Adaptive Sliding Mode Control of Functional Electrical Stimulation (FES) for Tracking Knee Joint Movement

Functional electrical stimulation (FES) has shown great potential in helping patients to achieve their joint movements. Since muscle system exists non-linear, time-varying and external disturbances, conventional controllers are difficult to achieve the precise control. In order to improve the accuracy of FES for knee joint movement, in this paper, a RBF neural network based sliding mode control method is designed. An electrical stimulation model of knee joint is first established, the nonlinear performance of RBF neural network is used to approximate the lower limb joint model uncertainties and external disturbances. To determine the width of hidden layer units and the architecture of the neural network, genetic algorithm is introduced to optimize the network structure parameters. Experimental results show that neural network sliding mode control based on genetic algorithm can accurately control the electrical stimulation to obtain the desired joint motion, and can be effectively compensated in the case of external disturbances.