Date of Award
11-28-2016
Document Type
Thesis
Degree Name
Master of Science (MS)
Department
Engineering Technology
First Advisor
Hamdi Allbunashee
Abstract
A myoelectric-controlled prosthesis holds the promise of renewed quality of life for the amputee population. When it comes to prosthesis control, muscle fatigue presents a significant challenge for all task discrimination algorithms due to changed muscle activation patterns and significant shifts in the spectral contents of the surface electromyogram (sEMG) signals. The effect of muscle fatigue on the classification accuracy of two supervised machine-learning algorithms, i.e. the linear discriminant analysis (LDA), and the muscle synergy based task discrimination (MSD) was investigated. Due to muscle fatigue, the classification accuracies declined from (>90%) to 41% and 59% for the LDA and the MSD algorithm respectively. Two real-time adaptive classification techniques were proposed to enhance the performance of the myoelectric prosthesis during normal and muscle fatigue conditions. Using the proposed techniques, the performance of the MSD algorithm was maintained at >90% in non-fatiguing and fatigue conditions.
Recommended Citation
Allbunashee, Hamdi Allbunashee, "Adaptive Classification Framework for Control Myoelectric Prostheses During Muscle Fatigue" (2016). Theses and Dissertations. 714.
https://research.ualr.edu/etd/714
