基于GM-PHD的低可观测编队目标跟踪方法

ET-GM-PHD Formation Target Tracking Method Under Low Obervability Condition

考虑到多编队在低可观测情况下存在的目标跟踪问题,提出了一种基于高斯混合概率假设密度(GM-PHD)滤波算法的编队目标跟踪方法。该方法是在修剪融合过程中,先保留剪掉的高斯分量,再对这些分量进行状态外推,利用JS(Jensen-Shannon)散度判断下一时刻状态估计值与外推状态值是否相似,以判断结果体现目标丢失情况,使得真实目标不丢失,解决了低可观测情况下目标易漏检带来的跟踪性能下降问题。然后,利用编队目标的特点,结合密度聚类方法估计出编队整体的状态,避免因状态估计集合中状态值过多影响算法性能。最终,仿真实验结果表明,该方法可以在低可观测情况下有效跟踪编队目标,具有较好的跟踪性能。

Considering the target tracking problem of multiple formations under low observability,a formation target tracking method based on Gaussian Mixture Probability Hypothesis Density(GM-PHD)filtering algorithm is proposed.In this method,during the pruning fusion process,the cut Gaussian components are first retained,and then these components are extrapolated in state.The JS(Jensen-Shannon)divergence is used to determine whether the next state estimate value is similar to the extrapolated state value.The judgment result reflects the loss condition of the targets,so that the real target is not lost,and the problem of tracking performance degradation caused by easy miss detection of the targets under low observability conditions is solved.Then,using the characteristics of the formation target,combined with the density clustering method is combined to estimate the overall state of the formation,and to avoid affecting the performance of the algorithm by too many state values in the state estimation aggregation.Finally,the results of simulation experiments show that this method can effectively track formation targets under low observability conditions and its tracking performance is good.