Improving Human-Robot Interaction Safety through Compliant Motion Constraints in Bilateral Upper Limb Rehabilitation.

Robot-assisted bilateral upper limb training helps to activate the secondary motor brain areas and improve arms' coordinative capabilities for patients with neurological injuries. Due to the shared workspace of the patient and the robot, it is necessary to ensure human users' safety in a compliant way. While several approaches have been proposed for safe human-robot interaction, few considered evaluation of human motor function to provide subject-specific and compliant motion constraints. This study proposes a safety metrics for bilateral training from two aspects. On one hand, a human-kinematics-based method is used to customize safe interactive workspace. On the other hand, a repulsive potential function strategy is employed to ensure movement safety when towards workspace boundary, and a performance-based fuzzy logic is developed as an adaptive law to deal with mechanical collision. The proposed strategy was preliminarily validated for bilateral upper limb training with an end-effector robotic system. Experimental results validated the effectiveness and potentiality of the proposed safety strategies.