Study on Hammerstein models of sparse nonlinear identification with LS-SVM

These instructions give you basic guidelines for preparing papers for conference proceedings. This paper put forward a kind of new method aimed at identification of Hammerstein models in nonlinear system. With the analysis of the problem of sparseness that based on least square support vector machine (LS-SVM), the predictive output function of static nonlinear loop was obtained, and applied state space model to dynamic linear loop, got a new nonlinear identification model of Hammerstein. Simulation results shows that the method has better precision of identification, high-speed of response. Its computing time approximately for other modelspsila 30%, improved identification efficiency obviously, and testified the validity and feasibility of the method.