Self Recurrent Wavelet Neural Network Based Direct Adaptive Backstepping Control for a Class of Uncertain Non-Affine Nonlinear Systems

This paper proposes an adaptive backstepping control strategy for a class of uncertain non affine systems using self recurrent neural networks. To assure the stable tracking of nonlinear non affine system, it is first converted to an affine like form and subsequently a wavelet based adaptive backstepping controller is developed. Self recurrent wavelet neural network (SRWNN) is used to approximate the uncertainties present in the system as well as to compensate the highly dynamic nonlinearities inserted by these uncertainties in the control terms. In addition robust control terms are also designed to attenuate the approximation error due to SRWNN. Based on the Lyapunov theory, the online adaptation laws and stability of the closed loop system are verified. A numerical example is provided to verify the effectiveness of theoretical development.