Study On Target Tracking Algorithm Of LEO Satellite Constellation

The ballistic missiles (BM) are becoming an important war weapon and strategic power pursuited by countries in the world. The powers are increasingly strengthening the BM early warning system which will provide earlier detection, entire-flight track, correct identification, timely trajectory prediction, and counterattack and interception assistance. The early warning system based on the LEO satellite constellation are drawing more and more attention, thanks to its advantages, such as high placement, superior visibility, earlier warning, especially owning the capability of detecting the boost phase object and identify the coast phase object.The research work in this dissertation focuses on developing and enriching the target tracking algorithm based on the LEO satellite constellation. The contributions and innovation points are as follows:(1). Studying the single target tracking of LEO satellite constellation. The Fisher information matrix (FIM) of sigle observation for current state is examined, together with its eigenvalus and eigenvectors, which illustrates that the single observation will provide no information contribution for target state in the direction from the satellite to the target. In the case of estimating the initial state using single satellite, an more optimal iterative algorithm is deduced, with the purpose to resolve the coupling problem of the measurement noise covariance and the evaluation angle directly. The CRLB of on-line target tracking with origin uncertainty is studied, which can assist the estimator's efficiency evaluation.(2). Studying the multi-target tracking (MTT) of LEO satellite constellation. Multitarget multisensor management is a basic problem in the MTT field. The PCRLB-based multi-sensor management algorithm is probed, with the benefit of establishing the basic framework and providing an available strategy for the revolution of the sensor management problem of LEO satellite constellation.(3). Studying the multi-target tracking (MTT) of LEO satellite constellation. The closely-spaced target tracking is the most complicated problem in the MTT field. The EM-based algorithm for the coast-phase closely-spaced target tracking is deduced, which will assist alleviating the track biase and improving the track precision, and broadening the application of the EM method.(4). Studying the target tracking algorithm in the complicated cases. The model of predicting handover for initial burnout target is probed, which will provide the numerical analysis. In order to identify target and decoy, a new data association algorithm, utilizing kinematic information only and based on batch processing and hard decision, is probed, with the contribution to the revolution of the target's identification problem in the presence of decoys.