Abstract: An improved on-axis tracking scheme is proposed to effectively improve the tracking accuracy of photoelectric tracking system. The iterative unbiased finite impulse response (UFIR) filter algorithm is used to replace the traditional Kalman filter (KF) algorithm for state estimation and prediction. Because the iterative UFIR algorithm does not rely on the noise prior statistical information, it can overcome the shortcomings of the traditional KF algorithm and maintain the ideal prediction accuracy when the noise statistics are unknown, which makes the improved system also achieve the high tracking accuracy in actual operating condition. Two simulations were carried out，in which the KF was compared with the iterative UFIR filter， and two on-axis tracking systems which use above two algorithms were compared. Simulated results verify the correctness of the theoretical analysis, and show that the proposed scheme based on the iterative UFIR filter is more robust than the traditional KF-based scheme and has greater application potential in actual engineering where noise statistical information can't be accurately obtained in advance. Key

Key words: photoelectrictrackingsystem, on-axistracking, trackingaccuracy, unbiasedfiniteimpulseresponsefilter, Kalmanfilter