A Novel Adaptive NN Control for a Class of Strict-Feedback Nonlinear Systems

Abstract — An adaptive neural network control(ANNC) is proposed for a class of strict-feedback uncertain nonlinear sys-tems with unknown system nonlinearities and unknown virtual control gain nonlinearities. Combining the dynamic surface con-trol(DSC) technique with minimal-learning-parameters(MLP) algorithm, a systematic procedure for synthesis of ANNC is developed based on the universal approximation of neural networks. An important feature of the proposed algorithm is that the number of parameters updated on line for each subsystem is reduced only to one, both problems of “explosion of complexity ” and “curse of dimension ” are solved simulta-neously, such that the computation load is reduced drastically and it is convenient to implement the controller in applications. It is shown that all closed-loop signals are semi-global uniform ultimate bound(SGUUB) via Lyapunov stability theory. Finally, simulation results are presented to demonstrate the effectiveness of the proposed scheme. Index Terms — Uncertain nonlinear systems, neural networks, adaptive control, dynamic surface control, minimal-learning parameters. I