The design of control methods for neuroprostheses to restore multiple degrees of freedom motor functions

In this paper, development of control methods for future generation of functional electrical stimulation (FES) systems is discussed. A reflex-like controller based on muscle EMG feedback signals is explored for control of elbow extension movement. The authors also propose a self-adjusting neural network controller that endows FES systems with ability to learn a local model and to optimize control with respect to an individual patient. Computer simulations of these control methods are performed, and clinical evaluation is under way