A Novel Telerehabilitation System Based on Bilateral Upper Limb Exoskeleton Robot

Abstract

In this paper, a novel bilateral upper limb telerehabilitation training system is presented. By wearing two MYO armbands, one worn in the middle of the upper arm and the other in the middle of the forearm, the therapist could remotely control a 12 degree-of-freedom (DOF) bilateral upper limb exoskeleton robot to perform post-stroke rehabilitation training for stroke survivors. Through IMU sensors and EMG sensors embedded in MYO armbands, the telerehabilitation system acquires the joint angles and EMG signals of the therapist during upper limb movements and reconstructs these movements on the rehabilitation exoskeleton robot to help the patient perform rehabilitation training. In order to confirm the effectiveness and availability of this system, we analyzed the angles change obtained by the IMU sensors during the upper limb movements and simulated these actions in MATLAB; we also compared the classification accuracy by KNN, LDA, logistic regression and SVMs. At last, giving priority to efficiency, combination of five principal components and Gaussian kernel SVM is meeting for the real-time classification requirements. The results of experiments prove the feasibility of the novel telerehabilitation system.