Bias compensation-based parameter and state estimation for a class of time-delay non-linear state-space models

This study presents, based on bias compensation, an integrated parameter and state estimation algorithm for a class of time-delay non-linear systems which are described by canonical observable state-space model. In technical development, the state-space system model is transformed into an input-output representation/realisation by eliminating the state variables, which is accordingly used as a feasible identification model. With such an input-output structure, directly data measurable to accommodate the estimation bias, an augmented least-squares algorithm (by adding the bias correction terms into the estimates) is proposed for estimating the parameters and states interactively. Regarding the estimator properties, the proposed algorithm is proved unbiased. The simulation results show that the proposed algorithm has good performance in estimating the parameters of state-space systems.