Target Tracking With State Constraints Or Non-uniform Sampling

The research of modern tacking system is very important in the modernization of national defense and the development of national economy, which has been gaining more and more attention. Funded by the national defense research program, this thesis mainly addresses the target tracking with state constraints,the target tracking in periodic nonuniform sampling system and the Cramer-Rao Bound for target tracking. The analysis and simulations are provided. The main contributions are as followsA new filtering algorithm named as Delayed Update Filter (DUF) is developed for mining the steady-state performance when tracking a constant input/bias marine target. The statistical convergence and stability properties of the DUF were insured by the steady-state analysis. And the constant-gain filter is designed accordingly. Three categories of filters with state constraints are summarized and their CRLBs are investigated.A fusion-based target tracker for the sector scan radar is developed, then most prevalent tracking algorithms in uniform sampling system can be readily applied for periodic nonuniform sampling(PNS) system.A cascade filter is proposed for higher-order and multi-phase PNS tracking system. Closed-form, generalized filter gains and relationships for its steady-state form are obtained. Further, a CRLB-based adaptive algorithm is presented for staggering time estimation when tracking a target in the PNS system and the sensors have different measurement noise variances.The study of Cramer-Rao Bound for target tracking is extended to a more practical scene with the combination of the information reduction factor method and the enumeration method, which supplies a potent tool to predict accurately the best achievable tracking performance.