Adaptive H∞ Control Using Backstepping Design and Neural Networks

Abstract

In this paper, the adaptive H∞ control problem based on the neural network technique is studied for a class of strict-feedback nonlinear systems with mismatching nonlinear uncertainties that may not be linearly parametrized. By combining the backstepping technique with H∞ control design, an adaptive neural controller is synthesized to attenuate the effect of approximation errors and guarantee an H∞ tracking performance for the closed-loop system. In this work, the structural property of the system is utilized to synthesize the controller such that the singularity problem of the controller usually encountered in feedback linearization design is avoided. A numerical simulation illustrating the H∞ control performance of the closed-loop system is provided.