RBF Networks-Based Weighted Multi-Model Adaptive Control for a Category of Nonlinear Systems With Jumping Parameters

This paper studies the tracking control problem for nonlinear system with largely jumping parameters. To deal with the large uncertainties of system parameters, an RBF networks-based weighted multi-model adaptive control (WMMAC) strategy is designed, in which the model set can be constructed to cover and approximate the variation range of the plant parameters. Correspondingly, the RBF network controller set generates the appropriate global control signals when parameters jumping, and better transient performance is obtained. Different from existing neural network learning methods, two novel learning rules based on the difference value between the current value of an objective function and its optimal value for the RBF networks are developed to achieve a fast convergence rate. The stability of the overall closed-loop system is validated by the virtual equivalent system (VES) theory, and favorable performance of the proposed control strategy is demonstrated by the numerical simulations.