ADAPTIVE FUZZY CONTROL BY REAL-TIME CHOOSING MULTI-MODEL ARCHITECTURE FOR UNCERTAIN NONLINEAR SYSTEM

Guang Hu and Zhenbin Du

Keywords

Multi-model architecture, fuzzy model, DRBF neural network, nonlinear system, tracking control

Abstract

An adaptive tracking control method for an uncertain nonlinear system was studied. A control scheme for real-time choosing multi- model architecture was given when the controlled object could not be robustly controlled by a single fuzzy controller. First, several linear models and linear controllers at corresponding balance working points were designed according to the nonlinear system, which has several balance working points. Next, adaptive fuzzy models and fuzzy controllers combined with dynamic radial basis function (DRBF) neural network designed for potential problems in the uncertain nonlinear system. Last, choosing rules for real- time multi-model architecture was given. With these rules and the running condition of the system, the best controller could be chosen in real time to control it, and in this system, noise could be eliminated by DRBF neural network. Simulation results have demonstrated that the control scheme was feasible.

Important Links:

Go Back