While le
19. Network RN 42 at 3Mbps and up to 20 meters Timestamps can be given to the signals due to the embedded clock in the Microcontroller This clock is highly configurable to use an extremely low battery MMA7361 Accelerometer The main board also contains Accelerometer MMA7361 with important features such as 3 5 1mm current consumption 400uA sleep mode 3uA scalable sensitivity y 1 5g 6g 33 Important applications for Accelerometer MMA7361 are 3D Gaming HD MP3 Pedometer amp Robotics as motion sensing Freefall detection through PC amp Laptop Sensor 21014 Fachistes soft power switching Antenna D gt e Figure 7 Internal Circuitry Parts of SHIMMER Accelerometer and Package Box 3 1 2 Daughter board Daughter boards are relatively smaller circuit boards for 3 party sensing NAS which can be connected with base board to function It includes gyro NS magneto ECG EMG GSR and other circuit boards Details can be read in the table 3 1 3 1 3 MicroSD Card 2GB Each base board comes with a built in MicroSD flash socket with 2GB card e to record data for offline processing Particularly useful to record data of experimentation for longer than a whole day 5 3 1 4 Leads ECG Cardio sensing to record electrical impulses of heart muscles are done AC through ECG leads 34 3 1 5 Live USB You can boot directly from the Live USB with TinyOS already
20. PC through a data acquisition system for interpretation and offline processing Depending on different requirements real time implementation through algorithm can be performed on real time motion analysis to trigger different healthcare facilities such as emergency alarms automated SOS calls and messages nursing or health givers an alertness etc The schema of our acquisition system is presented through the following figure N Data Real time Data processing Moti shimmer Acquisition and feature extraction otion Alarm Detection system m mes gt _ ireless x 95 amp Threshold gt 225 acclerometer offline real time 55 sampe Warning alarm csv file calculating Std x y z Threshold 330 Emergency alarm F Figure 2 Data Acquisition to Track Motion 1 15 Advantage and Future Prospect The major advantage of this algorithm is its reusability Algorithm is not hard coded because a user can set his own sampling rate and threshold value and check the results Specially minimum and maximum threshold values can be set through a GUI slider as shown in the Figure 18 This algorithm is further modifiable to trigger an airbag a security push button SOS calls messages a siren activation system automatic email forwarding health care alert system and many more The same algorithm with minor 13 modifications can be used for seizure detection in other disorders mainly in epileptic s
21. a 2 GB MicroSD card can be used to capture data in the SHIMMER accelerometer itself The use of accelerometer is more suitable due to the fact that we are capturing data from postural instability One two or combinations of more accelerometers can be put on different body parts While SHIMMER Gyroscope may be more suitable for jerky motion for diseases such as epilepsy Mostly accelerometers and gyroscopes are used for gait analysis 4 1 10 Short Description of Study Defining our research work this study is carried out on a patient with Parkinson s disease PD to study various gait parameters test wireless accelerometers on different body parts and implementing an algorithm to trigger a security alarm system by setting a threshold value Criteria for setting threshold value are calculating standard deviation and employed by different researchers like 5 11 1 11 Motivation Actually there is not such a smooth mechanism to monitor human gait We need to develop such an efficient and adaptive method so that a user can set own sampling rate and thresholds The main motivation to perform this experimental research work is to detect any gait deviation for the patient with PD Security alarms can be activated whenever a patient poses shakier gait Two types of alarms or sirens can be activated in the algorithm first to activate Warning Alarms when the value from motion data exceeds minimum threshold value and second to activate Emergenc
22. axis 1 for Y axis and 2 for Z axis IV Input Selector 3 parameters used are number of input 3 active input 1 checkbox is clicked Active on Select False 40 V Matrix Display in order to see XYZ acceleration VI FastDataBufferMatrix From Input Selector to save csv files WriteMatrixToFile int In order to save motion data into csv file VII Text Selector Important parameter is Docked True VIII String Display Important parameter is Docked False Another EyesWeb patch is also developed to save the CSV file for offline processing EyesWeb Patch Saveto File Figure 13 Real time EyesWeb Patch Save into file 4 0 Matlab Real time motion detection algorithm can be implemented in EyesWeb but we selected Matlab as a designing tool because it s signaling data handling capabilities are more efficient Shimmer Research has already provided Shimmer Instrument Driver Rev0 4 for Matlab Once again before proceeding with Matlab demo applications for Shimmer need to be tested such as plotandwriteexample m file There are two other Matlab files 1 ShimmerHandleClass m 2 twoshimmerexample m to handle and use two SHIMMERS with different settings at a time It is also recommended to read Shimmer Matlab Instrument Driver User Manual MATLAB v9 is used to calculate mean median standard deviation and different types of plots to analyze the motion data shown in the Tables 6 2 amp 6 3 Stand
23. below Sometimes slightly higher acceleration in SlowGait is due to jumping behavior 57 3s de a ee Me Pale Ps P af Pn A in zi 1 acl 200 aon 1500 Plot 1 Consciously slow gait during turning becomes more stable Our studies also demonstrated that the induced small step gait has significantly lower acceleration resulting a more stable gait So whenever a patient feels unstable he she can avoid postural instability during a shaky gait through small stepping 1 1 150 200 250 Plot 2 SitChair2Chair Gait Ploting in MATLAB Patient takes turn before sitting on the chair each time in the gait SitChair2Chair holding glass in both hands so threshold value should be set to discard unnecessary noise in the signal as shown in the Plot 2 Also note that this gait is formulated to check the sitting patterns of patient holding glasses in both hands without arm support In CircleGait Accel Y becomes more sensitive due to change of orientation of shimmer accelerometer such as between samples 676 700 Accel XYZ 213 83 237 77 161 66 In FastGait higher acceleration to be noted is 289 75 captured through EyesWeb as shown in the plot 3 below Data is captured for first 5 seconds to analyze the starting behavior It is also observed that when the patient starts stepping with energy her gait patterns are bit
24. e n 18 2 3 PARKINSON S DISEASE 19 2 3 1 Primary Symptoms Motor 1 19 2 3 2 Secondary Symptoms Non Motor Impairments esee 21 2 4 TREATMENTS deinen e n ER 22 24A Therapeutics ONE QE HA en Ee REV ERE eM 22 2 44 2 Surgical Treatment est tae e ia o a eo T E He ERROR Ye eR Vua 22 24 3 RIED OR QE AV TE REX REMANERE UT 23 224247 PIETER NUTT OM aive eei IE 23 2 4 5 Lifestyle Modification i adatta te a e a ee ER REV 24 2 5 FALLING IN OLDER PERSONS 2 60000 n n enhn enhn nnn 24 2 5 1 Factors Responsible for Falling eese eene eene enne 25 2 82 Fall Detectionz iciatis taj yere EE RE PUE URP Erde 25 29 3 TZ RERO 26 2 6 GATT PARAMETERS eh ettet tee eet thee 27 2 627 Step related Parameters eaqui tet edite e HER etg 27 2 6 Limbs related Parameters iccccccccccssesccceccccccsssssccecccccsussscececcccssusssesccceceusuaeseccecessusaaescesseeees 28 267 Body related Parameters aset dtt 28 26 4 Constant Parameters de eee tee cbe ee eee eus e to Pede eee eee de 28 2 6 5 lt
25. has overall good step count per unit time In short if the patient has to do some extra work during the walk it becomes problematic for her It becomes a source of reduced walking patterns and lack of concentration which later on may be a major reason for fall So if both hands are set free then probably patient will walk at normal speed with good meditation Sometimes fast but conscious gait with energetic stepping produces walking patterns better with low deviation in acceleration as shown in the following calculation in FastGait Exp B Std xyz 124 55 122 88 97 43 56 Ignoring Std z 97 43 because both values for x and y axis yield higher deviation The subject leans right during turning back and gives higher deviation in PickGlassBoth gait as shown below fmax 417 38 6 2 22 Gait Tables Different values for standard deviation are calculated by putting one accelerometer on the right waist position in the experiment 01 and two accelerometers in both waist positions for experiment 02 The positions where accelerometers are put on waist is another gait parameter Experiment 01 Standard Deviation of XYZ Acceleration detail results 4 meter Go amp Back 4GB X ACCELERATION SAMPLES teken 9374 12139 2298 total samples 1744 15 17 300 samples at rest 19337 32858 426 450 samples 126 06 22484 501 525 samples 148 38 26171 626 650 samples Stghtly higher values show bit jumping gait 15221 1668 total sample
26. installed on it Other features and environments such as CYGWIN TinyOS Source code developing tools such as 28 BioMOBIUS can also be installed without any network access Live USB can also be used virtual machine using VMware 3 1 6 USB Dock A USB cable is needed to connect the SHIMMER with PC through a docking station to program the shimmer It can also charge the device but take more time Green LED color shows that device is connected and fully charged while yellow shows not fully charged or during data transfer This light becomes light pink when device is being programmed 3 1 7 6 Gang Multi Charger This charger can hold up to six SHIMMERS to charge at a time Green light shows fully charged while yellow light shows partially charged SHIMMERS may not be configurable with low charging even connected with docking station to the PC 3 1 8 Straps and bands Wrist straps and chest or waist bands are used to put shimmer devices on these body parts 3 1 9 Bluetooth SHIMMER is used to communicate with PC or mobile devices through two radio communications such as Bluetooth and Roving Networks 802 15 4 Roving networks has bit a slow data rate as compared to Bluetooth but considered better with other performance parameters like it takes low power in sleep mode 264A 5 SDHOST 5 0 UARTO Dock Analog GPIO UART1 Figure 8 Shimmer Baseboard Interconnections and
27. may become more stable as in conscious gait Plot 1 From our visual examination and post processing analysis results show that after DBS surgical procedure the patient still experiences postural instability even she suddenly falls showing no sign specially when medication is not taken in time So it is evident to show that such patients may have reduced cognition even after surgery 6 8 Bibliography 1 2 3 4 5 6 7 8 9 10 11 12 13 El Gohary McNames J Chung Aboy Salarian A Horak F Continuous At Home Monitoring of Tremor in Patients with Parkinson s Disease BIOSIGNAL 2010 S Patel C Mancinelli A Dalton B Patritti T Pang S Schachter and P Bonato Detecting epileptic seizures using wearable sensors in Bioengineering Conference 2009 IEEE 35th Annual Northeast 2009 pp 1 2 U B Flansbjer and J Lexell Reliability of Gait Performance Tests in Individuals With Late Effects of Polio PM amp R vol 2 no 2 pp 125 131 Feb 2010 T Shany S J Redmond M R Narayanan and N H Lovell Sensors Based Wearable Systems for Monitoring of Human Movement and Falls IEEE Sensors Journal vol 12 no 3 pp 658 670 Mar 2012 Jonas STANDAERT Wouter SPEYBROUCK Implementing real time step detection algorithm in EyesWeb environment Arkiv EX Blekinge Tekniska H gskola Online Available http www bth se fou cuppsat
28. stable 58 data 1 200 1 1 1 1 1 1 1 1 o 50 100 150 200 250 300 350 400 450 500 Plot 3 FastGait in EyesWeb showing higher acceleration In CircleGait higher acceleration noted to be 261 24 This type of gait protocol is designed to give the patient some challenging environment as shown in the plot 4 below data 1 spline linear 1200 1000 Higher Std y 261 24 600 b w 126 150 samples 200 1 1 1 1 L L 1 L 0 50 100 150 200 250 350 400 450 500 Plot 4 CircleGait in EyesWeb showing higher acceleration 59 6 3 Results amp Discussion Experiment 02 Only Max Standard Deviation of XYZ Acceleration detail results in4 meter Go amp Back 4GB and circling gaits Z ACCELERATION SAMPLES tsken 151 175 1476 1500 181 68 22355 13442 7712 105 28 79 68 ples 2601 2625 426 450 samples 3 BathGlassGoit 34 24285 21793 12481 10327 5317 8055 876 900 326 350 samples 4 FastBothGlass 41 27292 23626 16227 154 00 5682 10725 301 325 826 850 samples Comments 5 21656 23234 86 90 180 86 63 62 7429 51 18427 228 88 178 02 93 89 10742 11529 m fortotal samples 98 172 119 Shows patientturns before sitting so gives y axis more sensitive more accelly 111 89 93 73 72 16 Total Samples 21160 193 96 116 82 4751 4775 samples 213 84 16243 23777 18615 161 67 12427 TS Total Std xyz gives clase values this Std x 211 as patient takes a l
29. step frequency These authors also describe that small rapid stepping Festination and backward steps can be detectable by 14 signature identifier Vertical acceleration can be detected through 3 axis accelerometer which should be sensitive to low frequency low amplitude motion and low energy consumable They also stress to use a gyroscope in combination with accelerometer for accurate horizontal movement of the individual s head Above authors suggest to use predefined gait features for an indication that a person is falling or prone to fall If an individual is about to fall then system is configured to notify a third party and warn individual with postural instability as well 6 Another related work is done by Toshiyo et al 23 in which a wearable automatic airbag opens when a fall occurs based on acceleration and angular velocity They incorporate a fall sensing algorithm in the form of a wearable smart jacket to protect head and hip when value exceeds to the threshold limit So this algorithm is proved helpful to avoid fall injuries thus saves lives at construction sites and other locations The authors have used accelerometers and gyroscope in their studies Our proposed method works on some of features mentioned above Some of these characteristics are as given Our efficient and sensitive algorithm can be used to detect a small deviation in the gait in all three axis setting threshold values using wireless triaxial acceleromet
