基于一种改进IMMJPDA算法的地面目标跟踪

对地面多目标的跟踪,由于地面目标的高机动性、杂波密集特点以及运动不确定性,交互式多模型联合概率数据关联无疑是一种好的跟踪算法,但是该算法需对与可行联合事件相对应的矩阵进行拆分,随着目标个数的增多,计算量会呈指数增长。为此提出一种基于模糊多门限的交互式多模型联合概率数据关联算法,该算法利用量测与目标的关联概率来替代可行联合事件概率的计算。Monte Carlo仿真结果显示了该算法在现实运动中的可行性和方便性。该算法减少了计算量,又改善性能,利用多模型特点解决了地面目标的高机动性所带来的运动模型匹配问题。

Ground target tracking based on an improved IMMJPDA algorithm

It is no doubt that Interactive Multi-Model Joint Probability Data Association（IMMJPDA） is a better way to track multi-target on ground due to the high maneuverability, dense clutter and movement uncertainty of the ground targets. However, the algorithm needs to split the matrix corresponding to the feasible joint events, and the calculation amount grows exponentially with the increase of targets. This paper presents an IMMJPDA algorithm based on fuzzy and multi-gate limit, which reduces the calculation amount and improves the performance by using the associated probability of measurement and target for calculation instead of the feasible joint events probability. The motion model matching problem owing to the high maneuverability of ground targets is solved by using the muhi-model characteristic. The results of Monte Carlo simulation show the algorithm is effective and convenient for the actual movements.