Muscle-Model-Oriented EMG-Based Control of an Upper-Limb Power-Assist Exoskeleton with a Neuro-Fuzzy Modifier

Many studies on power-assist exoskeleton robots have been carried out in order to assist daily activities and/or rehabilitation of physically weak persons. EMG-based control is one of the most effective control methods to realize the power-assist with the exoskeleton based on user's motion intention. In this paper, a muscle-model, which is adjusted by a neuro-fuzzy modifier according to the user's upper-limb posture, is introduced to realize an effective EMG-based controller of the power-assist exoskeleton. Force/torque generated between the user's wrist part and the tip of the exoskeleton is used to train the neuro-fuzzy modifier. The effectiveness of the proposed control method was evaluated by performing experiment.