Modelling and multi-innovation parameter identification for Hammerstein nonlinear state space systems using the filtering technique

This article focuses the problem of the modelling and identification for Hammerstein state space systems with coloured noise. In order to jointly estimate the system parameters and states, a filtering-based multi-innovation stochastic gradient algorithm is developed by combining the filtering technique with the multi-innovation identification theory. The key is that the estimation of the system parameters uses the estimated states, and the estimation of the states uses the preceding parameter estimates. The given examples confirm that the proposed algorithm can provide more accurate parameter estimates than the hierarchical multi-innovation stochastic gradient algorithm.