Non-iterative Estimation Algorithm for Time-delay Hammerstein State Space System

This paper aims at the parameter estimation method for time-delay Hammerstein state space system. The Hammerstein system studied is composed of a static nonlinear module approximated by a neural fuzzy model and a dynamic linear module modeled by a time-delay state space model, and input test signals which is composed of separable signals and random signals are applied to parameter separation estimation of the time-delay Hammerstein system. Firstly, according to the properties of the unit backward operator, the time-delay state space system is transformed into a transfer function, and the parameters of the linear module are estimated by using correlation analysis method based on Gaussian signals. Moreover, clustering and Taylor series expansion methods are used to estimate nonlinear module, that is, the neuro fuzzy model parameters. The simulation results show the proposed method can effectively estimate the time-delay Hammerstein state space system.