Determination of Thresholds of Target Maneuver Detection via Markov Chain Approach

Wen-Chieh Yang, Fan-Ren Chang and Yi-Hsueh Tsai

Abstract:	Target maneuvers refer to unpredictable changes in target motion, which may cause serious inaccuracies in modeling the system. As a result, loss of track may take place during the target tracking. From the surveillance point of view, the observers must detect the maneuvers quickly and correctly once the objects move in unexpected ways. Usually the overall criterion is quick detection of maneuvers under a fixed false alarm rate. In order to meet this criterion, the threshold corresponding to a specified false alarm rate must be accurately computed. The fading memory average has been proposed to detect target maneuver, which accumulates all past and present data of the normalized innovations squared produced from the Kalman filter. In most literature, the fading memory average is approximated to a simpler random variable by moment matching, through which the thresholds can be found. These methods cannot provide sufficiently accurate thresholds related to a fixed false alarm rate, however, as they do not take into account the dynamic property of the fading memory average. To meet the requirement of the fixed false alarm rate, an algorithm based on the Markov property of the fading memory average is proposed in this paper to accurately evaluate the threshold of fading memory average under a fixed false alarm rate. Monte Carlo simulation results verify that, in comparison to the conventional methods, the proposed Markov-chain-based approach can provide more accurate thresholds.
Published in:	Proceedings of the 58th Annual Meeting of The Institute of Navigation and CIGTF 21st Guidance Test Symposium (2002) June 24 - 26, 2002 Hyatt Regency Hotel Albuquerque, NM
Pages:	364 - 371
Cite this article:	Yang, Wen-Chieh, Chang, Fan-Ren, Tsai, Yi-Hsueh, "Determination of Thresholds of Target Maneuver Detection via Markov Chain Approach," Proceedings of the 58th Annual Meeting of The Institute of Navigation and CIGTF 21st Guidance Test Symposium (2002), Albuquerque, NM, June 2002, pp. 364-371.
Full Paper:	ION Members/Non-Members: 1 Download Credit Sign In