BCRLS Identification Method for Hammerstein-Wiener Model

In this paper, an improved two-stage on-line identification algorithm is presented to identify Hammerstein-Wiener systems with process disturbance. The proposed algorithm consists of two steps: Firstly, the bias compensation recursive least squares method is adopted to identify the parameter products of original system. By introducing a correction term in the estimate of recursive least squares, the estimation bias caused by process noise is compensated. Secondly, the average method is employed to separate original system parameters. The simulation shows that the proposed algorithm is effective.