基于空间分布特征的阵群目标数据关联算法

Group Target Data Association Algorithm Based on Space Distribution Feature

在低数据率条件下,对监视区域内的每个目标进行有效分析是非常困难的。针对该问题,将阵群目标作为研究对象,并给出了一种基于空间分布特征的阵群目标数据关联算法。首先,在近邻点集聚合的基础上,从传感器给出的单目标观测集合中提取出阵群目标观测(简称阵群观测);其次,按照自顶向下的方式逐次计算各个阵群观测不同子集空间分布的距离度量,并以此为基础计算阵群观测之间的关联度量;最后,通过在关联代价矩阵上应用二维分配算法得到不同时刻各个阵群观测之间的对应关系。仿真结果证实了该算法的有效性。

It is difficult to analyze every target effectively when the measurement data rate is relatively low. Aiming at the problem, a scheme was developed which regarded the group targets as the study object and a group target data association algorithm was proposed which took full advantage of their space distribution features. Firstly, group targets measurements （group measurements for short） were extracted based on the combination of the near neighbour point sets from the measurements set of individual member targets. Secondly, the association metric of two group targets’ measurements was calculated based on the distance metric between the subsets of the two measurements set that was implemented in a top-down manner. And finally, the correspondence relations between the group targets’ measurements obtained at different moment were found after the 2D assignment algorithm was applied to the association cost matrix. The simulation has proved the effectiveness of the algorithm.