A novel nonlinear model is proposed for nonlinear system blind identification with infinite memory. This model is a generalization and combination of both ARMA model and Volterra-Hammestein(NMA-Hammestein) model. In the process of blind identification, two steps are involved. First, we focused on the left part(AR) of model form,which could be identified by using the generalized BBR formula and MYW(modified Yule-Walker) normal equation. Then, the problem changes into the blind identification of Volterra-Hammestein model, thus several methods could be employed. The simulation results showed its good estimation performance for nonlinear system.
展开▼
