A Prosthesis Control System Based on the Combination of Speech and sEMG Signals and Its Performance Assessment

摘要

Surface electromyographic (sEMG) signals from the residual limb muscles after amputation have been widely used for prosthesis control. However, for the amputees with high-level amputations, there usually exists a dilemma that the sEMG signal sources for prosthesis control are limited but more limb motions need to be recovered, which strongly limits the practicality of the current myoelectric prostheses. In order to operate prostheses with multiple degrees of freedom (DOF) of movements, several control protocols have been suggested in some previous studies to deal with this dilemma. In this paper, a prosthesis control system based on the combination of speech and sEMG signals (Strategy 1) was built up in laboratory conditions, where speech commands were applied for the prosthesis joint-mode switching and sEMG signals were applied to determine the motion-class and execute the target movement. The control performance of the developed system was evaluated and compared with that of the traditional control strategy based on the pattern recognition of sEMG signals (Strategy 2). The experimental results showed that the difference between Strategy 1 and Strategy 2 was insignificant for the control of a 2-DOF prosthesis, but Strategy I was much better in the control of a prosthesis with more DOFs in comparison to Strategy 2. In addition, the positive user experience also demonstrated the reliability and practicality of Strategy 1.