ROBOT ASSISTED STROKE REHABILITATION: JOINT TORQUE/FORCE CONVERSION FROM EMG USING SA PROCESS

Abstract

This paper focuses on the implementation of robot-assisted stroke rehabilitation using electromyography (EMG) as the interface between the robot and subjects. The key issue in implementing EMG for this application is the conversion process of EMG signal into torque/force, which is used as a input to the control system. This paper presents a methodology of EMG signal conversion into estimated joint torque by using simulated annealing (SA) technique. Basic principle of SA, formulation, and implementation to the problem are discussed in this paper. Experimental studies with real life EMG data have been carried out for five subjects. These studies are used to evaluate the feasibility of the methodology proposed for robot-assisted stroke rehabilitation problem. Experimental investigations and results are discussed at the end of the paper.