Functional Link NN based Adaptive Fuzzy Control for Nonlinear Dynamic Systems

Abstract

Since their inception, fuzzy logic and its variants involving neural networks have witnessed tremendous applications in the area of identification and control of nonlinear dynamic plants. Fuzzy logic being the universal approximator becomes more powerful when combined with inherent learning capability of Neural Networks (NNs). This research presents a novel adaptive fuzzy control based on Functional Link NNs (FLNNs). The Laguerre orthogonal polynomials have been used for functional expansion of FLNNs. The parameter adaptation and thus the shape of the membership functions and weights of the polynomials of FLNNs are adapted online based on gradient descent optimization technique. Finally, the proposed control scheme has been checked for its performance using comparative evaluation with conventional control schemes applied to different nonlinear plants. The nonlinear time domain simulation results and their quantitative analysis validate the superior performance of the proposed adaptive fuzzy FLNN control.