Identification of Hammerstein model based on dynamical separation technology

A novel identification approach based on dynamical separation technology is proposed for a class of Hammerstein system. By using two different pseudo-random binary signals as the system input, the linear dynamical part of the Hammerstein system can be separated by simple algebraic operations, and the parameters of the linear part can thus be identified. The intermediate input between the nonlinear part and linear part is then reconstructed by using the observed output based on the estimated linear part. Finally, the nonlinear element is identified by using the test input signal and the reconstructed intermediate input. The effectiveness of the proposed algorithm is demonstrated by simulation examples.