Development of a general purpose computer-based platform to provide functional assistance to people with severe motor disabilities Franco Senatore A dissertation submitted to the Faculty of Engineering and the Built Environment at The University of the Witwatersrand, Johannesburg, in fulfilment of the requirements for the degree of Master of Science in Engineering. Johannesburg, March 2009 Declaration I declare that this dissertation is my own, unaided work, except where otherwise acknowl- edged. It is being submitted for the degree of Master of Science in Engineering in the University of the Witwatersrand, Johannesburg. It has not been submitted before for any degree or examination in any other university. Signed this day of 20 Franco Senatore. i Abstract Research and development into a generic assistive platform, which can accommodate a vari- ety of patients suffering from a wide range of motor disabilities is described. Methodologies were established, whereby the design could be made sufficiently flexible, such that it could be programmed to suit these people in terms of their needs and level of motor disability. This needed to be achieved without redesigning the system for each person. Suitable sensors were chosen to sense the residual motor function of the disabled individual, while being non-invasive and safe for use. These sensors included a dual-axis accelerometer (tilt switch), a 6-key touch sensor and a SCATIR switch (blink/wink sensor). The placement of the sensors, for the purpose of this study, were restricted to sensing arm (dual-axis ac- celerometer) or finger movements (touch sensors), head and neck movements (accelerom- eter) and blink/wink and/or eye-brow movements (SCATIR switch). These input devices were used to control a variety of different output functions, as required by the user, while being non-invasive and safe for use. After ethics approval was obtained, volunteers with various motor disabilities were subse- quently invited to test the system and thereafter requested to answer a series of questions regarding the performance and potential usefulness of the system. The input sensors were found to be comfortable and easy to use, while performing predictably and with very little to no fatigue experienced. The system performed as expected and accepted all of the input sensors attached to it, while repeating specific tasks multiple times. It was also established that the system was customisable in terms of providing a specific output for a specific and voluntary input. The system could be improved by further compacting and simplifying the design and operation, while using wireless sensors were necessary. It was thereafter concluded that the system, in general, was capable of satisfying the various users’ diverse requirements, thereby achieving the required objectives. ii Acknowledgements The following research was performed under the auspices of the Biomedical Engineering Department at the University of the Witwatersrand, Johannesburg, South Africa. I would like to thank my supervisors, Prof. David Rubin and Prof. George Gibbon, for their continued support, guidance and patience throughout the duration of this research. I would also like to thank all of the volunteers who came forward to assist with the performance and testing phase of the research and whose names are not mentioned here to maintain confiden- tiality, as well as Mr Vincent Gore who put me into contact with most of these volunteers. A special mention must be made to Mrs Celeste Mukheibir of Inclusive Solutions for kindly lending me one of the input sensors at no cost, as well as Brittan Healthcare for providing the safety testing at no cost, it is greatly appreciated. I would also like to extend a special mention to my colleagues at the Department of Elec- trical and Information Engineering A-Lab, namely Megan Russell for kindly assisting me with the testing phase of the research as well as Rolan Christian, Ryan Van Den Bergh and David Vannucci for helping with various aspects of the project. A thanks goes out to Mr Tom Lattimer, Mr David Tshabalala, Mr John Manchidi and Mr Sakhi Xaba of the Elec- tronics Workshop for their assistance with regards to technical help and support. Finally, I would like to thank the Human Research Ethics Committee (HREC) of the University of the Witwatersrand for reviewing the protocol and providing feedback on the study. iii Contents Declaration i Abstract ii Acknowledgements iii Contents iv List of Figures xiii List of Tables xvi 1 Introduction 1 1.1 Research Objectives . 