Skip to main content

ECE PhD Dissertation Defense by PhD Candidate, Ziling Zhu, via Zoom

Wednesday, April 21, 2021
1:00 pm
Floor/Room #: 
via Zoom (please email Ziling Zhu [] for the zoom link if you would like to attend)


Proportional, Simultaneous and Independent Control (PsiCon) of Upper-Limb Prostheses for the Limb-Absent Population



Limb-absent subjects can generate motor commands that are communicated to remnant muscle tissue, which contracts and provides a measurable electromyogram (EMG). EMG is a powerful tool utilized in various medical assessments, rehabilitation, and myoelectric control of prostheses and orthoses. Proportional myoelectric control of one degree of freedom (DoF) prosthesis tasks has been available commercially for decades—including systems which support sequential switching between distinct DoFs. More functional control of multiple DoFs is mostly found in research systems and is primarily limited to 2-DoFs. Such control has typically been facilitated via multiple EMG sites or advanced machine learning algorithms.


This study developed a proportional, simultaneous and independent 2-DoF hand-wrist prosthesis control (PsiCon) for able-bodied and limb-absent subjects, from offline to online testing.


(1)   A fundamental challenge for developing EMG-force models in limb-absent subjects is that end-effector force is not available as the output of supervised model training. We explored three alternative approaches by using contralateral force of target movement as outputs with different visual feedback types for both able-bodied and limb-absent subjects. An extension of this study exhibited a direct measure as to how well forces in the contralateral limb are representative of forces in the dominant limb of able-bodied subjects.

(2)   We developed a complete prosthesis real-time control system based on a regression algorithm. We tested 2-DoF control with three different control strategies: 2-DoF mapping control, 2-DoF direct control (each with 6 or 12 electrodes), and conventional 2-site control with co-contraction mode switching for both able-bodied and limb-absent subjects. Three tasks were tested to judge the performance with different control strategies and different number of electrodes, showed that conventional 2-site control is still necessary in 1-DoF control, while new mapping and direct control techniques had great advantages and prospect in 2-DoF control.


Research Advisor:

Prof. Edward A. Clancy

ECE Department, WPI


Committee Members:

Dr. Xinming Huang

ECE Department, WPI

Dr. Todd Farrell

Liberating Technologies Inc., Holliston, MA