30. 3767 M M Skelly and H J Chizeck Real time gait event detection for paraplegic FES walking IEEE Transactions on Neural Systems and Rehabilitation Engineering vol 9 no 1 pp 59 68 Mar 2001 L Wang T Tan W Hu and H Ning Automatic gait recognition based on statistical shape analysis IEEE Transactions on Image Processing vol 12 no 9 1120 1131 Sep 2003 M Goffredo Bouchrika J N Carter and M S Nixon Self Calibrating View Invariant Gait Biometrics IEEE Transactions on Systems Man and Cybernetics Part B Cybernetics vol 40 no 4 pp 997 1008 Aug 2010 R Andrew Swartz Deokwoo Jung Jerome P Lynch Yang Wang Dan Shi Michael P Flynn Design of a Wireless Sensor for Scalable Distributed In Network Computation in a Structural Health Monitoring System Proceedings of the 5th International Workshop on Structural Health Monitoring Stanford CA USA September 12 14 2005 A Basharat N Catbas and M Shah A framework for intelligent sensor network with video camera for structural health monitoring of bridges in Third IEEE International Conference on Pervasive Computing and Communications Workshops 2005 PerCom 2005 Workshops 2005 pp 385 389 J M Havinga Kauw A Tjoe M Marin Perianu and J P Thalen SensorShoe Mobile Gait Analysis for Parkinson s Disease Patients Jun 2007 Online Available http doc utwente nl 64122 Accessed 31 Jul 2012
31. 40 www shimmer research com access june 2012 69 4 Meter Go amp Back Test 4GB There are 7 types of gaits which are monitored by the sensors and can be validated through video recording at the same time Gait Name Gait Parameters Body related parameter Total No of Steps Total Time Cadence Steps Sec Distance Normal Gait e g Either sides of the e g 18 19 e g 11 Seconds e g e g 8 Meters Slow_Gait waist Fast_Gait Fast_Both_Hands_Glass Sit_Chair2Chair_Glass Circle_Gait Circle_Glass_Gait Fast_Pick_Glass_Gait BMI Height Gender Subjects Last Therapy ate Surgery date Medicated Disease condition Disease Body Physique Non addicted no impairment in any neurological disorder backbone slip vision hearing cognition memory loss hallucination extremely tired etc can be put into Exclusion Criteria i Body Related Parameters can be anywhere such as upper and lower limbs chest ankle belly bellybutton trunk waist etc 1 If no of subjects is more than one separate test for each subject 11 Any or all of Gait Names can be repeated depending upon specific requirements iv Other Special Parameters can be included changed of walkway in Go amp Back L Shaped U Shaped depends on specific requirements v Circle_Gait in a 1 meter diameter with no hurdle i Du
32. Master s Thesis Computer Science September 2012 Real Time Gait Analysis Algorithm for Patient Activity Detection to Understand and Respond to the Movements Inam ul Haq M Adnan Jalil School of Computing Blekinge Institute of Technology SE 371 79 Karlskrona Sweden This thesis is submitted to the School of Computing at Blekinge Institute of Technology in partial fulfillment of the requirements for the degree of Master of Science in Computer Science The thesis is equivalent to 20 weeks of full time studies Contact Information Authors Inam ul Haq Address Blekinge Institute of Technology E mail inam bth gmail com M Adnan Jalil Address Blekinge Institute of Technology E mail adnanjalil1985 yahoo com University Advisor s Supervisor Jenny Lundberg PhD School of Computing Blekinge Institute of Technology Sweden School of Computing Internet www bth se com Blekinge Institute of Technology Phone 46 455 38 50 00 SE 371 79 Karlskrona Fax 46 455 38 50 57 Sweden ABSTRACT Context Most of the patients suffering from any neurological disorder pose ambulatory disturbance at any stage of disease which may result in falling without showing any warning sign and every patient is different from another So there is a need to develop a mechanism to detect shaky motion Objectives The major objectives are i To check different gait parameters in walking disorders using Shimmer platform ii We
33. P Corsini J Light and A Vecchio Monitoring of Human Movements for Fall Detection and Activities Recognition in Elderly Care Using Wireless Sensor Network a Survey in Wireless Sensor Networks Application Centric Design Y K Tan Ed InTech 2010 G Diraco A Leone and P Siciliano An active vision system for fall detection and posture recognition in elderly healthcare in Design Automation Test in Europe Conference Exhibition DATE 2010 2010 pp 1536 1541 V Vishwakarma C Mandal and S Sural Automatic detection of human fall in video in Proceedings of the 2nd international conference on Pattern recognition and machine intelligence Berlin Heidelberg 2007 pp 616 623 M Alwan P J Rajendran S Kell D Mack S Dalal M Wolfe and R Felder A Smart and Passive Floor Vibration Based Fall Detector for Elderly in Information and Communication Technologies 2006 06 2nd 2006 vol 1 pp 1003 1007 C J Robinson M C Purucker and L W Faulkner Design control and characterization of a sliding linear investigative platform for analyzing lower limb stability SLIP FALLS IEEE Transactions on Rehabilitation Engineering vol 6 no 3 pp 334 350 Sep 1998 M J Mathie A C F Coster N H Lovell and B G Celler Accelerometry providing an integrated practical method for long term ambulatory monitoring of human movement Physiological Measurement vol 25
34. a In particular accelerometers and gyroscopes are used for gait analysis 4 Figure 5 Gait Analysis 2 6 4 Constant Parameters Some parameters are not directly related to gait and cannot be changed Rather these are to be kept under observation These parameters include number of total participants N BMI age disease age sex the height body physique gait regularity medicated undergone surgery therapy disease condition normal moderate or severe and the number of subjects Other closely related disorders are also noted such as backbone disk slip mentally retarded and similar The detail description is given in the 4GB Test at the end of the report Our subject is 70 year old a healthy lady recently undergone medical surgery reduced gait small steps no or minimum tremor 2 6 5 Assumed Parameters It is also assumed that no other impairments or disorders are present These parameters are impairment during hearing vision cognition memory loss hallucinations non alcoholic and non addicted and can be kept under exclusion criteria 28 2 6 6 Special Parameters There could be put some activity or induced patient to perform task such as to walk over a straight line U shaped L shaped Sit Stand Stand Go in the shape of different tests such as TUG time up amp go test 4GB 4 meter Go amp Back slow or fast gait and different scales such as ADL holding something during walking etc can be used Healthy subjec
35. ance 26 Implementation of proper and timely intervention to decrease the fall risk is highly important and desirable Following are some 24 of important fall avoidance prevention techniques Fall Avoidance Techniques Technology advancement in the field of electrical instrument has resulted in different reliable techniques to prevent fall Some of the fall prevention techniques are given below D Intelligent Shoe System is designed to prevent slip during the walk using an enlarged sole area up to 7 cm comprising a parallel link with a servo motor The degree of sole area changes with a real time analysis taken from the data acquisition system 24 II Foot Mat installed on the floor can be easily 25 applied at home In foot mat there is a triaxial accelerometer that measures low high frequency and vibration in three dimensional and a sensor that acts as a sensing resistor in the electrical circuit Another advantage of using foot mat is that it has data acquisition system 2 6 Gait Parameters As this thesis project is directly related to Fall Detection so FoG fall risk assessment postural instability and trembling of legs can also be included based on interested features to extract from Generally different parameters or features related to gait have been described such as body velocity time ratio stride time and length swing stance phases maximum acceleration at toe off etc Different researchers try to include or exclud
36. ard deviation with different gait parameters such as sampling rates accel sensitivity accel placement gait types time duration amp distance seems suitable for making it as the basis to find how far a 41 value lies from mean value Different researchers have used standard deviation as basic criteria to compare it with threshold values It will be discussed in the later chapter in more detail Detail description about using Matlab will be discussed in Chapter 6 THE END OF CHAPTER 4 7 Findings MATLAB tool has proved to be very efficient in dealing with calculations related to signals from the SHIMMER So it is a signal processing tool which is easily be employed in such applications where there is a requirement to capture and track the real time motion as data acquisition 42 5 GAIT ANALYSIS TOOLS amp METHODS Gait analysis is an important component of neurological examination after physical examination During analysis patient balance is checked as a part of Parkinson s disease standard measurement Despite SHIMMER platform there exist multiple instruments and tools of various complexities to 32 analyze the gait such as T amp T Medilogic Medizintechnik GmbH measurement device marker set zebris Medical GmbH an electronic controlled carpet with integrated pressure sensors GAITRite CIR Systems Inc a camera guided 3 D kinematic system VICON Oxford Matrix or a camera operated video system Peak Performa
37. aring SHIMMER wireless sensors on hip waist and chest to check which one is the most suitable 111 To draw effective conclusion results based on calibrated data in real time and offline processing in EyesWeb Matlab To develop an effective mechanism algorithm for security warning and activating alarm systems Methods Our thesis project is related to analyze real time gait of the patient suffering from Parkinson s disease for actively responding to the shaky movements Based on real world data we have developed a mechanism to monitor a real time gait analysis algorithm to detect any gait deviation This algorithm is efficient sensitive to detect miner deviation and not hard coded i e user can set Sampling Rate amp Threshold values to analyze motion Researchers can directly use this algorithm in their study without need to implement themselves It works on pre calculated threshold values while initial sampling rate is set to 100MHz Results Accelerometers putting on the chest shows high unnecessary acceleration during fall suggest putting on waist position Also if a patient initiates steps with energy his her gait may become more stable as shown in the conscious gait Results show that after DBS surgical procedure the patient still experiences postural instability with fall So it is evident to show that such patients may have reduced cognition even after surgery Another finding is that such patients may lean left or right during turnin
38. ctively achieving the goal of designing an effective technique to actively respond to the shakier motion setting threshold values Our proposed algorithm is easy to implement reusable and can affectively generate healthcare alarms Additionally the system might be used by other researchers in their studies for real time tracking of the motion of the patient with PD The proposed method is sensitive to detect fall therefore objectively can be used for fall risk assessment as well We have also validated that putting wireless accelerometers on waist position gives efficient result and this is performed through literature review pilot experimentation The major advantage of this algorithm is its reusability Algorithm is not hard coded because a user can set his own sampling rate and threshold value in order to check the results More specifically minimum and maximum threshold values can be set through a GUI slider as shown in the figure 18 6 5 Future Work This algorithm is further adjustable to trigger an airbag a security push button SOS calls messages siren activation system automatic email forwarding health care alert and many more The same algorithm with minor modifications can be used for seizure detection in other disorders mainly epileptic seizers to alert health providers for emergency In this way in combination with Gyroscopes related disorders and seizures can also be detected A market device can be equipped with button optio
39. d under this project Other services to track information are ECG heart rate and different sensors This project is cooperated with MIT RGI the BIOM Institutes of TUT and other companies 5 3 2 Sensors for Medicine and Science Sensors Medicine and Science A concept of an integrated Glucose Monitoring System is given by the Sensors for Medicine and Science Ins Sensors for Medicine and Science is developing different glucose monitoring products in order to facilitate diabetic patients 47 to regularly monitor their blood sugar with ease This system provides highly accurate results with longer life along with analysis for O2 CO pH and Lactate 5 3 3 ZigBee ZigBee 62 pod It provides Certified Products for maintaining Health and Fitness of the patients In this long list some products in the Health amp Fitness category include weight scales BP monitors in home displays wireless watches traffic management systems different kinds of sensors and many more devices in order to help people live non dependent The importance of these ZigBee applications is that they require low data rate 250 Kbps longer battery and secure networks for periodic or continuous tracking of data 5 34 PUHVI 5 3 5 WIN Human Recorder Co Ltd 5 4 A short range PAN well being project for conducting research possibilities of applying wearable technology in wellness and healthcare providing long term health monitoring I
40. daert M Akay J Dy M Welsh and P Bonato Monitoring Motor Fluctuations in Patients With Parkinson Hx0027 s Disease Using Wearable Sensors IEEE Transactions on Information Technology in Biomedicine vol 13 no 6 pp 864 873 Nov 2009 N Giladi H Shabtai E Rozenberg and E Shabtai Gait festination in Parkinson s disease Parkinsonism amp Related Disorders vol 7 no 2 pp 135 138 Apr 2001 A C Lo V C Chang M A Gianfrancesco J H Friedman T S Patterson and D F Benedicto Reduction of freezing of gait in Parkinson s disease by repetitive robot 67 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 assisted treadmill training a pilot study Journal of NeuroEngineering and Rehabilitation vol 7 no 1 p 51 Oct 2010 Marc Bachlint Meir Plotnikl Daniel Roggent Noit Inbarl Nir Giladi Jeffrey Hausdorffl Gerhard Trostert Parkinson s disease patients perspective on context aware wearable technology for auditive assistance Greene B R O Donovan A Romero Ortuno R Cogan L Ni Scanaill C Kenny R A Falls risk assessment through quantitative analysis of TUG A M O Halloran N P nard A Galli C W Fan H Robertson and R A Kenny Falls and falls efficacy the role of sustained attention in older adults BMC Geriatrics vol 11 no 1 p 85 Dec 2011 S Abbate M Awenuti
