An effective approach to nonlinear Hammerstein model identification using evolutionary neural network

In this paper, a new approach to nonlinear system identification using evolutionary Neural Networks and LMS algorithm has been proposed. System in our method consists of a static nonlinear function in series with a dynamic linear function, which has been refers to as Hammerstein model. NN, in the form of nonlinear function, is implemented to approximate nonlinear term, where GA is responsible for finding optimal weights of the NN. GA also offers linear system order, which is used to estimate linear system coefficients through LMS. AIC is used as the fitness function of the GA. Chehychev's polynomials and Taylor's power series are also employed, where simulation results present the effectiveness of the NN with respect to latter functions.