Bias compensation based recursive least squares identification for equation error models with colored noises

It is well known that the least squares estimation of ARMAX models is biased. In this paper, by combining the principle of bias compensation and hierarchical identification, a new identification is established for this equation error model with moving average noises. The proposed estimate of the system parameter is given by the least squares estimate modified by a correction term. A numerical example is employed to show the advantage of the proposed estimation algorithm.