Nussbaum-Type Neural Network-Based Control of Neuromuscular Electrical Stimulation With Input Saturation and Muscle Fatigue

Abstract

Abstract Neuromuscular electrical stimulation (NMES) is a promising technique to actuate the human musculoskeletal system in the presence of neurological impairments. The closed-loop control of NMES systems is nontrivial due to their inherent uncertain nonlinearity. In this paper, we propose a Nussbaum-type neural network (NN)-based controller for the lower leg limb NMES systems. The control accounts for model uncertainties and external disturbances in the system and, for the first time, provides a solution with rigorous stability analysis to the adaptive NMES tracking problem with input saturation and muscle fatigue. The proposed controller guarantees a uniformly ultimately bounded (UUB) tracking for the knee-joint angular position. To evaluate the control performance, a simulation study is taken, where the performance comparison with a NN controller inspired by Ge et al. (2004, “Adaptive Neural Control of Nonlinear Time-Delay Systems With Unknown Virtual Control Coefficients,” IEEE Trans. Syst., Man, Cybern.-Part B, 34(1), pp. 499–516) is given. The simulation results show a good tracking performance of the proposed controller regardless of the time-varying muscle fatigue and moderate input saturation. The adaptation mechanism of the Nussbaum-type gain and the controller's robustness to the muscle fatigue and input saturation are discussed in details along with the simulations.