A Novel Robot-Aided Therapy for Shoulder Rehabilitation after Stroke: Active-Assisted Control of the RehaArm Robot Using Electromyographic Signals

Abstract

In this work, we introduced a novel experimental protocol for shoulder rehabilitation after stroke using RehaArm, a three Degrees of Freedom compliant robot. We implemented a control framework based on sensorimotor integration by developing an electromyographic (EMG) driven operation of the robot. EMG signals from five muscles were collected during experimental protocol consisting of four principal rotations of the shoulder joint. The subject movement intention was detected by monitoring the EMG activity of the primary agonist muscle for the selected task. Whenever the EMG activity was above the threshold set to 20% of maximum voluntary contraction, the robot provided assistive forces towards the target position. The system was tested in four healthy subjects and one stroke survivor. All subjects were able to produce continuous EMG activation in target muscles in order to smoothly control the robot. Healthy subjects exhibited normal on-off pattern of activity, while in the stroke patient an abnormal activation was observed characterized by a loss of selective recruitment of some muscles. The results of this preliminary evaluation suggested that the developed closed loop framework is a suitable platform for novel robotic treatments for shoulder rehabilitation after stroke.