Recursive Estimation of the Takagi-Sugeno Models II: Estimation of Hammerstein Models

Abstract

This paper considers the identification of a class of nonlinear systems. It is assumed that the model have block-oriented structure. The Hammerstein model will be considered within this model structure. Static nonlinearity is a polynomial function of the input signal. A linear part is described by discrete transfer function. The Hammerstein model is approximated with finite collection of linear models by Takagi-Sugeno model. In order to implement this model, it is needed to perform the fuzzy decomposition of the entire input signal space. Using Gustafson-Kessel algorithm and least squared method the membership functions are determined. Also, the least squared method is used for estimation of Takagi-Sugeno model.