Reserch On Target Tracking Based On The Random Finite Set

Target tracking is to estimate states of targets by the measurements from sensors, which has been widely applied to military and civilian fields. Most target tracking algorithms assumed a "standard" model, which each target generates at most one measurement and each measurement generated by one target or clutter in every scan. However, in many practical target tracking applications, many targets violate the suppose of the "standard" model, such as extended targets, unresolved target, multipath target, multiple sensors target etc, which are known as a "non-standard measurement" target. The problem of "non-standard measurement" target tracking is more complex and realistic significance. In recent years, the target tracking algorithms based on random finite set (RFS) have attracted more attentions, which have become the main algorithm of target tracking. The article focuses on the problem of the tracking of "non-standard measurement" target and the target tracking algorithm based on the RFS. The main contents of this article are as follows:1. For the problem of multiple sensors target tracking, a novel approach of the multi-sensor Bernoulli filter (MSBF) based on RFS is proposed. First, the target state is modeled by Bernoulli RFSs, and then the FISST is used to derive the function of MSBF. Simulation results show that the proposed MSBF has excellent performance compared with the traditional MSBF based on the'iterated corrector approximation'.2. For the problem of dynamic model uncertainty in maneuvering target tracking, combining the MSBF with the multiple models (MM) algorithm, a multiple-model MSBF (MM-MSBF) has proposed. The simulation results demonstrate that the MM-MSBF can track a single maneuvering target effectively.3. In order to solve the multipath propagation problem in Over-the-horizon radar (OTHR), which one target can generate multiple measurements via different propagation paths, this article proposes a novel tracking algorithm based on RFS called the multipath probability hypothesis density (MP-PHD) filter. First, the target state is modeled by RFSs, and then the FISST is used to derive the function of MP-PHD filter, and then Gaussian mixture (GM) is introduced to derive the closed-form solution of the MP-PHD filter. Eventually, the simulation result shows that the MP-PHD filter can solve the multipath propagation problem effectively in OTHR.4. For the problem of multipath multitarget tracking in OTHR, further research of the target tracking algorithm based on the RFS has conducted, and a novel multipath cardinality balanced multitarget multi-Bernoulli (MP-CBMeMBer) filter is proposed in this article. First, the target state is modeled by multi-Bernoulli RFSs, and then the FISST is used to derive the function of MP-CBMeMBer filter, and then both the particle filter and GM method are applied to the MP-CBMeMBer filter. Simulation results show that the MP-CBMeMBer filter further improves the performance of multipath target tracking in OTHR.