Pattern recognition-based myoelectric control of upper limb prostheses has been made clinically available to individuals with more proximal upper limb amputations and can restore intuitive control of a prosthetic hand. This control method has yet to be implemented for individuals with amputations distal to the wrist (i.e. partial-hand amputations) and who constitute over 90% of all upper limb amputees. Unique to this population of amputees is the presence of a functional wrist. The purpose of this work was to evaluate strategies that would facilitate the novel use of pattern recognition-based myoelectric control of electrically powered, partial hand prostheses and allow partial-hand amputees to maintain wrist function while effectively controlling the prosthesis. My initial offline studies showed that a functional wrist significantly diminishes the performance pattern recognition algorithms (p < 0.001). To resolve this challenge, this dissertation evaluated (1) different paradigms to collect decoding algorithm training data, (2) contributions of discriminatory information from intrinsic and extrinsic EMG muscles sources, (3) classification algorithms, (4) time and frequency domain EMG features and (5) the use of wrist kinematics to improve the decoding of hand grasps in different wrist positions in non-amputee and amputee subjects. Finally, real time control tests with four partial-hand amputees using the Touch Bionics i-limb digits were performed and demonstrated that pattern recognition myoelectric control allows for the control of electrically-powered partial hand prostheses while allowing the user to retain residual wrist function.

Last modified

• 02/13/2018

Creator

• Adenike Adewuyi

DOI

• https://doi.org/10.21985/N26673

Subject

• Biomedical Engineering

Keyword

• Pattern Recognition
• Partial hand amputee
• Intrinsic hand muscle
• myoelectric control

Date created

• 2016-01-01

Resource type

• Dissertation

Rights statement

• In Copyright