Biomechanics Aware Collaborative Robot System for Delivery of Safe Physical Therapy in Shoulder Rehabilitation

In this work, we explore using computational musculoskeletal modeling to equip an industrial collaborative robot with awareness of the internal state a patient safely deliver physical therapy. A major concern robot-mediated therapy is that robots may unwittingly injure patients. For patients shoulder injuries typically means risk tearing rotator-cuff muscle tendon. Risk reinjury hampers both human and therapists it main reason for conservative Advances in modeling, however, can additional perception potential risks. While ultimate goal improve safety, range-of-motion activity receive through therapy, aim letter develop test framework enables understand execute movements demonstrate low injury achieve large subjects. We build on prior work human-robot interaction via impedance control, but take next level by including manipulating model parallel patient. Taking most common impairments (i.e., tears) as example, planned, model-based trajectories minimize strain these muscles corresponding healthy Our experiments suggest promising approach plan therapeutic are safe effective robot.