41. dation Different types of gait parameters such as constant and assumed parameters are clearly mentioned along with special instructions to follow before taking measurements Some important gaits are as follows e Sit Stand Chair no arm support e Circle Gait with amp without glasses in both hands e Simple Slow and Fast Gait e Pick Both Glasses Gait Detail description is provided at the section 6 1 6 2 amp Appendix A Existing Gait Analysis Methods Some of the existing gait analysis methods and techniques are presented here A real time gait event detector is proposed to automatically control FES functional electrical stimulation during paraplegic locomotion The algorithm is based on fuzzy logic to estimate patient s current state of gait 34 The term Gait Recognition has recently gained significant attention which is strongly motivated by the need for automated person identification systems at a distance in visual surveillance and monitoring applications The authors propose an efficient automatic gait recognition algorithm using statistical shape analysis In this analysis for each image sequence an improved background subtraction procedure is used to extract a moving silhouettes of a walking figure from the background 35 46 A new method for viewpoint independent gait biometrics is proposed In this method the system relies on a single camera without camera calibration and works with a wide range of camera vi
42. e different gait parameters depending upon their own specific requirements However the 26 article describes different combinations of gait parameters in its own style under the category of Gait Data Here we try to present different gait parameters dividing into various categories Some of these are listed below 2 6 1 Step related Parameters All parameters related to feet such as step length and time frequency maximum acceleration at toe off or heel strike an ankle swing rate and steps unit time A Sliding window method is 27 generally used to periodically measure gait parameters in the real time environment An algorithm based on Double Pendulum is employed to measure stride 28 length during walking 27 2 6 2 Limbs related Parameters Upper and lower limbs thigh shank finger hand wrist mostly to detect and analyze tremors 2 6 3 Body related Parameters Sensing devices can be worn on different body areas to capture motion data Some of these body areas are different positions on hip belly bellybutton chest trunk waist and sternum Higher acceleration has shown when wireless accelerometers are put on Chest as compared to using accelerometers on Hip or Waist during Freezing of Gait and falling Our subject was wearing a small wireless tri axial accelerometer in her either sides of the waist as shown in the Figure 5 Reason for choosing waist position is just in order to avoid tracking unnecessary motion dat
43. e 71 Appendix C FlowChart Offline Gait Analysis teen ees 72 Appendix D Parkinson s disease Symptoms esses 73 Appendix E FlowChart Real Time Gait 1 8 2212 74 Appendix Offline Code Gait 1 8 22 75 Appendix G Offline Code Gait 1 021 76 Appendix Scientific Discussion Validation on Results 78 Permissions Thesis authors grant permission to BTH for unconditional usage of this thesis work partially or fully in any way or form suitable to any person researcher organization or institution This report is a part of thesis project conducted under BTH Errors and omissions may occur which will be removed later on Inam ul Haq amp M Adnan Jalil Ronneby September 2012 Acknowledgement This thesis project could not be completed without the efforts and pain taken by Jenny Lundberg being supervisor and Lars Lundberg Here prayers of our friends Wouter Tavakoli many researchers in the related field for getting inputs and validation on results Definitely we would like to appreciate sincere efforts of parents who contributed a lot and we express good level of appreciation for them In the last I admire the support from my wife who prayed for my work For granting hardware facilities and software access we are muc
44. earch that you are testing against If so you should look to those findings as a marker for validation We have captured raw acceleration from the sensors accelerometers and devices may vary will it be an issue It sounds like a reliability issue If two different people are taking measurements with two different devices on the same exact phenomenon those data should match If the devices vary you MUST make sure that they are giving you reliable data If they do it is not a problem If they do not yes it will be an issue You also want to determine that human error is eliminated Are the instrument calibrated if they can be Are the researchers trained in the same manner I have looked at the attachments but would still like specifics on the type of input that you are looking for It may help to know that I am a qualitative researcher who studies age friendly cities and policies I am not experienced in dealing with models or algorithms but would be more than happy to provide you with suggestions about housing and environments for older adults and people with disabilities You asked are there characteristics of the built environment that are important to consider such as housing elements What does this mean If I have understood it probably you are talking the possibility to build this algorithm in some wearable device If so then I am happy that this can be possible This algorithm can be housed in the dress to track the information
45. ed Pants algorithm 30 e Online FOG Detection Algorithm based on Moore s low latency principle 14 e Berg Balance Score and TUG 15 29 Names of related techniques for Fall Avoidance are described in section 2 5 3 END OF CHAPTER 2 7 Findings It is not necessary that every patient with PD possesses the same symptoms at different severity levels But falling patterns somehow can be same either forward side backward Symptoms of tremor or Bradykinesia may progress faster than rigidity Also that every patient has a unique set of symptoms and this may trigger some researchers to work on FOG for fall detection or fall risk assessment It is also to be noted that patients with PD lose their balance control and are more likely to fall During experimentation and measurement it is also observed that falling might be more likely to happen during turning Also wearing accelerometers on the chest gives unnecessary noise data as compared to waist or hip As citizens at age 70 are likely get ill soon or fall so naturally there is a need to work on this group of people for their well being Detail description for Gait Parameters is introduced the first time in such way There does not exist an exact technical device to detection FoG 30 Treating FoG can be done by involving brain to perform multiple actions such as singing marching ordering like get set go etc This deceiving therapy to the brain helps somehow for patients to unfreeze The
46. eizers to alert health providers in case of emergency 1 16 Overall Overall this report presents the analysis of the experiment to measure the usability of wireless accelerometer data to monitor patient activity suffering from Parkinson disease Our research and experimental work can be quoted towards fall risk assessment 1 17 Organization of Chapters The rest of the report is organized as Chapter 1 gives a basic introductory description of the work Chapter 2 discusses symptoms of PD and Gait Parameters Chapter 3 4 deals with software and hardware involved in this experiment Chapter 5 discusses some existing tools and methods while Chapter 6 discusses our actual experimental environment with a result interpretation 1 18 Major Achievements e Real time algorithm implementation e Detail Description of Gait Parameters e 4GB Test 1 19 Related Work A US patent device for monitoring real time deviations in the patients gait is proposed This device is useful to wear partially inside the auditory canal equipped with accelerometer and gyroscope Subsequent current gait features are extracted continuously to compare with accumulated gait statistics to find the deviation in gait patterns Functioning of this device is based on some important facts One such fact is that neurological disorders possess unique identifiable characteristics For example small shuffling steps can be detected from forward velocity vertical acceleration and
47. ents through a network of wearable accelerometers in unsupervised environments in 68 29 30 31 32 33 34 35 36 37 38 39 2011 Annual International Conference of the IEEE Engineering in Medicine and Biology Society EMBC 2011 pp 2233 2236 J Stamatakis J Cremers D Maquet B Macq and G Garraux Gait feature extraction in Parkinson s disease using low cost accelerometers in 2011 Annual International Conference of the IEEE Engineering in Medicine and Biology Society EMBC 2011 pp 7900 7903 K Niazmand K Tonn Y Zhao U M Fietzek F Schroeteler K Ziegler A O Ceballos Baumann and T C Lueth Freezing of Gait detection in Parkinson s disease using accelerometer based smart clothes in 2011 IEEE Biomedical Circuits and Systems Conference BioCAS 2011 pp 201 204 N Gy rb r A Fabian and G Hom nyi An activity recognition system for mobile phones Mob Netw Appl vol 14 no 1 pp 82 91 Feb 2009 W H Wu A A T Bui M A Batalin L K Au J D Binney and W J Kaiser MEDIC Medical embedded device for individualized care Artificial Intelligence in Medicine vol 42 no 2 pp 137 152 Feb 2008 F Yu A Bilberg and E Stenager Wireless medical sensor measurements of fatigue in patients with multiple sclerosis in 2010 Annual International Conference of the IEEE Engineering in Medicine and Biology Society EMBC 2010 pp 3763
48. ers Step counts vertical acceleration to detect fall notifying a third party amp individual itself are features described in our proposed system as well Another step detection algorithm for PD is implemented This method is based on calculating standard deviation as the basic parameter which will be lower in case of shuffling gait In the post processing environment the threshold value is calculated to be 25 of total standard deviation while it is 20 after trying different values in range 20 33 in real time scenario In order to track the motion initially for 60 seconds a concept of rotating windows is used to make the calculation light This rotating window uses array to hold 100 values to take next action and for each specific second 200 new values are received Thus for each 60 second of the period 12 000 values are received 5 Our method is also based on calculating standard deviations at different time periods during each gait in order to set the threshold values in post processing as shown in tables 6 2 amp 6 3 Here it is observed that a patient may fall during 1 4th of a second thus algorithm will take every 25 values to take next action as shown in the Appendix F Sampling rate 100 Hz means 100 sample values in each second A 3 d array for holding 25 values is declared in the code for this purpose as shown in Appendix G An activity recognition system based on neural networks for mobile phones is presented 31 that rec
49. ertain features of the system Gy orb r Spatial Media Group University of Aizu Aizu Wakamatsu Fukushima ken 965 8580 Japan According to Norbert some possible mobile applications could be GPS tracking in correlation with gait finding out which places routes cause most problems for patients Suggesting alternative routes e g gait analysis shows the route the patient takes to a shop is difficult Similar scenarios can be devised for in house but you d need indoor positioning for that Emergency features if a critical value is reached the phone can automatically call signal an emergency call Collecting of long time statistics tracking improvements worsening condition Consider whether using any other phone s sensors including camera and microphone could be beneficial I haven t done any work on gait analysis so not sure if these fit your interests but perhaps they can be adapted or could trigger other ideas Nigel Lovell UNSW Scientia Professor Graduate School of Biomedical Engineering University of New South Wales UNSW Sydney NSW 2052 Australia Nigel says that our research work sounds interesting but in terms of translating something that works in a lab to working in real life you would need to do quite a bit more work including clearly defining the clinical need for such a device At the moment it seems to be trying to do everything from gait analysis to falls prediction and pre
50. eurological and physiological disorders Released by nerve cells Dopamine is a transporter of electrical signals from one neuron to another In the brain there are five known dopamine receptors D1 D5 Men have 2 times more chances than women for getting the disease 2 3 Parkinson s Disease Symptoms 2 3 1 Primary Symptoms Motor Impairments Tremor The Tremor is 2 most apparent symptom after 9 Bradykinesia in PD Resting tremor is also common in which body part continuously shake when at rest In tremor one or more body parts continuously move round a fixed pattern 3 6Hz depending upon the severity Majority studies do not focus to differentiate between action and resting tremor The Tremor usually 10 starts shaking limbs on either side 3 out of every four subjects of the body and spreads all over the body in severe cases affecting jaws lips tongue Severe tremors in legs are the great cause of gait shuffling Details on Appendix D Bradykinesia Brady means slow and Kinesis means movement is the failure of basal ganglia which affects commands to move sometimes shuffling or sliding of feet on the ground Bradykinesia is the most common symptom in PD Other name for Bradykinesia is motion slowness Opening amp closing of hands can be used to estimate the severity of Bradykinesia It is the slowness of movements both spontaneous and automatic making simple or routine tasks bit more difficult to perform 11 In the early stage
51. ews 36 Many other gait analysis methods and techniques with can be described 5 3 Wireless Measurements 5 3 1 WISEPLA As the name wireless means no wires so it means comparatively less burden as compared with Wired Measurements in which enormous quantities of wires are used Wireless sensor network is good non spatial method to use Different types of gait of the patient with PD are tested using SHIMMER wireless sensor nodes SHIMMER Shimmer provides an excellent platform for gait a analysis as independent living technology 34 Different shimmer daughter boards can be attached to the main board for motion detection and save the information into a central data acquisition system for later pre or post processing The features such as recording information for more than 24 hours in the form of CSV files into 2GB MicroSD card in one time full charge of the battery are quite efficient low power options More over it provides the functionality to track information about ECG EMG GYRO Strain Gauge Heart Rate GSR Magnetometer etc So it gives the SHIMMER as good option to utilize it in the field of biomedical sensing rehabilitation gesture amp posture tracking gait analysis and motor disorder monitoring etc A Short range wireless sensor platform for gaits and it can be used in different implantable applications by Tekes A 3 D packaging of electronics and biocompatible encapsulation technologies are likely to be develope
