Neuromuscular electrical stimulation (NMES) is a technique that is widely used as a tool for rehabilitation and restoration of basic functions for people suffering from upper motor neuron lesion (UMNL). Closed-loop methods have shown a potential for improving the effectiveness of NMES. In this paper, uncertainties in the muscle contraction dynamics are taken into consideration when compensating for the muscle contraction dynamics. Accounting for the muscle contraction dynamics is a challenge because of uncertainty, nonlinearity and the fact that the contraction states are not measurable. A neural-network (NN)-based controller together with a dynamic NN-based identifier is designed to enable semi-global uniformly ultimately bounded tracking of a desired limb trajectory and on-line estimation of the limb acceleration. The overall stability of the identifier-controller system is analyzed through Lyapunov methods. Simulation results are provided to illustrate the controller performance.

This content is only available via PDF.
You do not currently have access to this content.