Research On Maneuvering Target Tracking Algorithms Based On Particle Filter

Maneuvering target tracking has been a popular research area for years, this dissertation first introduces its elemental principle and methods, such as particle filter and multiple model method. Then, based on the analysis of the difficulty and character of the problem, this dissertation presents two improved maneuvering target tracking algorithms with particle filter. The first proposed algorithm combines the advantages of standard particle filter (SPF) and multiple model particle filter (MMPF): the SPF is adopted when the target is non-maneuvering and the MMPF is adopted when maneuvering. The proposed algorithm can detect the target maneuver effectively by a fuzzy logic controller and uses a backward correction sub-algorithm to alleviate the performance degradation caused by detection delay. The second proposed algorithm is a novel grey prediction based particle filter (GP-PF), which combines both model-based and model-free advantages. The basic idea of the GP-PF is that new particles are sampled by both the state transition prior and the grey prediction technique. Since the grey prediction is a model-free method, GP-PF is able to sample sufficient contributing particles even when the model is not accurate enough. The tracking performance of the two algorithms is compared with SPF and MMPF, in terms of tracking accuracy, computational complexity and stability, the simulation results indicate that the proposed algorithms have superior overall performance.