52. f the Accelerometer X axis signal handles iAccelYShimmer handles shimmer getsignalindex Accelerometer Y Determine the column index of the Accelerometer Y axis signal handles iAccelZShimmer handles shimmer getsignalindex Accelerometer Z Determine the column index of the Accelerometer Z axis signal signalNameArray handles shimmer getenabledsignalnames Get the list of enabled signals signalNamesString char signalNameArray 1 1 Create a single string signalName sString for i 2 length signalNameArray which lists the names of the enabled tabbedNextSignalName char 9 char signalNameArray 1 i Add tab delimiter before signal name signalNamesString strcat signalNamesString tabbedNextSignalName Concatenate signal names delimited by a tab 76 end o dlmwrite handles fileName signalNamesString s Write the signalNamesString as the first row of the file else delete handles figurel end SSS SSS SSSSSSSS End 5 5 5 5 5 5 5 5 5 5 9555555555555 Startl 5 5 5 5 5 if handles shimmer start TRUE if the shimmer starts streaming processed buff size round handles shimmer getsamplingrate 4 siren low siren low fs wavread SIREN1 WAV siren high siren high fs wavread SIREN2 WAV arr counter 1 plotDataBuffer newData elapsedTime 0 Reset to 0 tic amp Start timer while elapsedTime lt handles captureDuration pause hand
53. for unbalance shaky gait for wellness But for this of course need some funding opportunities to work further over it 80
54. g Conclusions We have presented a real time gait analysis algorithm capable of detecting the motion of the patient with PD to actively respond to the shakier motion setting threshold values Our proposed algorithm is easy to implement reusable and can affectively generate healthcare alarms Additionally this system might be used by other researchers without the need to implement by themselves The proposed method is sensitive to detect fall therefore objectively can be used for fall risk assessment as well The same algorithm with minor modifications can be used for seizure detection in other disorders mainly epileptic seizers to alert health providers for emergency Keywords Gait Analysis Parkinson s disease Wireless Sensors Fall Risk Assessment Aging Wellness and Gait Event Detection TABLE OF CONTENTS Contents ABSTRAG PB A 1 dU DES EGO OM cR 2 1 25402250 9 1 1 AN E EE E EEN 9 1 2 NEUROLOGICAL DNO D R e n E EEEE E EE NEEE E E pap eei 9 1 3 GATE PARAMETER S 55 eee eee E E 9 1 4 10 1 5 AREA OB STUDY si n 10 1 6
55. gacy bootstraps contain three types of samples and each sample contains the following set of programs 1 AccelECG shimmer ihex Please select progremmer s port Reload COM ports v Select by file name D Staring Thesis Jenny Lundberg ShimmenFimware boatstraps Shimmerar sam Please selectbootstrep sd Open bootstrap file Please selectbootstep Lookin JE 2010 09 03 Shimmer2r Accel Sensitvities gt 4 El File name Date modifiec 2 AccelGyro shimmer ihex AccelECG stimmer2r 100Hz 15Gihex 6 16 2011 81 AccelECG_shimmer2r_100Hz_6G ihex 6 16 2011 8 1 2 AccelGyro shimmer2r 50Hz 1 5G ihex 6 16 2011 81 3 AnEx_shimmer ihex AccelGyro shimmer2r S0Hz 6G ihex 6 16 2011 81 HostTimeLogging_shimmer2r_50Hz_1 5G ihex 6 16 2011 8 1 i 4 ECG_shimmer ihex Fitename _ RccatGyro_ahimmacr 156 Fiesoftype files z Cancel 5 EMG_shimmer ihex L2 6 GSR shimmer ihex Figure 10 AccelGyro_shimmer2r_50Hz_1 5G 37 There are also special Shimmer Accel Sensitivities samples with most important are 1 AccelECG_shimmer2r_100Hz_1 5G 2 AccelECG_shimmer2r_100Hz_6G 3 AccelGyro_shimmer2r_50Hz_1 5G 4 AccelGyro_shimmer2r_50Hz_6G 5 HostTimeLogging shimmer2r 50Hz 1 5G 6 JustFATLogging shimmer2r 50Hz 1 5G 7 JustFATLogging shimmer2r 50Hz 6G For project we used BoilerPlate_shimmer2r ihex as default f
56. h thankful to BTH to allow us to work on thesis Wouter Speybrouck proved to be helping guy and always ready to assist whenever we faced problem during experimentation Inam ul Haq amp M Adnan Jalil Ronneby September 2012 Aim amp Objectives The Overall goal of this thesis work is to study different gait parameters to detect gait disturbance in the patients with Parkinson s disease by exploiting wireless sensor technology ie Shimmer to actively respond to patient s movements through implementation of algorithm Objectives To check gait related parameters in walking disorders using Shimmer platform e Wearing SHIMMER wireless sensors on hip waist and chest to check which one is the most suitable place e draw effective conclusion results in both real time and offline processing in EyesWeb and MATLAB e develop effective mechanism for security warning and activating alarm systems Algorithm 1 INTRODUCTION 1 1 Gait The way how a person walks is called gait Another good definition can be a coordinated action of neuromuscular and skeletomuscular systems Gait disorders resulting in uncontrollable walking patterns may be due to diseases such as Parkinson Epilepsy Arthritis Fracture Injuries Collision and many other diseases can cause neurological disorder or motor paralysis These disorders can be either in one body part or throughout the body without confining to any specific age group 1 2 Ne
57. hair CircleGait CircleGlassGait These above activities are somewhat similar to 10 53 6 1 4 Data Acquisition Data acquisition can be pre processing or post processing Pre processing is a real time scenario to track the information related to the gait of the patient While post processing is an offline analysis of data saved into the computer most commonly it is CSV file to calculate the standard deviation at different points in the gait to set threshold values We checked two software ShimmerConnect and the EyesWeb patch The parameters for configuring shimmer are Software burnt BoilerPlate_shimmer2r ihex Software burnt for EyesWeb AccelGyro_shimmer2r_50Hz_1 5G ihex Sampling Rate 100 MHz Accel Range 1 5g Motion data is saved into CSV Comma Separated Version files with relevant gait names in all experiments for post processing offline analysis to set threshold values based on the calculations of standard deviation at different points in each gait This threshold value is used later on in the development of real time gait analysis algorithm to detect and actively respond to the gait assessment process EyesWeb Patch We have developed EyesWeb patch to capture motion data with a constant generator for particular axis motion with value 0 1 2 x y z respectively Hence experimental results are collected on different gait parameters to capture data on specific axis This EyesWeb patch is developed to capture only o
58. he cost of falling is estimated 28 2 billion during 2010 Attaining one of the goals of wellness older people should be avoided from falling through proper in time planning 2 5 1 Factors Responsible for Falling Despite Parkinson s disease there are some other factors responsible for falling Weak muscles and bones Week vision Getting aged Environmental factors hurry slippery Drug addiction Qu CON EU qus Back bone diseases 2 5 2 Fall Detection Fall could be characterized as a potential change in the horizontal movement of the patient A phenomenon of unintentionally falling is caused due to different movement disturbances of gait Technologically fall detection can be divided into three main categories 17 Fall Detection Techniques 2 5 2 1 1 Vision based Fall Detection 25 It includes different types of fix cameras which can continuously record the patient s activity A Vision based fall detection system contains data acquisition system to collect images implementing different algorithms that can recognize fall patterns and may trigger alarms An advantage of vision based fall detection technique is the patient has no need to additionally wear anything However there are some privacy concerns and probably difficult to use outdoors Few of these methods are l A 3D Centroid Position system for fall detection with the detection of inactivity 18 2 Automatic Detection of human fall in v
59. he value from motion data exceeds maximum threshold value 2 Later on airbag can be put on the patient s hip position to avoid him her from injury and hip fracture The results show the proposed system is fairly simple to implement in the real time environment flexible to adjust to any necessary change in the future The major advantage of this algorithm is its reusability Algorithm is not hard coded because a user can set his own sampling rate or threshold value or both and check results This algorithm is further modifiable to trigger airbag a security push button SOS calls messages siren activation system automatic email forwarding health care alert and many more The same algorithm with minor modifications can be used for fall avoidance or health care assurance on other disorders mainly in epileptic seizers to alert health providers in case of emergency can be used for other seizures and disorders such as epilepsy etc Overall this report presents the analysis of an experiment to measure the usability of wireless accelerometer data to monitor the activity of the patient suffering from Parkinson disease Our research and experimental work can be quoted toward fall risk assessment THE END OF CHAPTER 6 7 Findings Although accelerometer on the chest shows higher acceleration during fall hence unnecessary body motion create more noise in the calibrated signals 66 Also if a patient initiates steps with energy his her gait
60. ideo 19 2 5 2 1 2 Environmental This approach is based on the installation of devices such as sensors in a certain area that monitor the patient s activity Such environmental approach has some benefits such as patient do not need to wear anything However that approach may confine to limited range and difficult to use in outdoors Few of these methods are i A Floor Vibration based Fall Detector 20 1 Slip fall Detection System using the sliding linear investigative platform 21 2 5 2 1 3 Wearable These devices usually contain motion detectors called sensors such as gyroscopes and accelerometers One such platform is provided by SHIMMER technologies Using such systems sensors can transmit data to the Data Acquisition System apply different fall detection algorithms and then may trigger alarm Wearable approach has some advantages such as it can be used both in indoor and outdoor easy installation smaller and light weight Few of these method methods are G Integrated Approach of Waist mounted Accelerometry 22 ii A Wearable Airbag to Prevent Fall Injuries 23 Which technique is better Well it is observed that the third fall detection technique seems better than other two because it has more advantages such as light weight smaller compact can record motion even for days data acquisition in CSV files enable any functionality like Gyroscope Magnetometer etc with daughter board 2 5 3 Fall Avoid
61. ies Researchers believe these bodies to be one of the reasons for causing Parkinson s disease e patient with PD has impaired neurons in the brain area called substantia nigra These neurons produce a chemical called Dopamine and due to its reduced level it affects smooth communication between muscles for movements The result is a loss in the ability to control the proper functioning of body movements Parkinson Symptom Motor impairments Tremor Bradykinesia T Impairments Non Motor Rigidity Postural Instability Festination Freezing of Gait FoG Neuropsychiatric Symptoms Autonomic Nervous System Figure 4 Primary Symptoms in Parkinson s disease e Why level of Dopamine is reduced is still unknown without any doubt e PD may be inherited although controversies exist e Some evidence shows that certain toxins in the environment such as manganese carbon monoxide carbon disulfide and some pesticides herbicides may cause PD 18 e Oxidation is thought to damage neurons e Some other drugs can cause symptoms similar to Parkinson s disease It is still an open research problem to solve the exact cause of PD In fact Dopamine is a chemical in the brain responsible for functioning as neurotransmitter among the neurons Its reduced level causes different n
62. iii Antidepressants such as Duloxetine for anxiety and Gabapentin for neuropathic pains iv Sildenafil for erectile dysfunction It may have some negative impacts on kidneys Most common side effect of drugs is Hallucination Especially 7 10 L DOPA is reached to the brain due to the fact that Benserazide cannot cross blood brain barrier while remaining amount is assimilated into the other body parts causing nausea and dyskinesia So Benserazide in combination with L DOPA is used to treat Parkinson s disease Loss of balance or postural instability cannot be treated with medicine Deep Brain Stimulation Generally DBS is recommended in those circumstances where use of Levodopa has reduced or no effect in the brain So it is good option and sometimes gives good results Problem with DBS is that it cannot decrease the progression of the disease Other side effects may be brain bleeding cognition decline strokes infection and minor symptoms of disease may last However it can relieve some symptoms and enhance wellness up to five years after surgery 73 Appendix E Read Acceleration xyz Load data into File d 25 values received Save File csv Yes M Calculate std m Message Device is not Working Yes Std m Threshold2 551 Activate Emergency Alarm Std m gt Threshold1 417 Activate Warning Alarm No
63. ini Mental State Examination MMSE MMSE is a brief questions questionnaire to assess patient s level of cognition It is normally used at the later stages of disease when the patient poses weaker mental status such as a short time memory loss and dementia Some important brief questions are What Is the Date Today Month gt day year correct score 3 e What Is this Called Watch Please Repeat the Following No Ifs Ands or Buts Perfect 1 Please copy this drawing code 6 if low vision Hospital Anxiety amp Depression Scale HADS HADS is a 14 questions 7 anxiety and 7 depression one page clinical scale to assess the level of anxiety and depression of a patient 45 5 1 ok 5 1 5 2 9 Fall Prevention Home Assessment Chart It is a simple Yes No questionnaire for fall prevention when the patient stays at home Important parts include questions using Bathroom Bedroom Kitchen Porch and Living Areas An answer No means need improvement 10 Personal Risk Factors Checklist It is a simple 10 question Yes No questionnaire to assess the risk of fall with recommendations to each Yes responses 11 4 Meter Go amp Back Test 4GB 4GB is a 4 meter walkway designed by us to monitor the unbalanced gait of the patient with wearable sensors Different types of gaits are formulated in order to give the patient more challenging environment during walking Video is recorded for offline vali
