Stiffness control of variable stiffness joint using electromyography signals

We have developed a multi-functional joint mechanism with variable joint stiffness and the ability to sense external forces. In this joint, silicone rubber cushions are employed to adjust the joint stiffness by controlling the pressure in the cushions. In this study, we constructed a basic control system to synchronize joint motion and stiffness with those of a human, for potential application in human-assistive robotic technologies. This system consists of a human-motion measurement tool, a system to control joint position, and a pneumatic control system for joint stiffness. Human joint stiffness was estimated by inputting electromyography signals into the mathematical musculoskeletal model proposed by Shin[9]; the result was inputted into the joint stiffness control unit. To synchronize the robot motion and joint stiffness with that of a human, we introduced a compensator into the joint stiffness control loop. This element recovers the delay and insufficient gain of a pneumatic system. Through several experiments, we confirmed that this basic control system satisfactorily follows human dynamic motion, reflecting the change in joint stiffness. In addition, the experimental results showed that the adjustment of joint stiffness is necessary for a robot to achieve human-like dynamic motion.