Identification of Hammerstein Models Based on Support Vector Regression

This paper presents a method for the identification of Hammerstein models based on support vector regression (SVR). First, the intermediate linear model was established through converting the nonlinear equations of Hammerstein to a class of linear one by the function expansion. Second, training samples for intermediate linear model were obtained by operating measured data synthetically, and coefficients of the intermediate model were obtained by the SVR algorithm. Then, through the relations of the coefficients of intermediate model and that of Hammerstein model, the nonlinear static part and linear dynamic part were identified simultaneously. Finally, The efficiency of the proposed algorithm was demonstrated by simulation examples.