Multi-model approach for nonlinear system identification by EM algorithm

Abstract

The problem of system identification for nonlinear system is studied in this paper by using EM algorithm, and a stochastic scheduling parameter which follows a Markov jump process is considered. First, multi-model approach is addressed to describe the nonlinear process, where each linear parameter system is represented by an auto regressive exogenous model, and then, EM algorithm is used to do estimation with the help of stochastic scheduling parameter. A simulation example is given to illustrate the effectiveness of the approach proposed.