64. is very big Also change in orientation accelerometer may change readings so this issue also affects the test retest reliability of the measurements To have a robust outcome you either have to rely on features of the signal that are invariant to the orientation of the sensor e g the norm of the acceleration vector or develop a method to compute and compensate the orientation of the sensor in relation to the body either automatically or using a protocol of functional tests He quoted that idea of using the 4GB test for this purpose is good but is not enough Such tests are good at assessing sensitivity of the detection method but are not enough to assess specificity or the positive predictive value In other words these protocols do not give you a reasonable estimate of false positive as well as true negative rates He also commented that GaitAnalyzer may not be suitable for certain types of gait such as stumping gait There are also possibilities that during gait analysis you might record very small variation of acceleration signal e g when a PD subject is walking very slowing specially during OFF state User selectable thresholds might appear to be a good idea but in 78 fact are an important limitation Who is going to set the thresholds Patients Clinicians Also depending on the type of activities thresholds might need to be changed Thus ideally system should use adaptive thresholds that are automatically selected based on c
65. ittle jump Here Y axis becomes sensitive means stopping and fairly conscious walking pattern 51 75 2501 2525 24054 15432 170 16 154 19 147 72 108 60 Note 7 Gaits are given No 1 is standard deviation of total samples for xyz acceleration Majority of values of interest naturally fall under Y acceleration column due to accelerometer orientation Maximum accelerations are underlined X axis is more sensitive Le itis verticular to the surface Exp A Exp B Exp A Bare conducted on same day with half hour pause Both A B put into Experiment 02 Table 4 Experiment 02 shows 6 different types of gaits with higher std deviations after each 25 samples In this chapter we propose a method for tracking the real time motion of the patient using wireless sensors The method is based on analysis of calculating standard deviation on different gaits The collected data is analyzed and various interesting features are extracted Based on these features algorithm with name GaitAnalyzer is implemented to track and monitor the gait of the patient The algorithm has been evaluated with three experiments on different time slots The good aspect of the method is that it contains a smooth slider for setting threshold values 1 and 2 Both these values can be set to different standard deviations to capture the motion and saving into the computer for later offline processing Plotting in parallel and ON OFF for alarms representing at the end are a
66. king but also during sit stand from the chair getting out for water or taking clothes from laundry Patients cannot completely avoid freezing but can adopt certain strategies medications therapies nutrition and exercises to help controlling freezing About 50 patients with advanced PD experience FoG more in men 10 in mild PD and 80 in severe PD 14 Detail on Appendix D 2 3 2 Secondary Symptoms Non Motor Impairments Patient has not control over few symptoms such as impairment of cognition behavior voice and brain and these are commonly included in non motor disorders Brain disorders include weakness of thoughts and dementia Neuropsychiatric Symptoms It deals with all those symptoms that are related to Neurology and Psychiatry It involves change in mood thinking behavior reasoning perception judgment Depression Depression is one of the main problems faced by almost all the patients It is estimated that nearly half of patients with PD experience depression due to one or many reasons such as losing hope lesser interest and little enjoyment in life Re occurring of disease after medication with Levodopa is another cause of anxiety Mild Cognition Loss Some patients experience mild cognition loss at early stages especially it affects the sense of self planning and management On later stages these patient experience a severe condition known as dementia Other cognition related symptoms in patients with Parkinson
67. larm c Alarm system 8 Figure 14 Data Acquisition to Track Motion This study is carried out on a patient with Parkinson s disease to observe various gait 5 6 Findings for warning and emergency alarm activation THE END OF CHAPTER parameters to test wireless accelerometers on different body parts and to implement an algorithm triggering security alarm system by setting threshold values Criteria for setting threshold value are based on calculating standard deviation and finding two threshold values 4 Meter Go amp Back 4GB test has been designed and tested This test may be good in planning and taking the measurements necessary for experiment The different type of gaits gait parameters and instructions are also provided in the test 50 6 IMPLEMENTING ALGORITHM amp RESULTS In this chapter we would document how we proceed in conducting experiments taking measurements calculating standard deviation and implementing the algorithm 6 1 Experiment 6 1 1 Patient History medical and fall history The lady with PD has undergone two surgeries during 2009 Subject is 70 years an old healthy patient with PD and has undergone Deep Brain Stimulation surgery The detail about DBS is written in Chapter 2 The reason of DBS is that the patient is not feeling comfortable with medication and hence posing severe shaky motions which has reduced after surgery but symptoms still exists Perhaps the use of Le
68. le may freeze for up to a minute In short Rigidity is the loss of arm swing and facial expression 5 tiredness freezing of limbs or shoulders and reduced initiating of muscle movement Freezing of Gait Freezing is not necessarily associated with all patients but there are more chances if the patient is older remains depressed and is being medicated with Levodopa for a longer time Sometimes stepping becomes too fast which results in freezing or falling So try to slow down stepping in a rhythmic pattern But the biggest risk associated with falling is unpredictable Somehow a patient can only be judged by examining his gait behavior turning patterns and sit stand positions Chance of falling in elderly people gt 60 is up to 30 and therefore associated injuries and cost of falls become high 15 Nearly 8 patients observe frozen shoulders as early symptoms sometimes feeling stiffed and painful It is closely resembled with Postural Instability where a patient with a shuffling gait loses his balance and co ordination So due to poor coordination between reflexes PD patients have an increase risk of falling especially in later stages There does not exist an exact technical device to detection FoG 30 Reason for freezing is yet not known Treatment Rasagiline and Selegiline basically belongs to MAO B a class of antidepressant drugs are also used to increase Dopamine level in brain basal ganglia Amantadine for treating early tremors
69. les DELAY PERIOD Pause for this period of time on each iteration to allow data to arrive in the buffer newData handles shimmer getuncalibrateddata Read the latest uncalibrated data from shimmer data buffer if isempty newData TRUE if new data has arrived dlmwrite handles fileName newData append delimiter XE E o file in Append the new data to the a tab delimited format 555555555555 StartIl newData handles iAccelXShimmer newData handles iAccelYShimmer newData handles iAccelZShimmer X l std_x std double x std_y std double y std_z std double z std arr arr counter 1 std x std arr arr counter 2 std y std arr arr counter 3 std 7 arr counter arr counter if std x gt handles Threshold2 Il std y gt handles Threshold2 Il std z gt handles Threshold2 handles shimmer stop Stop data streaming display Alarm High ON wavplay siren high siren high fs handles shimmer start Start data streaming elseif std x gt handles Threshold1 std y gt handles Threshold1 ll std z gt handles Threshold1 handles shimmer stop Stop data streaming display Alarm Low wavplay siren low siren low fs handles shimmer start Start data streaming else display Alarm OFF end T Appendix H Scientific Discussion Validation on Results As our supervisor Dr Jenny Lundberg directed to take inpu
70. loss of balance or postural instability cannot be treated with medicine Problem with DBS is that it cannot decrease the progression of the disease Other side effects may be brain bleeding cognition decline strokes infection and minor symptoms of disease may last However it can relieve some symptoms and enhance wellness up to five years after surgery Gastroparesis is one such symptom in which food stays in the stomach for more than two hours Mediterranean diet can be introduced in the patients with PD 30 31 3 HARDWARE Wireless sensors such as accelerometers are used in this thesis project for providing sensing solutions to track motion data As our thesis project is related to the area of the wireless sensor network and body area network so brief description for these terms is given below Wireless Sensor Network WSN is a low cost low power a scalable network of nodes that consists of spatially distributed sensors used to monitor physical conditions such as temperature pressure distance etc WSN has the ability to pass data to a central server or data acquisition system which can be further analyzed to take necessary measures Body Area Network BAN is a group of tiny sensors wearable by an individual with some central unit to regulate communication A wireless BAN uses PAN for scalability and the vital applications include logging information about cardiology diabetes neurological disorders etc 3 1 SHIMMER Platfor
71. lso good aspects of this algorithm Results and simulation show that method is fairly simple and sensitive to track the real time gait In order to take the necessary measurements systematically a test or scale was necessary So 4 meter Go amp Back test 4GB is designed and it is observed that this test may provide help for future researchers to take the necessary measurements in their experiments to get expected results 60 Some of the features described in Table 6 2 first three gait types show fairly healthy a little jumping fashion When the patient is carrying two glasses in each hand in GlassBoth Gait a fairly balanced gait is observed But at 5 Sit_Stand_GlassBoth gait fmax 551 31 is observed when the patient takes a jerk during sitting This higher value of standard deviation also suggests setting threshold to higher values But we need to analysis data captured during experiment 02 In the table 6 3 experiment 02 again first three gait types are fairly balanced however FastBothGlass gait gives fmax 272 92 This is a bit higher acceleration Two gaits BothGlassGait and CircleGait also yield higher acceleration CircleGlassGait however yields maximum acceleration i e fmax 240 64 Above gait does not have higher acceleration as it was expected It can be helpful in suggesting setting threshold value While in physical examination patient was observed losing concentration during circling
72. m SHIMMER stands for Sensing Health with Intelligence Modularity Mobility and Experimental Reusability Ireland based SHIMMER may help 40 researchers to use a core IT technology to create medical applications and devices for the full range of kinematic modules such as 3 axis Accelerometer 6 axis Accel Gyro and 9 degrees Accel Gyro Magnetometer for freedom of motion It also includes wide range of sensing boards devices such as ECG GSR Magnetometer Stain Gauge EMG Heart Rate etc Some of important applications of using SHIMMER is given below e Motion tracking and streaming of biomechanical data e Gestural computing and human computer interface research shimmer Figure 6 Shimmer XYZ e Sport technique analysis and athlete development e Rehabilitation assessment e Motor disorder monitoring e Gait analysis e Navigation and tracking of objects or people 32 light indication when going out of range is a unique feature heart Low offset precision amplification Realtime Data collection storage and display NAME PURPOSE SPECIFICATION PHOTO APPLICATIONS 3 axis vibration MSP430 microcontroller 1 5 6g sensitivity Baseboard 3 colors LEDs reset 5 p ERSTE 555 Realtime Data collection storage and display Capture acceleration e 8Mhz 10 RAM 48KB Flash 2GB MicroSD AR Accelerometer xyz E cc Measures tri axial acceleration 802 15 4 Radi
73. motion analysis In the wireless sensor network SHIMMER platform provides wireless Body Area Network BAN to capture motion data This data can be saved in CSV Comma Separated Version file for post processing or a 2 GB MicroSD card can be used to capture data in the SHIMMER accelerometer itself The use of accelerometer is more suitable due to the fact that we are 65 capturing data from postural instability One two or combinations of accelerometers can be put on different body parts SHIMMER Gyroscope is more suitable for jerky motion with disease such as epilepsy Mostly accelerometers and gyroscopes are used for gait analysis 4 Defining our research work this study is carried out on the patient with Parkinson s disease PD to study various gait parameters test wireless accelerometers on different body parts and implementing an algorithm to trigger a security alarm system by setting a threshold value Criteria for setting threshold value are calculating standard deviation and employed by different researchers like 3 The main motivation to perform this experimental research work is to avoid the patient with PD from falling during unstable shaky gait Security alarms can be activated whenever a patient poses a shakier gait Two types of alarms or sirens can be activated in the algorithm First to activate Warning Alarms when the value from motion data exceeds maximum threshold value 1 and second to activate Emergency Alarms when t
74. n this system the user can on a portable sensor that is attached to the body as shown in the figure whereas information of its readings can be tracked through a computer It can measure ECG signals heart rate brain signals velocity body temperature and respiration with a long life battery for up to four days So this system provides a good and convenient service to remotely monitor the patient s health care activities especially for the elder people Wired Measurements It is a universal understanding over the fact that wired networks always provide fast and reliable communication of information Yet these networks are becoming out dated due to the reason that these are expensive up to thousands of dollars 37 38 and difficult to install especially retroactively into an existing structure Some such wired systems are given below 48 5 4 1 Nanol7mm 6 axis Force Torque Sensor A novel protocol called Advanced Sensing for Assessment of Parkinson s disease ASAP is used to 5 measure the patient s grip force trying to follow a sinusoidal force target wave For this purpose a 9 gram 17 mm diameter measuring for all 6 axis force torque is presented which is the smallest in the world Some typical applications may include e Haptics Rehabilitation e Robotic Hand amp Surgery Telerobotics e Finger force research This sensor contains an intelligent data acquisition system for data interpretation thr
