Aim: 

1. Develop a 9-DOF Underactuated Upper limb exoskeleton (UULE) for chronic stroke patients with impaired distal joints

2. Provide reliable sensing feedback for VR with visual-inertial sensors and encoders

3. Develop control frameworks for underactuated exoskeletons in bimanual ADL training 

Technical Problem: 

1. Existing exoskeletons suffered design limitations like heavyweight and no independent joint assistance for specific impaired joints

2. VR reconstructs the daily scenarios for ADL training, but the sensing errors limit the effectiveness 

3. Impaired joints assisted by active joints may not coordinate with passive joints freely mobilized by patients

Novel Approach:

1. Lightweight cable-driven joints assist impaired joints with joint impedance control, while passive joints conforming to less-impaired joints

2. A sensing system consisting of visual-inertia sensors and rotary encoders is used in kinematics measurement and VR

3. Independent joint assistance with visual guidance can assist patients in ADL and self-coordinate their unassisted joints

Experimental Result: 

1. Joint assistance was verified by 15 healthy subjects with reductions in muscle activation and motion error

2. With our sensing system, 8 healthy subjects have similar joint motion and muscle activation in VR and the real world, so patients can train in VR daily scenes

3. Robot assistance can be observed by the reduction of EMG and joint angle difference in 15 subjects 

Mechanical design with 3D CAD model, and 2D diagram for fabrication with GD&T and China GB Standard

Robot dynamic simulation with MATLAB

Electronic components selection and PC104 (Linux) implementation, including actuators, encoders, inertial measurement units (IMUs), and camera

Robot controller implementation, including PID, impedance control, and robot weight compensation

Development of an interactive virtual reality (VR) environment for human-robot interaction in Unity3D

C/C++/ C# programming skills for PC104 (Linux), Teensy, and Unity3D

Experiment planning to verify the project’s aim

Data collection, processing, and analysis, including robot sensing signal and human biosignal (Electromyography (EMG), and kinematic data (Vicon system))

Thomas M. Kwok, “Development Of An Underactuated Upper Limb Exoskeleton For Post-Stroke Bimanual Activities-Of-Daily-Living Training,” Ph.D. dissertation, National University of Singapore, 2023

Thomas M. Kwok, Haoyong Yu, “A Novel Bilateral Underactuated Upper Limb Exoskeleton for Post-stroke Bimanual ADL Training,” IEEE Neural Systems & Rehabilitation Engineering (TNSRE), 2024.

Thomas M. Kwok, Haoyong Yu, “Asymmetric Bimanual ADL Training with Underactuated Exoskeleton using Independent Joint Control and Visual Guidance,” IEEE Access, 2024.

Thomas M. Kwok, Tong Li, Haoyong Yu, “A Reliable Kinematic Measurement of Upper Limb Exoskeleton for VR Therapy with Visual-inertial Sensors,” in 2023 IEEE/ASME International Conference on Advanced Intelligent Mechatronics (AIM), 28-30 June 2023, pp. 584-590.

Thomas M. Kwok, Haoyong Yu, “Apparatus for Assisting Movement of Upper Limb Of User,” International Patent Application (WIPO) No. WO2022216230A1, 2022 

Thomas M. Kwok, Haoyong Yu, “Device for Assisting the Movement of the Upper Limbs of a User (用于辅助用户的上肢运动的装置),” Chinese Patent Publication No. CN 114246761 A, 2022

Thomas M. Kwok, Haoyong Yu, “Apparatus for Assisting Movement of Upper Limb Of User,” SG Non-Provisional Application No. 10202103465U, 2021