Neural Adaptive Dynamic Surface Control via Sampled-data Observer and Nonlinear Feedback and Application to Automatic Train Operation

In this paper, a neural adaptive dynamic surface control approach based on the sampled-data observer and nonlinear feedback function is proposed for a class of uncertain nonlinear systems. The sampled-data observer is employed to estimate the unmeasurable states using available sampled data output information, and neural networks are applied to approximate the unknown nonlinear functions. The dynamic surface control technique can circumvent the explosion of complexity caused by the repeated differentiations of nonlinear functions in the traditional backstepping methodology. Nonlinear feedback scheme is designed for the intermediate and actual controllers based on a continuous differentiable nonlinear function, which synthesizes the merits of high-gain and low-gain control. Finally, the proposed control scheme guarantees that the tracking error converges to a small neighborhood of the origin and the closed-loop signals are semi-globally uniformly ultimately bounded. Simulation results illustrate the effectiveness of the proposed scheme, and satisfactory speed tracking and position tracking can be guaranteed in Automatic Train Operation (ATO) systems.