인간-기계 인터페이스를 위한 근육 진동 신호 기반 팔꿈치 굽힘 힘 예측

Abstract

The mechaomyography (MMG) is generated by the muscle surface oscillations due to pressure waves originating from the lateral expansion of the contracting muscle fibers. Since MMG signals reflect important aspect of muscle activities including the number and firing rates of recruited motor units (MUs). MMG signals could be utilized for predicting the exerted force by skeletal muscles. The objectives of this study were to investigate the feasibility of predicting isometric force from MMG signals and to explore the effect of muscle fatigue in performance of the force prediction. MMG signals from three contributed muscles during isometric elbow flexion were recorded as well as the force at wrist. Artificial neural network (ANN) with muscle fatigue compensation model were constructed and trained using the MMG signals collected for each subject, It was found that the isometric elbow flexion force could be accurately predicted by proposed methods, based on the values of normalized root mean square error (NRMSE) and the correlation coefficient (CC), with the performance of the ANN with fatigue compensation model being significantly better (NRMSE = 0.106±0.049, CC= 0.918±0.083) that those of its ANN (NRMSE = 0.160±0.045, CC= 0.710±0.211).