Adaptive fuzzy neural controller design

A new design approach for adaptive fuzzy logic controller (AFLC) based on neural-net concept is presented in this work. The main ideas of our proposed strategy include: (1) connectionist implementation of fuzzy logic controller (FLC) using a simplified fuzzy inference network (SFIN); (2) initialisation of the AFLC using supervised learning control; and (3) online adaptation of the AFLC using GPC performance index. A nonlinear plant with nonminimum phase characteristics and uncertainties is employed in our simulations. The observed results demonstrate the effectiveness at the presented AFLC