IMM-PDA滤波器的模型概率加权跟踪门研究

Model Probability Weighed Gating Techniques for IMM-PDA Filter

杂波情况下跟踪机动目标的一个重要问题就是跟踪门问题。文中研究了两种现行的跟踪门技术,模型跟踪门和集中跟踪门。提出了新的跟踪门技术,分别为模型概率加权跟踪门和两极模型概率加权跟踪门(TS-MPWG),TS-MPWG考虑了可能的模型误差,且当建立第二个波门时在模型误差变量上用到了上界。仿真里采用了一种高机动目标的基准轨迹,用RMS误差、跟踪丢失百分比和计算量比较了讨论的跟踪门技术的性能。从仿真结果可以看出,在杂波环境下用IMM-PDAF进行机动目标跟踪,TS-MPWG跟踪门技术优于其他方法。

One of the central issues in tracking a maneuvering target in clutter is related to gating. We reviewde two existing gating techniques,namely , Model based gating（MBG） and centralized Gating（CG）. In addition,we proposed new gating techniques, called model probability weighed gating and two-stage model probability weighted gating, respectively. The latter takes possible model errors into account and uses an upper bound on the model error varianee when establishing the sencond gate. A benchmark trajectory are used. The gating techniques are compared in terms of RMS errors, percentage of track loss and computational load, Simulation results suggest thar for maneuvering target tracking in a cluttered enviroment using IMM-PDAF, gating via the TSMPWG method is preferable to other methods.