Hierarchical control of human joint motion simulators

Abstract This paper presents a hierarchical adaptive algorithm for controlling upper extremity human joint motion simulators. A joint motion simulator is a computer-controlled, electromechanical system which permits the application of forces to the tendons of a human cadaver specimen in such a way that the cadaver joint under study achieves a desired motion in a physiologic manner. The proposed control scheme does not require knowledge of the cadaver specimen dynamic model, and solves on-line the so-called “indeterminate problem” which arises because human joints typically possess more actuators than degrees of freedom. Computer simulation results are given for an elbow/forearm system and wrist/hand system under hierarchical control, and demonstrate that any desired normal joint motion can be accurately tracked with the proposed algorithm. Additionally, these simulations indicate that the controller resolves the “indeterminate problem” redundancy in a physiologic manner, and show that the control scheme is robust to parameter uncertainty and to sensor noise.