Identification of Wiener Model with Output Colored Noise Based on Separable Signal Sources

Most of research works discuss the identification problem of Wiener model parameters without considering the colored noise of the actual industry process and lacking the physical basis. In order to solve the problem, a Wiener model identification algorithm based on separable signal sources is proposed in this paper to identify the Wiener nonlinear subsystem with output noise interference. The separable signals are used to separate the dynamic linear subsystem from static nonlinear subsystem. Then, the correlation analysis approach is adopted by using a set of Gaussian signals to identify the parameters of dynamic linear subsystem. Further, the parameter identification algorithm based on recursive generalized extended least square method is proposed to estimate the parameter of the dynamic nonlinear subsystem and the noise subsystem by using a set of binary signals as the input signals.