Tracking control for a class of uncertain nonlinear systems using Legendre Neural Network

Abstract

This work presents a Legendre Neural Network (LNN) based adaptive tacking controller for uncertain nonlinear systems. Due to the features of orthonormality, compact support and enhanced spectral response with increment in degree these polynomials posses promising function approximating capabilities. Proposed LNN utilizes the Legendre polynomial as activation function thereby improving the approximation capability of conventional feed forward neural network The effectiveness of LNN has been shown by considering the non linear system identification for an uncertain non linear dynamic system. The performance of the proposed network is found superior to that of a conventional neural network Adaptation laws are developed for the online tuning of the neural network parameters.