Adaptive control of functional electrical stimulation for upper limb stroke rehabilitation

In this paper a multiple model adaptive control scheme is designed and evaluated for the electrically stimulated isometric human arm. Experimental data are first used to identify models that represent various stages of muscular fatigue. The control scheme is then implemented on four unimpaired participants supplying zero voluntary effort. Results illustrate the ability of the scheme to track the time-variance of the true plant and provide improved torque tracking compared with a fixed controller. An advantage of the proposed method is the ability to achieve similar levels of performance across multiple participants without the need to perform identification tests.