Combined Signals Based Identification Scheme for the Hammerstein-Wiener System with Process Noise

An identification scheme of the nonlinear Hammerstein-Wiener system with colored noise with the help of combined signals is proposed. The combined signals are used to active the Hammerstein-Wiener system, resulting in the identification process are significantly simplified. Firstly, the correlation analysis algorithm is employed to identify the output nonlinear block parameters and linear block parameters, thus the process noise can be compensated by correlation function. Furthermore, the immeasurable internal variable and the unknown noise term in identification system are replaced with the output of designed auxiliary model and estimated residual, therefore, auxiliary model based recursive extended least square identification scheme can be obtained to learn the input nonlinear block parameters and noise model parameter. Simulations results are given to demonstrate the advantage of the identification scheme.