Auxiliary model method for transfer function estimation from noisy input and output data

In order to get the unbiased parameter estimation of transfer function from noisy input and output data, this paper presents an auxiliary model based method. The basic idea is to replace the unmeasurable outputs in the information vector of the system by auxiliary model outputs, and to modify the recursive least squares parameter estimation by the errors between the system output and auxiliary model output. The proposed method improves the precision of parameter estimates, and has wide application in modelling and control engineering without strict mathematical assumption. Finally, the advantages of the proposed approach are shown by simulation tests.