Development and human-like control of an upper limb rehabilitation exoskeleton using sEMG bio-feedback

Abstract

The paper presents our results in development and human-like control of 5-DOFs upper limb rehabilitation exoskeleton. We build the robot dynamics and propose human-like learning control using low-feedback gains and sEMG bio-feedback for the exoskeleton such that it can flexibly interact with different human users. Due to the harmonic drive transmission (HDT) of the exoskeleton, its motion is precise for co-manipulation tasks of human and robot because of its backdrivability, precise positioning capabilities, and zero backlash. However, it is challenging for making the development of an accurate dynamic model of the robot extremely because of the compliance and oscillations inherent in harmonic drive systems, and the lack of any technical information on the internal dynamics of the transmission. Moreover, it's impossible to get a precise human arm model for each user. The novelty of this work lies in the development of rehabilitation robot and extracting the model parameters of a harmonic drive transmission in the robot, together with human-like learning control and sEMG bio-feedback, to facilitate co-manipulation task. We have conducted the experiments on human body to verify the effectiveness of the proposed approaches.