75. nce 1 3 Measurement devices or marker sets which are attached to body parts will affect the natural undisturbed gait A bit related work is done by Fei Yu et al 33 to develop wireless medical sensor devices for ECG EMG BST eye movement motions continuously while the software for signal processing and data recording is implemented in LabVIEW There can be a number of assessment methods for human motion and fall Some of them are video capture visual examination interviews keeping diaries physical measurements questionnaires and wearable sensors In wearable sensors most common are accelerometers gyroscopes magnetometer strain gauge pedometer actometer etc 4 The Difference between Gait Assessment amp Gait Analysis Gait Analysis can be defined as Gait analysis is the study of the biomechanics of human movement aimed at quantifying factors governing the functionality of the lower extremities This is crucial for the detection of gait disorders 26 Gait Analysis is the evaluation of walking style by observing human as he walks in a straight line 5 In a broader sense gait analysis includes assessment planning and treating people with conditions affecting their gait So gait analysis allows to assess the gait for different walking disorders Gait assessment is the evaluation of walking patterns in order to differentiate dysfunction It includes different assessment tools and scales as mentioned below
76. ne axis and we are interested in y axis setting constant generator 1 in experiment 01 and x axis setting constant generator 0 in experiment 02 So when Constant Generator 2 it means that we are capturing Z axis motion The XYZ acceleration in Real time EyesWeb Patch has been shown above in the Figure 12 13 respectively The purpose of using EyesWeb is that we want to validate our motion analysis results So this tool is another method to verify the results in parallel to the ShimmerConnect default software Algorithm Implementation in MATLAB Some important Matlab commands and features used in the gait assessment are as under csvread FILENAME CSV 1 1 This command is used to read CSV file leaving the first row and column and placing data into variable M for later use Plot M Analyzing graph for variable M 54 Std M Std m shows how far data is distributed from mean such as if STD 238 and maximum for x axis then this shows that there is more deviation in this axis more acceleration Plot M 1200 1300 2 END Ploting for graph between samples 1200 1300 only Std M 500 600 2 END To see standard deviation between samples 500 600 only Hist M Hist m command is used to see which axis has more acceleration in histogram through pictorial representation Curve Fitting Toolbox This graphical tool is used for curve fitting and surface to data for the plots This toolbox is important to use for post processing of da
77. ns on it for fast slow and other gaits setting threshold value to inform the patient about his postural instability Risks 1 Should also try other Shimmer firmware samples amp battery time gyroscope 2 Change of placement of accelerometer may change results pilot study 3 Algorithm may not give efficient results during stop restart session same time let the algorithm run smoothly at a distance gt 25 steps away Blue Tooth range Admittance We admit that patients with PD can also be treated more efficiently with Electroencephalography diet behavioral study medication and therapies However our 64 study can have benefit for individual researchers and industry to devise a tool to assist patients in postural instability suffering from different neurological disorders 6 6 Summary Any malfunctioning of neurons in the nervous system is called Neurological disorder Over 100 neurological disorders have been discovered throughout the world In our study we have chosen one disorder Falling in Parkinson s disease Experiments can be performed on different gait parameters like body velocity time ratio ground slope stance swing body gestures and gait patterns Sensors can be put on hips knees thighs limbs neck head chest or any other suitable body part to capture motion data for further pre and post processing Pre processing is real time gait analysis through time domain and frequency domain to trigger vario
78. o steps sec steps sec Mean of A B 1 NormalGait 1 57 1 73 1 31 1 52 2 FastGait 1 61 1 54 1 13 1 33 55 3 BothGlassGait 1 41 1 3 1 3 1 30 4 Sit Stand Walk 1 29 5 Pick up Glass 1 10 6 FastBothGlass 1 54 1 58 1 56 Table 2 Step count Number of steps taken by the Subject We compare Experiment 01 two weeks before with Experiment 02 Column wise in the above table it is clear that the patient looks healthier in experiment 01 as compared to experiment 02 This fact is also validated by matching the results with video recordings Moreover Experiment B yields reduced steps sec because the patient becomes tired less interested or loses concentration in taking measurements Row wise in the table when the patient is forced or have to do some extra task to perform like holding glasses in both hands in BothGlassGait and Pick up Glass then the number of steps per second decreases sharply This suggests that hands of patients with PD are needed to be free for smooth walking In the Sit Stand Walk gait patient poses reduced walking patterns and same is the case with Pick up Glass gait as shown in the table above Underlined CircleGait is matchless here because the patient is resistant to follow tags in the circle and she has to do extra hard work to concentrate Comparing CircleGait and CircleGlass with other gaits it is clear that the patient
79. o Roving Networks RN 42 Sapper gyro magneto Size 53 32 15mm weight 22g Sleep mode Batterylile uu s Capture gyro data Tri axial angular sensing Range 500 sec Ao Compier morc racing uf Gyroscope br is biomechanical data 2 Perpendicular z axis Daughter board reset button M gnetic Sensing ditat compass IC for sensing slow motion Tracking objects biomechanics Magnetometer T Gain 390 1620 counts milligauss 15g pinhole reset i im 2551 xyz FH rehabilitation sports training healthcare r 15 times less power as compared to gyro Daughter board 3 leads ECG EKG digital signals Daughter board 28g ECG Electrical activity of Frequency range 05Hz 159Hz Collect and display electrical signal of heart Stain Gauge Stress measurement 28g low high gain output Fall Risk Assessment Frequency range 2 Excitation voltage 2 8V 5 Connections 3 5mm jack daughter board Pressure measurement muscular skeletal amp sport research Electrical activity of Daughter board 28g current 180uA Gain 682 Measures electrical activity of muscles EMG 2 t muscles Frequency range 5 482Hz nerves fatigue tremor amp Ergonomics Daughter board 28gFrequency range DC 15 9Hz GSR Stress Measurement e Frequency range 15 9Hz Current 60uA 4 Stress Detection amp Analysis Measurement range
80. ognizes and records motional activities of a person For this MotionBand sensors are attached to the smart phone to capture accelerometry data Among 15 Six activities to be detected are running walking and cycling The authors suggest that this idea of activity recognition can be applicable to other applications such as harmonizing the music with the activity we are engaged in for example low music at rest and fast music during running Another interesting application can be to automatically learn personal habits such as disturbance can be reduced in the office using the concept of Polite Calling in which system redirects phone calls to voice mail showing the person is busy THE END OF CHAPTER 16 2 PARKINSON S DISEASE 2 4 Introduction Parkinson s disease is a progressive incurable most commonly observed in elderly population above 50 The patients with Parkinson s disease show a reduced unbalanced and robot like walking pattern While other results show that PD patients keep 29 slower velocity and nearly 2096 reduced stride length than the Healthy Elders while all PD 7 subjects show abnormal gait patterns In Genetic 8 PD there can be more than ten genes associated with PD In Sweden out of 14000 PD patients 550 700 have inherited background no data less than 10 10 20 20 30 30 40 40 50 50 60 60 70 70 80 pare em Parkinson s Disease World Map 90 100 100 1000 more than 1000 Figu
81. oping industrial educational mathematical engineering and biomedical applications For this facility Intel has chosen EyesWeb in its project Independent Living in 2008 Hundreds of applications can be designed for audio video media math 3D rendering and auditory Before proceeding with any new software it is always good to start with its demo application For EyesWeb a demo application is SHIMMER_Demo_Patch eywx 4 5 1 Graphical Development Environment it provides the functionality to create EyesWeb blocks in Workspace as shown in the Figure 11 7 7 777777 Evesweb Development Environment EywPatch1 2 0 mx File Edit View System Tools Window D Uag amp 10 da ew Lbrary View Properties s x EywPalchi C F No Fitter ET Params Descr Profiling gt j Z Switch m yle Trigger inputs changed i henny AlphaFilter Catalog View FIRFilter BB HibenTransi y Lineaifjaer nr iher d FO Converter F Generator Property View d Double matrix In generat Patch View Fb ReadMatixFromF ReadMatinF omit VectorGenerator Message Console a ter Item Selected 1 indexComparein Select an tem bon Bl Matix Arithmetic 7 nm Ready Figure 11 EyesWeb GDE Main View 4 5 2 EyesWeb Blocks amp Patches
82. or ShimmerConnectV2 0 for fall detection in Parkinson disease While for EyesWeb we developed customized application with AccelGyro shimmer2r 50Hz 1 5G ihex as shown in the Figure 10 4 2 SHIMMER Connect This software is used to connect SHIMMER with PC or mobile device It contains shimmer configurations sampling rate sensors to sample Accel range GSR range shows graphs plot to see real time motion data in time domain and Saves to CSV an Excel extension 4 3 Demo Applications and Samples In order to check a device for the first time it is a good idea to run with demo applications provided Shimmer has provided different such demo applications as discussed above We also tested demo programs on Matlab BioMOBIUS and EyesWeb 4 4 Mobile Symbian Application World has changed towards handy wireless communication devices such as smart phones So Shimmer has provided a customized application for Symbian s60 mobile phones to store amp display Accelerometer and ECG motion data Future work could be to design customized applications for specific motion data on mobile phones 38 4 5 EyesWeb Developed by InfoMus Lab EyesWeb is open source software to design real time applications for physiological kinetic signals and Gesture Analysis A user can develop his own application using EyesWeb Development Environment which gives the facility to create multi model interactive programs EyesWeb provides scientific research for devel
83. ough the data acquisition card plug into the pc It also provides high signal noise ratio 5 4 2 Digitising Tablet Another measurement method is represented by a 5 digitizing tablet to analyze the Micrographia of PD patients Micrographia is impairment in 3 poor small shaky and abnormal handwriting Another advantage is the facility to record the upper limbs tremor of PD patient in 2 D information at accuracy rate 200 lines per inch Tremor rate cannot exceed 12 Hz while this tablet operates at a sampling rate of 50 Hz 5 5 Our Gait Analysis Mechanism Technically wearable wireless sensors provided by the SHIMMER are put on the different areas of body to remotely track the information about the movement of the patient with PD This information is captured and saved into the PC through the data acquisition system for interpretation and offline processing Depending on different requirements real time implementation through algorithm can be performed on real time motion analysis to trigger different healthcare facilities such as emergency alarms automated SOS calls messages health givers alertness etc 49 shimmer Wireless M acclerometer ES Data Acquisition Offline real time csv file Real time Data processing and feature extraction amp algorithm 25 samples calculating Std x y z Motion Detection Threshold gt 225 Warning alarm Threshold 330 Emergency a
84. oving daily life health facilities and wellness solutions 1 7 Problem Statement Most of the patients suffering from any neurological disorder at their later stages of disease pose ambulatory disturbance Such patients may even fall without showing any warning sign Falls are considered one of the major causes of injury disability cost and mortality For above different gait related parameters need to study There may need a systematic tool to take measurements as well 10 Finally there is a need to develop a mechanism to detect any gait deviation notifying 3 party in the form of security alarms 1 8 Research Questions So particular research questions may be formulated as Q 1 What are how to use gait parameters for gait analysis Q 2 Based on specific parameters how to a Formulate effective mechanism to monitor movements b Trigger messages alarms button for patient care 1 9 Evaluation of Gait Analysis In order to assess amp evaluate gait analysis accurate reliable amp consistent measurement tools need to be utilized Even slight deviation in data monitoring through measurement tools is not encouraged to use 3 Gait disturbance can be measured using 3 axis accelerometers like SHIMMER for real time motion analysis In the wireless sensor network SHIMMER platform provides wireless Body Area Network BAN to capture motion data This data can be saved in CSV Comma Separated Version file for post processing or
85. pecial devices and facilities can be used such as shower stools grabbing sticks and wall handles 2 5 Falling in Older Persons Falls 11 15 are considered one of the major causes of injury disability cost and mortality Half of the PD population aging more than 65 has a tendency to fall and nearly 80 patients need hospitalization in this age group Fall prevention can be achieved to some 24 extent if a patient is analyzed with his gait pattern Studies show that the reduced walking rate and weak grip are two other contributing factors towards falling Normally falling can happen during Festination and Postural Instability The patients are also likely to fall during freezing in their gait Psychologically these patients are prone to low fall rate and feel fear of falling Both low fall efficacy and fear of falling are correlated yet have different dimensions Keeping a diary for fall happening 16 during FoG is another good idea given by Albert et al and this record keeping 13 can be proved helpful for doctors researchers and therapists The patients suffering from osteoporosis are more likely to get their hips fractured after fall so studies suggest these patients to build strong muscles and bones through exercise protein enriched nutrition along with medication and undergo fall risk assessment Statistics projects that the population of age group 65 is growing up to 4 3 during 2050 according to US Bureau of Census even then t
