A fast algorithm without static error for generalized predictive control of Hénon chaotic system

Abstract

A kind of fast generalized predictive control algorithm without static error for Hénon chaotic system is proposed. Firstly, the chaotic system is identified by improved recursive least squares parameter of time-varying forget factor. Then the softness factor is adjusted on line through combining MP neural network with BP algorithm, thus the fast tracking without static error of reference signa1 is realized by introducing the feedforward gain matrix and softening coefficient matrix into regular performance index function of generalized predictive control. The algorithm avoids the matrix inversion computation and has strong ability of tracking the reference signal. The simulated results show the effectiveness of this algorithm.