Identification of a nonlinear block-oriented model

Abstract

Key term separation principle, auxiliary model and a modified particle swarm optimization (MPSO) algorithm are applied to identify parameters of a block-oriented model represented by Hammerstein model with two-segment piecewise nonlinearities. Expressing output of the nonlinear Hammerstein models as a regressive equation in all parameters via the key term separation principle and an auxiliary model. Consequently, the problem of nonlinear system identification is changed into a function optimization over parameter space, and then a proposed MPSO algorithm is adopted to solve the optimization problem. Finally, numerical simulation experiments demonstrate the feasibility of the proposed identification algorithm.