Torque Estimation of Elbow Joint Using a Mechanomyogram Signal Based Biomechanical Model

Abstract

To perform appropriate assistive control, a wearable robotic exoskeleton needs to acquire the wearer's muscle activity directly and in real time. In this paper, we propose a elbow-joint torque estimator that uses a simplified musculotendinous model together with mechanomyogram (MMG) signals of the biceps brachii and triceps brachii muscles, and describe the procedures for determining the estimator's parameters. We then evaluate the accuracy of our joint-torque estimator through experiments comparing it with an estimator using electromyogram (EMG) signals. Experimental results demonstrate that the estimation accuracy of our estimator with MMG signals is almost equivalent to that using EMG signals and is adequate for the application of our estimator to assist-torque control of robotic exoskeletons.