EMG-based control for lower-limb power-assist exoskeletons

Abstract

Power-Assist for lower-limb is expected by many physically weak persons to assist their daily activities. It is important for the power-assist robots that the assist motion is generated based on users motion intention. This paper presents a muscle-model-oriented EMG-based (electromyogram-based) control method to activate a lower-limb power-assist robot according to the users motion intention since the EMG directly reflects the users muscle activities. In the proposed method, a matrix which expresses the relationship between the muscle activities and the generating joint torque is applied to estimate the users motion intention in real-time. The effectiveness of the proposed control method was evaluated by performing experiments.