Iterative learning identification of time-varying parameter based on global newton method

Abstract

Iterative learning identification algorithms of time-varying parameters for nonlinear systems are presented in this paper. The iterative learning control based on Newton method is extended to the identification model of nonlinear systems. A Newton-type iterative learning identification scheme with time-varying parameters is proposed. The convergence of this algorithm is analyzed and proved. In order to improve the performance of choosing initial parameters, the iterative learning identification procedure is established to develop its extension in the iteration domain by the extension method. The global convergence of the iterative learning identification algorithm is given and proved. The proposed iterative learning identification algorithm based on global Newton method is applicable to converging globally and choosing the initial time-varying parameter arbitrarily.