An optimal algorithm for identification of rapidly time-varying systems

An algorithm for identification of deterministic rapidly time-varying systems was proposed by G. Davidov et al. (Proc. IEEE, vol.75, no.8, Aug. 1987). This algorithm is based upon applying a linear combination of the least-squares method equations. It is shown that this algorithm differs from the optimal algorithm, being a degenerate case of the Kalman filter equations if disturbances in the parameter update law and observation equation are absent.<
>