Optimal Impedance Control of an Upper Limb Stroke Rehabilitation Robot

Abstract

There is an increasing prevalence of stroke, spinal cord and brain injuries worldwide. Robotics are becoming a popular form of rehabilitation, but their effectiveness is yet to be proven. Recent motor learning theories motivate us to improve the control algorithms of these robots. This work proposes an optimal impedance control (i.e., a linear quadratic regulator mapped from the operational space) to control a 2 degree of freedom (DOF) upper extremity stroke rehabilitation robot. We used an integrated model of the robot and human (a 2 DOF musculoskeletal arm model) to develop and fine-tune the controller.