UNEXPECTED RESULTS OF EXTENDED FRACTIONAL KALMAN FILTER FOR PARAMETER IDENTIFICATION IN FRACTIONAL ORDER CHAOTIC SYSTEMS

摘要

The Extended Fractional Kalman Filter (EFKF) for nonlinear discrete stochastic fractional order systems is studied in this paper. Perfect synchronization of chaotic systems is achieved by the EFKF algorithm in the presence of channel additive noise and processing noise. However, the EFKF has a limitation for the unknown parameter identification. The parameters do not converge to their real values in many circumstances. The reason for these failures is analyzed. Finally, based on the fractional Chen system, two numerical examples are provided to illustrate the effectiveness of the proposed method.