State-Predictor-Based Adaptive Neural Dynamic Surface Control of Uncertain Strict-Feedback Systems with Unknown Control Direction

This paper presents a state-predictor-based adaptive neural dynamic surface control (ANDSC) scheme for uncertain strict-feedback systems with unknown control direction. In practice, there always exists a compromise between the system transient performance and stability. In order to attain satisfactory performance, state-predictors are designed to produce two time scales, which improve the transient performance and reduce the tracking error magnitude without generating high-frequency oscillations. And the Nussbaum-type gain technique is introduced to handel the problem posed by unknown control directions. Besides, radial basis function neural networks (RBFNN) are used to deal with system uncertainties. By Lyapunov stability theory analysis, the proposed control scheme is proved to be capable of ensuring that all closed-loop signals are uniformly ultimately bounded (UUB). A numerical simulation example is given for illustrating the correctness and effectiveness of the theoretical result.