Human Motion Capture and Identification for Assistive Systems Design in Rehabilitation

ISBN-13: 9781119515074 / Angielski / Twarda / 2021 / 240 str.

Pubudu N. Pathirana

Spis treści

1 Introduction 91.1 Human Body - Kinematic Perspective 101.2 Musculoskeletal Injuries and Neurological Movement Disorders 111.2.1 Musculoskeletal injuries 111.2.2 Neuromascular disorders 121.3 Sensors in Tele-rehabilitation 151.3.1 Opto-electroic Sensing 151.3.2 RGB Camera and Microphone 201.3.3 Inertial Measurement Unit (IMU) 221.4 Model based state estimation and sensor fusion 241.4.1 Summary and challenges 241.5 Human Motion Encoding in Tele-rehabilitation 251.5.1 Human motion encoders in action recognition 251.5.2 Human motion encoder in physical tele-rehabilitation 261.5.3 Summary and challenge 271.6 Patients' Performance Evaluation 281.6.1 Questionnaire based assessment scales 281.6.2 Automated kinematic performance assessment 291.6.3 Summary and challenge 302 Kinematic Performance Evaluation with Non-wearable Sensors 312.1 Introduction 322.2 Fusion 322.2.1 Introduction 322.2.2 Linear Model of Human Motion Multi-Kinect System 342.2.3 Model-based state estimation 372.2.4 Fusion of Information 372.2.5 Mitigation of occlusions and optimised positioning 372.2.6 Computer simulations and hardware implementation 382.3 Encoder 492.3.1 Introduction 492.3.2 The two-component encoder theory 512.3.3 Encoding Methods 522.3.4 Dealing with Noise 542.3.5 Complex Motion Decomposition using Switching Continuous Hidden Markov Models 562.3.6 Canonical Actions and the Action Alphabet 572.3.7 Experiments and Results 592.4 ADL Kinematic Performance Evaluation 672.4.1 Introduction 672.4.2 Methodology 682.4.3 Experiment Setup 712.4.4 Data Analysis and Results 752.5 Summary 833 BioKinematic Measurement with Wearable Sensors 853.1 INTRODUCTION 863.2 Kinematic Model 863.3 Introduction to Quaternions 863.4 Wahba's Problem 873.4.1 Solutions to the Whaba's Problem 883.4.2 Davenport's q method 893.4.3 Quaternion Estimation Algorithm(QUEST) 913.4.4 Fast Optimal Attiude Matrix (FOAM) 923.4.5 Estimator of the optimal quarternion (ESOQ or ESOQ1) method 923.5 Quaternion Propagation 923.6 MARG (Magnetic Angular Rates and Gravity) Sensor arrays based Algorithm 933.7 Model based estimation of attitude with IMU data 933.8 Robust optimisation based approach for Orientation estimation 963.9 Implementation of the Orientation estimation 973.9.1 Extended Kalman Filter based approach 973.9.2 Robust Extended Kalman Filter Implementation 983.9.3 Robust Extended Kalman Filter with Linear Measurements 993.10 Computer Simulations 993.11 Experimental Setup 1003.12 Results and Discussion 1023.12.1 Computer Simulations 1023.12.2 Experiment 1043.13 Conclusion 1054 Capturing Finger movements 1074.1 Introduction 1084.2 System Overview 1114.3 Accuracy Improvement of Total Active Movement and Proximal Interphalangeal Joint Angles 1114.4 Simulation 1154.5 Trial Procedure 1164.6 Results 1174.6.1 Concurrence validity 1174.6.2 Internal Reliability 1194.6.3 Time efficiency 1194.7 Discussions 1204.8 Approaching finger movement with a new perspective 1224.9 Reachable Space 1244.10 Boundary of the Reachable Space 1274.11 Area of Reachable Space 1314.12 Experiments 1334.13 Results and Discussion 1344.14 Conclusion & Future Work 1405 Non-contact measurement of respiratory function via Doppler Radar 1415.1 Introduction 1425.2 Fundamental Operation of Microwave Doppler Radar 1445.2.1 Velocity and frequency 1445.2.2 Correction of I/Q Amplitude and Phase Imbalance 1465.3 Signal Processing Approach 1495.3.1 Respiration rate 1495.3.2 Extracting respiratory signatures 1515.3.3 Low Pass Filtering (LPF) 1535.3.4 Discrete Wavelet Transform 1545.4 Common Data Acquisitions Setup 1555.5 Capturing the dynamics of respiration 1575.5.1 Normal Breathing 1585.5.2 Fast Breathing 1585.5.3 Slow Inhalation - Fast Exhalation 1585.5.4 Fast Inhalation - Slow Exhalation 1585.5.5 Capturing Abnormal Breathing Patterns 1605.5.6 Breathing Component Decomposition, Analysis and Classification 1615.6 Capturing Special Breathing Patterns 1645.6.1 Correlation of Radar Signal with Spirometer in Tidal Volume Estimations 1655.6.2 Experiment Setup 1665.6.3 Results 1665.6.4 Motion Signature from Doppler radar 1755.6.5 Measurement of Volume In (Inhalation) and Volume Out (Exhalation) 1765.7 Removal of Motion Artefacts from Doppler Radar based Respiratory Measurements1795.7.1 Experimental verification 1815.7.2 Results and Discussion 1825.7.3 summary 1855.8 Separation of Doppler Radar based Respiratory Signatures 1865.8.1 Respiration Sensing Using Doppler Radar 1865.8.2 Signal Processing -Source Separation (ICA) 1875.8.3 Experiment Protocol for Real Data Sensing 1895.8.4 Two Simulated Respiratory Sources 1905.8.5 Experiment involving real subjects 1905.8.6 Separation of Hand Motion 1935.8.7 Conclusion 1956 Appendix I 2036.1 Static Estimators 2046.1.1 Least Squares Estimation 2046.1.2 Maximum likelihood estimation 2046.2 Model based estimators 2056.2.1 Kalman filter (KF) 2056.3 Particle filter 2076.3.1 Robust filtering with linear measurements 2086.3.2 Constrained optimisation 210

O autorze

Pubudu N. Pathirana, PhD is a Senior Member of the Institute of Electrical and Electronics Engineers (IEEE) and received his PhD in Electrical Engineering from the University of Western Australia. He was a Postdoctoral Research Fellow at Oxford University, Oxford, England; a Research Fellow at the School of Electrical Engineering and Telecommunications at the University of New South Wales, Sydney, Australia; and a consultant to the Defense Science and Technology Organization (DSTO). Currently, he is a Professor and the Director of the Networked Sensing and Control group at the School of Engineering, Deakin University, Geelong, Australia. His current research interests include bio-medical assistive device design, human motion capture, mobile/wireless networks, rehabilitation robotics, and radar array signal processing.Saiyi Li, PhD received his PhD from Deakin University in 2016 through a sponsorship by the NICTA (National Information Communications Technology Australia, now Data61). His research interests include rehabilitation engineering, signal processing, and machine learning.Yee Siong Lee, PhD received his PhD from the Deakin University in 2016 through a sponsorship by NICTA. His research interests include biomedical applications and signals processing, machine learning, operations research, sensors networks, and radar signal processing.Trieu Pham, PhD received his PhD through a sponsorship by NICTA from Deakin University, Victoria, Australia in 2017. His research interests include machine learning, signal processing, computer aided rehabilitation, and kinematics of human motion.