Parameter estimation of Wiener-Hammerstein models via genetic algorithms

Conventional methods of estimating model parameters have difficulties with both nonlinear systems and with systems operating in noisy environments. In this paper, a modified genetic algorithm is used as a procedure to solve the parameter identification problem of the nonlinear Wiener-Hammerstein models. Numerical simulations are presented to illustrate the effectiveness of the proposed algorithm based on different input signals, and different noise-to-signal ratios of the output. Also, the algorithm is applied to model a DC generator with some nonlinear characteristics.