Improvement of Estimation Accuracy of Body Motion Based on Muscle Potential Signal

Abstract

The purpose of this research is to develop a man-machine interface that changes machine impedeace based on surface myoelectric potential. The impedance model of the wrist joint is constructed from the measured data to derive the moment of inertia, viscosity coefficient and rigidity of the human’s wrist joint, investigate the relationship between human joint impedance and electromyography (EMG), according to the change of EMG from the relationship, We have developed an interface to change the mechanical impedance of the robot arm. We investigate the relationship between human joint impedance and EMG, and apply the sparse modeling method to the modeling process. We have created an interface to change the mechanical impedance of the robot arm based on the model.As a result, myoelectric potential and force data have been acquired. Also, mechanical impedance has been acquired.In this study, we improve the estimation accuracy of physical motion by selecting characteristic parameters and terms using sparse modeling, which is considered to be feasible by estimating the force that occurs at the same time as the position at the time of physical movement.