86. peech swallowing handwriting dressing hygiene falling salivating turning in bed walking cutting food Section Evaluation of Motor Examination Section Complications of Therapy Section V Modified Hoehn amp Yahr Staging Section VI Schwab amp England Activities of Daily Living Scale 44 5 1 4 Instrumental Activity of Daily Living Scale ILADL IADL is almost same as MFE scale with the difference between both is that it is formulated to evaluate more complex daily routine tasks with detail interview based on a questionnaire It contains 8 items 10 15 minutes 0 8 scoring Some of required questions in detail are Mode of Transport Food Preparation Housekeeping and Ability to Handle Finance Activities of Daily Living ADL ADL is actually an Index of Independence in Activities of Daily Living It is designed to assess normal patient daily life activities based on dependence 1 point independence 0 point with total 6 points means the patient is highly independent and 0 point means patient is highly dependent Daily life activities include bathing shopping sit stand patterns housekeeping cooking etc Timed Up amp Go Test TUG TUG is a famous time based test to assess patient s moveability Normally patient activity of sit stand and vice versa on the chair with and without arm along with turning patterns is evaluated comparing with the criteria of time in the test M
87. ps and messages for patient care Post processing is offline analysis of motion data in different tools such as EyesWeb BioMOBIUS and Matlab for calculations analysis and plotting of motion to take decisions to formulate a mechanism for patient activity detection and monitoring 1 5 Area of Study The area which we choose is pretty interesting pertaining to rehabilitation wellness and healthcare for older people If a research string or query is formulated then following keywords may be helpful using one or combination of more than one WSN BAN or WBAN biosensors neurological disorders gait analysis fall detection fall avoidance Parkinson s disease wireless accelerometer ambulatory monitoring freezing of gait wearable sensors real time gait analysis remote data acquisition health monitoring monitoring motor fluctuation gait fluctuation shaky gait gait disorder gait event detection wearable embedded sensors movement disorders neurodegenerative disorders fall risk assessment technology for elderly 1 6 Scope of Study Wireless Sensor Network WSN technologies has become business of billions of dollars 25 Billion in 2012 for healthcare services for indoor patients monitoring chronic disease and elderly persons Few in such applications are wireless ambulatory cardiac amp diabetic monitoring systems and tracking different neurological disorders These applications are increasing during the past few years for impr
88. re 3 Parkinson s disease World Map with Courtesy Wikipedia Mainly Parkinson s disease can cause an effect on both motor and non motor movements Five most common symptoms related to Motor movements include Tremor Bradykinesia Rigidity Festination and Postural Instability While in non motor movements autonomic malfunction and neuropsychiatric problems are common as shown in the Figure 4 At advanced stages of PD patients have more slowness and stiffness in walking and reduced muscle movements As rigidity increases and mobility decreases patients feel more pains during motion In normal gait heel strikes the ground before toe but in case of Parkinson s disease flat foot strikes with small stepping during the stance phase and reduced lifting of feet during swing phase 17 In general a patient with PD has an unstable stooped walking pattern tends to lean right or left during turning head down small stepping and reducing arm swing 2 2 Etiology of Parkinson s disease The exact cause of the disease is still unknown or idiopathic yet there may be some other factors which somewhat can contribute toward the disease These factors can be genetic environmental toxins or low Dopamine level in the brain These and some other factors that can constitute toward happening of the motor symptoms of the disease are stated below e Abnormal collections of tiny microscopic proteins inside nerve cells of the brain are called Lewy bod
89. ring Circle_Gait subject may fall or feel dizzy or get tired ii Good to perform trials before actual measurements iii Save files immediate after each gait with same Gait Name iv Person operating camera for video shall call names of each gait in case of repeating of gaits Note any sudden uncertain condition during after taking measurements Comments Recommendations Feedback 70 Appendix B PSEUDO CODE Real time 1 Start 2 Set Std 2 0 3 Set Threshold 250 4 Set Threshold2 220 5 Initialize VAVLUE counter 0 6 Read acceleration xyz matlab 7 Load data into File temp 8 While VALUE counter 24 9 Calculate std temp 10 Save into file 11 IF Std gt Threshold2 12 Activate Emergency Alarms 13 Print Patient is about to fall 14 Set VALUE counter 0 15 ELSE 16 IF Std gt Threshold 17 Activate Warning Alarms 18 Print Take Care 19 Set VALUE counter 0 20 ELSE 21 Print Device is not working 22 End Temp is File 3D array i e int temp 24 Sampling rate 100Hz 100 values sec 25 values per 1 4th sec Emergency Alarms active 10 seconds Warning Alarms active 5 seconds Algorithm continues reading next 25 samples in either case 71 Appendix C Comments Std Standard Deviation Threshold1 417 Threshold2 551 CSV File m Note Read File csv v Calculate Std xyz
90. s 152 06 126 150 samples 347 01 301 325 samples Sightly higher values show bit jumping gait 8842 6251 3516 total samples 6522 25 70 126 150 samples 8158 11111 2676 2700 samples 5872 15464 2726 2750 samples Fairly balanced gait ie No 12 Slightly higher values have all accelerations almost close show bit jumping zait 8851 2774 samples 9271 501 525 samples No 1 2 3 Fairly balanced z 11837 2761 total 3437 351 375 sit stand 25214 1001 1025 samples 17320 1776 1800 samples 55131 2626 2650 stand sitsiect No 6 Value 551 suggests to have higher threshold gt 551 15942 12551 14809 2493 total 11871 18120 10015 751 775 samples 5081 3046 6742 431 473 glass pickup 1226 1250 samples Leans right during turning 1476 1500 27 glass pickup 295 02 41738 25186 39 78 5257 52 90 Comments Both glass picking up is fairly balanced 1 glass pickup 1 second bit smooth 417 shows risk to fall Note amp Gaits are given No 1 is standard deviation of total samples for xyz acceleration Majority of values of interest naturally fall under Y acceleration column due to accelerometer orientation as shown in the Figure Maximum accelerations are underlined Table 3 Experiment 01 shows 6 different types of gaits with higher std deviations after each 25 samples Consciously slow gait during turning makes the patient s gait bit stable so it is suggested to be careful more during turning or circling in any case as shown in the Plot 1
91. s nsf all a3adcbc475746ee1c12578b700532 b77 OpenDocument Accessed 02 Aug 2012 C L Fancourt and J G Aceti System and Method for Detecting Deviations in Nominal Gait Patterns A Salarian C Zampieri F B Horak P Carlson Kuhta J G Nutt and K Aminian Analyzing 180 turns using an inertial system reveals early signs of progression of parkinson s disease in Annual International Conference of the IEEE Engineering in Medicine and Biology Society 2009 EMBC 2009 20089 pp 224 227 J Barth J Klucken P Kugler T Kammerer R Steidl J Winkler J Hornegger and B Eskofier Biometric and mobile gait analysis for early diagnosis and therapy monitoring in Parkinson s disease in 2011 Annual International Conference of the IEEE Engineering in Medicine and Biology Society EMBC 2011 pp 868 871 G Rigas A T Tzallas D G Tsalikakis S Konitsiotis and D Fotiadis Real time quantification of resting tremor in the Parkinson s disease in Annual International Conference of the IEEE Engineering in Medicine and Biology Society 2009 EMBC 2009 2009 pp 1306 1309 G Rigas A T Tzallas M G Tsipouras P Bougia E E Tripoliti D Baga D Fotiadis S G Tsouli and S Konitsiotis Assessment of tremor activity in the Parkinson s disease using a set of wearable sensors IEEE Trans Inf Technol Biomed vol 16 no 3 pp 478 487 May 2012 S Patel K Lorincz R Hughes N Huggins J Growdon D Stan
92. s of Bradykinesia signs of slow motion may not 19 appear or less prominent But individual body movements become gradually slower with aging in the disease and muscles start to freeze Detail on Appendix D Rigidity Rigidity is muscle stiffness muscle inflexibility or movement inflexibility muscle tiredness or muscles ache In fact Rigidity is the change of tone in muscle movement or a sense of resistance to limb motion Rigidity may occur in shoulders hip neck and can reduce range of motion All muscles have opposing muscles and during rigidity these opposing muscles relax to pose hindrance to the movements Even some patients may not find signs of rigidity in the early stages but can be detected on physical examination On later stages it becomes uncontrollable and more reoccurring Details on Appendix D Postural Instability Other name is the loss of body balance The way of standing is called posture It is losing of body balance leading to fall It is the 4 most common symptom in PD and can be examined by a test called pull test for watching the behavior when a patient is pulled backward by the shoulders PI is the most common symptom of fall The patients with PD often fall left right or backward during standing or turning These patients behave PI even they are on medication have undergone some surgery exercising or therapies It is reduced or nominal loss of maintaining body equilibrium that may become se
93. study somehow Internal Validity Internal validity includes instrumentation amp maturization which could have effect on casual link of outcome and treatment It includes pilot experimentation for verifying WAIST position 4GB test Flow Charts of algorithm and Gait tables 1 amp 2 A false positive alarm could have effect on internal validity but if threshold is calculated and set accurately these false alarms can be avoided External Validity External validity is related to generalization This proposed algorithm can be fairly used for other patients with PD However for seizure detection or gait analysis for some other neurological disorders it may need test retest reliability Construct Validity Our 4GB test provides valid basis for motion data after algorithm implementation threshold is tested to set between 200 300 30 in offline and real time processing Conclusion validity We have a valid 4GB test protocol to take measurements We concluded that threshold value falls between 200 300 every time whenever algorithm is executed Reliability Reliability means consistency in data collection and data analysis We observed a bit consistent results in setting thresholds between 200 300 in offline and real time environments and it is proved in the test retest process 63 6 4 Conclusions We have presented a real time gait analysis algorithm which is capable of detecting the motion of the patient with PD and to obje
94. supervised conditions In the first experiment one accelerometer is used while two accelerometers are used on either sides of the waist in the second experiment Right vertically positioned accelerometer is used to capture motion data for the EyesWeb patch while left horizontal accelerometer is used to capture motion through Shimmer Connect default software We also tested accelerometers putting on different body positions such as limbs chest and waist Results from pilot testing encourage putting accelerometer on waist in order to track motion more accurately and actively When putting wireless sensors on chest amp upper lower limbs unnecessary motions are captured which is not required Another reason to put accelerometers on both waist positions is that it gives efficient information of the patient for falling 4 Every patient with PD may be different in symptoms from another but falling can have bit similar patterns SHIMMER tri axial accelerometers with Sampling Rate 100Hz 1 5g sensitivity and burn with BoilerPlate_Shimmer2r in BSL430 ShimmerConnect_V 2_Win software is used to capture the motion data We are interested to capture the higher acceleration above threshold value for fall risk assessment To this contrast FoG f 4 210Hz or even below 30 fmax Maximum frequency which is maximum standard deviation in a particular acceleration in xyz axis 6 1 3 Protocol We conducted two series of experiments 2
95. t from the researchers in the field we contacted different individuals and got following comments feedbacks Barry Greene PhD Applied Technology amp Design Intel labs Greene says that I think you are developing a potentially useful system While it may be beyond the scope of your project I would suggest testing your system on a target population e g Parkinson s patients older adults at risk of fall etc and using these data to further refine your methods Fei Yu Alsion 2 6400 S nderborg Denmark Fei Yu suggested using Fourier Transform besides using STD to compare results He also suggests using ROC and Classification Methods to evaluate result accuracy Wouter Speybrouck Technology Consultant Hewlett Packard Belgium Diegem Machelen Brussels Belgium Wouter says that observation you have made about the steps with energy is in my opinion very true As long as the patient knows that you are testing his gait he will consciously or unconsciously pay attention to his steps I guess this is worth mentioning in the report and in your findings He also suggests taking 50 100 samples second to avoid false positive alarms Arash Salarian Department of Neurology Oregon Health amp Science University Portland Arash suggests that working directly on raw acceleration may give less accurate results So there is a need to calibrate the signal before any processing because as variation between devices
