Blind nonlinear system identification based on a constrained hybrid genetic algorithm

System identification is an important issue in communication, instrumentation, and control systems. In this paper, we proposed a method with higher-order cumulant fitting for nonlinear system identification. Compared with the conventional method, which uses second-order cumulant as a constraint, the proposed method uses fourth-order cumulant in order to smooth out the additive Gaussian noise. Since the cost function with higher-order statistics has local minima, we also propose to use a hybrid method of simplex and genetic algorithms to minimize the cost function. The applicability of the proposed method is demonstrated by the computer simulations.