NN-based adaptive control for a class of MIMO nonaffine nonlinear systems

Based on neural network (NN) and robust control, an adaptive controller is developed for a class of MIMO nonaffine nonlinear systems. First, take the Taylor series expansion of the original system in the neighborhood of the operating trajectory. Then estimate the unknown functions of the system using NN. Finally, use robust control to overcome the affects of NN estimated error terms and high order terms of the Taylor series expansion. All the signals in the closed-loop system are proven to be uniformly ultimately bounded (UUB) and the mean square and L-infinite tracking error bounds are given in this paper. The simulation results show the effectiveness of the proposed approach.