1 1.2 Approach . 2 1.3 Guide to the Report . 2 2 Background Information 4 2.1 Motor Disabilities . 4 2.1.1 Spinal Cord Injury . 4 2.1.2 Classification of the Level of Lesion . 7 2.2 Motor Control . 10 iv 2.2.1 Open versus Closed Movements . 11 2.2.2 Controlling Externally Powered Devices . 11 2.3 Functional Electrical Stimulation (FES) . 14 2.4 Existing Technologies . 14 2.4.1 Single-switch Access . 14 2.4.2 Sip-and-puff switch . 15 2.4.3 Oversized trackball mouse . 15 2.4.4 Adaptive Keyboard . 15 2.4.5 Eye Tracking . 16 2.4.6 Voice Recognition . 16 3 Subject Classification 17 3.1 Types of Motor Disabilities . 17 3.1.1 Level of Lesion Classification . 17 3.2 Body Movement Control . 18 3.3 Motor Response Classification . 18 4 Design Rationale 19 4.1 Design Objectives . 19 4.2 Basic Considerations in Motor Behaviour and Control Systems . 19 4.2.1 Reliability . 20 4.2.2 Objectivity . 20 4.2.3 Validity . 20 4.2.4 Minimising Energy Required . 20 v 4.2.5 Minimising Disruption . 20 4.2.6 Safety . 21 4.2.7 Cosmetics . 21 4.2.8 Practicality . 21 4.3 General Concerns . 21 4.3.1 The Signal to Noise Ratio (SNR) . 21 4.3.2 Ambient (Transmission Line) Noise . 22 4.3.3 Signal Distortion . 22 4.3.4 Inherent Instability of the Signal . 23 4.4 Real-time Systems . 23 4.5 Input Sensor Selection . 23 4.6 Instrumentation Requirements . 24 5 Design and Implementation 25 5.1 Stage 1 - Power Supply and Conditioning . 27 5.1.1 Input Hardware Power Supply . 27 5.1.2 Data Acquisition Card and PC Power Supply . 27 5.2 Stage 2 - Input Stage (Input Sensor Design) . 27 5.2.1 The ADXL250 Dual-axis Accelerometer . 28 5.2.2 Six-Channel Touch Sensor . 29 5.2.3 The Self-Calibrating Auditory Tone Infrared (SCATIR) Switch . 30 5.3 Stage 3 - Signal Selection and Signal Conditioning . 30 5.3.1 Signal Selection . 31 vi 5.3.2 Signal Conditioning . 32 5.4 Stage 4 - Data Acquisition (DAQ) and Control . 35 5.4.1 Prior data acquisition research . 35 5.4.2 NITM PCI-6221 DAQ Card . 36 5.4.3 Analogue-to-Digital Conversion . 37 5.5 Stage 5 - Software Design and Implementation . 38 5.5.1 Matlab R ............................... 38 5.5.2 Object-Oriented Design using MSTM Visual Studio C#.NET . 38 5.5.3 National InstrumentsTM Measurement Studio and DAQmx 8.0 . 39 5.5.4 Requirements Engineering . 39 5.5.5 Functionality Diagram . 39 5.5.6 Software Implementation . 41 5.5.7 Software Testing and Validation . 45 5.6 Stage 6 - Output Stage . 47 5.6.1 Hardware/Physical Outputs . 48 5.6.2 Software Outputs . 48 5.7 Stage 7 - Design Implementation and Functionality Testing . 49 5.7.1 Implementation . 49 5.7.2 Operation and Functionality Testing . 49 6 Testing and Results 51 6.1 Ethics Approval . 51 6.2 Safety Precautions and Safety Testing . 52 vii 6.2.1 ±12V Battery . 52 6.2.2 The NITM PCI-6221 DAQ card . 52 6.2.3 The Isolation Transformer . 53 6.2.4 Overall Safety Testing . 53 6.3 Risks . 54 6.4 The Testing Procedure . 54 6.4.1 The Test Subjects . 54 6.4.2 The Input Sensors . 54 6.4.3 System Setup . 55 6.4.4 The Output Functions . 55 6.4.5 The Questionnaire . 56 6.5 Results . 56 6.5.1 The Input Sensors . 56 6.5.2 The Overall System . 60 6.5.3 User Feedback: General Comments and Possible Improvements to the System . 62 6.6 Analysis of Results . 63 6.6.1 Input Sensor Characteristics . 63 6.6.2 The Overall System Characteristics . 65 6.7 Suggested System and Design Improvements . 66 7 Conclusion and Recommendations 68 7.1 Recommendations for Further Research . 68 7.2 Conclusion . 69 viii A Classification of Spinal Cord Injuries 71 B Supplementary Prior Design Information 73 B.1 Introduction . 73 B.2 Generic Assistive System Principles and Components . 73 B.2.1 Input Sensors . 73 B.3 Filtering . 77 B.3.1 Analogue Filtering . 77 B.3.2 Digital Filtering . 79 B.4 Spectral Analysis . 79 B.4.1 Background Information . 79 B.4.2 Fast Fourier Transform (FFT) . ..