Identification of Hammerstein nonlinear dynamic systems using neural network

For nonlinear single-input single-output (SISO) Hammerstein model, a novel method for nonlinear system identification is proposed by using a special neural network structure. The identification problem is converted into the training problem of neural network, and the error back propagation algorithm is then adopted to solve the iterative training problem. Lastly, the parameters of memory-less nonlinear gain and linear dynamic subunit in Hammerstein model can be identified synchronously. The applicability of this estimate technique is demonstrated by simulation results. The results also show that the proposed method is simple and efficient, so it can be easily popularized.