96. ta detailed exploratory data analysis regression analysis using linear and non linear models and providing custom equations Detail description about algorithm flow charts and code is available in the Appendix at the end of this report 6 2 Gait Assessment A wearable gait measurement system can be 27 successfully developed to perform gait analysis in neurological disorders specially fall risk estimation for the elderly population Different measurements with standard deviations are calculated using Matlab tool as shown in the Table 6 1 Standard deviation is defined as the statistic used to measure how much variance from mean exists The criterion based on calculating standard deviation seems good as different researchers 5 29 10 have employed it in their experimentation Some other calculations standards rating scales measurement tools and parameters are also used in the gait analysis procedure So spatiotemporal gait features are calibrated for offline processing to calculate and set threshold value which can be used for real time processing through algorithm implementation In the following table step count is used as gait parameter to count the number of steps per seconds It can also be called step frequency All walks are 4 meter Go amp Back 4GB except circle gaits Experiments A B are conducted on same the day with half hour pause 6 2 1 Step Count Sr Experiment 01 Experiment A Experiment 02 Gait Name N
97. ter amp Back Test 4 46 3 2 EXISTING GAIT ANALYSIS METHODS cessent enne ene en tenter entente entente ene 46 5 3 WIRELESS MEASUREMENTS eit eroe thigh gece ter ee ves PEE E du PHYS EE 47 2 941 WIEISEBEAS cec er e e d e eR IE US 47 5 3 2 Sensors for Medicine and Science 47 23 9 9 ZZipBeecs assai ERE Ie XS ERE ERREUR FAVERE E 48 213 47 a 91 E A A S a i EES EEE a HE E Ee Y 46 5 3 5 WIN Human Recorder Co seen eet nest enter nennen 48 5 4 WIRED ge Hu 48 5 4 1 Nanol7mm 6 Force Torque 49 DADs gt Digitising Tablet e o EH ERR EE 49 25 OUR GAIT ANALYSIS MECHANISM sccesccessseceseeecsseceeeeecaeceeneecaaeceeeecsaeceneecsaeceeeeeeaeeeee 49 5 6 FINDINGS D 50 6 IMPLEMENTING ALGORITHM amp RESULTS 0 ccssscssssccsssssssscsssscssscsssscssssssssecssssees 51 6 1 EXPERIMENT iore re EVO REIS E bee ER SEES ORE ERES 51 6 1 1 Patient History medical and fall history eese 51 NUS AREE 52 6 1 3 uoo E 52 6 1 4 Acquisition sse tae tete epe ee sae e Tec eee EORR eee Rd 54 6 2 GATT ASSESSMENT niin i
98. ts can also be included for comparison called control subjects In our experiment we encourage subject to walk 4 meters go and return fashion We also observed that our subject poses shakier gait when holding glasses in both hands and she leans left or right during turning 2 6 7 Universal Gait Parameters Include body velocity time ratio distance range session or repetition of gait if necessary and video recordings We recorded all gait session with a high quality mobile for validation 2 6 8 Features Extraction Interested features from different gaits are extracted for further pre or post processing for analysis For example in our experiments we are more interested to extract the high acceleration features during the walk and capture more acceleration when the patient behaves above threshold value such as during turning patient leans towards right or left Analysis shows that patient with PD observes significantly longer turning duration 7 Various methods and techniques are used for feature extraction Some of them are Principle Component Analysis high or low pass band filters pattern recognition in Machine Learning and others Here we are giving some details about the purpose of using such techniques 2 6 9 Techniques Used for Feature Extraction Time Frequency Analysis technique is used for FOG detection can be difficult to detect when the patient performs feet sliding gait 29 Other techniques and methods are e MiM
99. urological Disorder Any malfunctioning of neurons in the nervous system is called a Neurological disorder Over 100 neurological disorders have been discovered throughout the world In our study we have chosen one disorder Parkinson s disease for falling Etiology of many disorders is still unknown although a lot of research has been performed Parkinson s disease is a neurodegenerative disorder in the central nervous system Three most common Parkinson symptoms are resting Tremor Slowness of Movements and Rigidity A Tremor is one of the 1 most common types of Parkinson s disorder which can be monitored using inertial sensors Epilepsy is an electrical storm in the brain producing set of seizures So only one seizure is not an epileptic seizure rather it is combined effect of many seizures Almost 1 4th of epileptic disorders cannot be treated using available therapies 2 The modern advanced world is suffering more with Parkinson s disease as compared to developing countries 1 3 Gait Parameters Experiments can be performed on different gait parameters like body velocity time ratio ground slope stance swing body gestures and gait patterns Sensors can be put on hips knees thighs limbs neck head chest or any other suitable body part to capture motion data for further pre and post processing 1 4 Pre and Post Processing Preprocessing is real time gait analysis through time and frequency domain to trigger various security ste
100. us security steps and messages for patient care Post processing is offline analysis of motion data in different tools such as EyesWeb BioMOBIUS and Matlab for calculations analysis and plotting of motion to take decisions to formulate a mechanism for patient activity detection and monitoring The area which we choose is pretty interesting pertaining to rehabilitation wellness and healthcare for older people Other related keywords may include keywords may be helpful using one or combination of more than one WSN BAN or WBAN biosensors neurological disorders gait analysis fall detection fall avoidance Parkinson s disease wireless accelerometer ambulatory monitoring freezing of gait and fall risk assessment Most of the patients suffering from any neurological disorder in later stages of disease pose ambulatory disturbance especially falling Such patients may fall without showing any warning sign and every patient is different from another So there is a need to develop a mechanism to detect shaky motion to avoid such patients from falling Therefore a real time gait analysis algorithm is implemented to trigger security alarms In order to assess amp evaluate gait analysis accurate reliable amp consistent measurement tools need to be utilized Even slight deviation in the data monitoring through measurement tools is not encouraged to use 21 Gait disturbance can be measured using 3 axis accelerometers like SHIMMER R for real time
101. vention You need to simplify your research aims As well look at such things as test retest reliability studies to see if the algorithms work on the same subject when the device is attached in somewhat different positions and the likelihood of false positives from activities of daily living if used for falls Alan DeLaTorre Project Manager Age Friendly Portland Institute on Aging School of Community Health Portland State University Aland says your work with accelerometers and movement disorders is very interesting and I would like to be of assistance Please let me know if there are specific questions that I may be able to assist with 79 I must admit that many of these questions are tough to answer as I am not an expert in the area that you are focusing on nor do I understand the development of algorithms Nonetheless included responses to several of your questions here 1 How can accuracy and specificity of algorithm be validated as we have tested on one female with PD Regarding validity I would assume that it is important to try and control for certain human variable For instance are you able to test with a person who has PD and someone who does not have PD If you could control for physical characteristics and or functional ability that would be helpful Look to attachments Lawton and ADL for variables that you may be able to hold constant Also are there specific PD theories of previous res
102. vere at later stage Festination This word has Latin background meaning Too Hurry Sudden rapid shuffling of small steps or hurrying in walking is called Festination Opposite to FoG a motor uncontrollable symptom Festination may lead to fall Festination is caused due to hyper tonicity of the muscles So in this way it is more dangerous as compared to PI and FoG Festination can be observed in gait speech and the handwriting etc Some patients feel disturbance during speaking called Oral Festination Festination Gait is clearly associated with longer duration of PD symptoms not with disease severity as reflected in the motor part of the UPDRS So a change of gait and 12 physical therapy may help in controlling the conditions of Festination Other name for Festination is Parkinsonian Gait Freezing of Gait FoG 20 Walking inactivity is called freezing of the gait FoG is one of the most disturbing symptoms of Parkinson s disease Freezing can also occur during speaking arm swing or head nodding Among these freezing of legs is most distressing and annoying which may result in falling and injury It is an episodic inability of taking step It is 13 one of the most disturbing problems in which patients observe feet get glued 14 frozen or stuck to the ground stoop forward and likely to fall Nearly 30 within 5 years and 60 within 10 years experience FoG more often at later stages It may happen not only during wal
103. vodopa has a low effect on her disease But after DBS surgical procedure patient still experiences postural instability resulting in falls showing no sign specially when medication is not taken in time Studies show that such patients after surgery may have reduced cognition In an interview she tells that although falls rates are rare before surgery but after surgical procedure the fall rate increases significantly varies day to day When asked replied that she falls early on the day before taking measurements in both experiments Her fall history shows that she often falls backward pertaining to Postural Instability Early in the day of measurement she observes fall at right Other constant parameters are Subject 70 year General physique healthy Gender female Support walking sticks at both hands Surgery history recently undergone surgery DBS Step related parameters reduced gait small steps Medicated yes PD Symptoms no or minimum tremor 51 Assumed symptoms Non alcoholic no backbone problem no addicted MFE this data was gathered through Modified Fall Efficacy MFE scale It is a 14 questions assessment sheet to assess the patient s tendency to fall during daily life activities while 4GB test designed by us may be helpful in taking measurements with gait parameters 6 1 2 Method Wearable waist mounted tri axial accelerometers are used to measure movement of human motion during walking under
104. wavplay siren2 siren fs break else display Alarm OFF end end plot std arr 1 hold on plot std arr 2 r hold on plot std_arr 3 g 75 Appendix G Code Real Time Gait Analysis 559595555555555 Start 55 55 55 cle handles NO SAMPLES IN PLOT 500 Number of samples that will be displayed in the plot at any one time handles DELAY PERIOD 0 25 A delay period of time in seconds between data read operations handles comPort 3 handles captureDuration Inf handles Thresholdl 200 handles Threshold2 400 handles fileName ShimmerData dat handles active 0 set handles pushbutton1 String Start set handles sliderl Value handles Thresholdl set handles slider2 Value handles Threshold2 handles shimmer ShimmerHandleClass handles comPort Define shimmer asa ShimmerHandle Class instance with comPort if handles shimmer connect TRUE if the shimmer connects Define settings for shimmer handles shimmer setsamplingrate 100 Set the shimmer sampling rate to 51 212 handles shimmer setinternalboard None Set the shimmer internal daughter board to None handles shimmer setenabledsensors Accel 1 5 Enable the shimmer 1 accelerometer handles shimmer setaccelrange 0 Set the accelerometer range to 0 f X258 handles iAccelXShimmer handles shimmer getsignalindex Accelerometer X Determine the column index o
105. wo different types of proteins need to be discouraged because it may have negative effects on medicines If the patient feels swallowing impairment dysphagia changes in eatable are recommended Levodopa is a type of protein so two types of proteins should not be taken in one meal as it reduces the effect of the drug So Levodopa is recommended to take 30 minutes before meal 2 4 5 Lifestyle Modification Rehabilitation also helps to decrease the progression of the disease Proper education and awareness provided by different private and public bodies can help to fight with the disease in an effective way Regular daily exercise to increase flexibility and moveability can be beneficial Rehabilitation through life style modification can enhance the lives of people with Parkinson s disease Special therapies can also contribute to gain a goal of rehabilitation Such as physical therapy and occupational therapy can guide people to behave better in daily life For independent living certain necessary initiatives and precautionary measures can be adopted such as e How to sit stand on the chair bending circling e How to get out of bed more easily e Howto grip things books glass stationary mobile etc e How to be conscious during turning e What stuff is needed to be removed in the way What and how to eat drink dress using special utensils such as spoon or fork Patients with PD are more prone towards falling in the wash rooms S
106. y Alarms when value from motion data exceeds maximum threshold value 1 12 Methodology Our research methodology is experimental in connection with literature review process RQ 1 is based on finding and exploring the related base papers and pilot experimentations Research question 2 is totally experimental starting from designing protocol in the form of 4GB Test flowcharts of Algorithm offline and real time Gait analysis tables 1 amp 2 Then finally implementation is done for real time in MATLAB Some detail about the flow of research process is given by the figure 1 below Literature Problem Review Statement Pilot Experiment Experimental Design Implementation offline pre Figure 1 Research Methodology 12 1 13 Results As shown in the gait assessment tables 6 2 6 3 below results show that the proposed system is fairly simple to use in real time situations flexible to adjust to any necessary change in the future In GUI environment the concept of slider seems fairly suitable to control min max thresholds for warning and emergency alarms plotting real time motion and saving file with name Threshold can be set between 200 300 in test retest process 1 14 Technical Description Technically wearable wireless sensors provided by the SHIMMER can be put on the different body parts to remotely track the information about the movement of the patient with PD This information is captured and saved into the

Download Pdf Manuals

Related Search

Related Contents

Bestron Puro  Phonix BEATLEW mobile headset  MS 4400 RS  Armado Carretera (96)  This device complies with part 15 of the FCC rules  GUIDE D`UTILISATION DU SITE DE L`AECSE (Association des  iogear GCS1004 KVM switch  Oregon Scientific BA900 User's Manual  

Copyright © All rights reserved. Failed to retrieve file