推广的Dempster合成规则在航迹关联中的应用

Application of the extended Dempster combination rule in track-to-track correlation

目前已有的航迹关联方法虽然能很好地解决二传感器二目标或者二传感器多目标航迹融合问题,但处理多传感器多目标航迹关联问题时效果较差.因航迹关联问题实际上是判决融合问题,作者提出用Dempster合成规则的思想来解决之,但必须要对合成规则有新的推广.推广后的方法不仅能解决二传感器航迹关联问题,而且还能解决多传感器航迹关联问题.当传感器证据相互冲突或不够准确的时候,经典的Dempster合成规则无法应用,而推广的Dempster合成规则可解决该问题.最后作者给出了一个四传感器二航迹关联的例子,来解释和说明推广的Dempster合成规则在航迹关联中的应用.计算机仿真结果说明推广的Dempster合成规则比经典的Dempster合成规则更适用.

At present, the track correlation methods perform well in solving two sensors two targets or two sensors more targets track fusion, but work badly in multi-sensor multi-target track correlation. Since the track correlation problem is actually a decision fusion issue, Dempster combination rule can be used to solve such track correlation problem. However, to be suitable to solve the track correlation problem, new extension to the classical Dempster combination rules is necessary. This extended method can solve not only the problem of two sensors track correlation, but also the problem of multi-sensor track correlation. In this paper, when the sensor evidences are conflicted mutually or not quite accurate, the authors propose the extended Dempster combination rule to handle the above issue. Finally, a exampie of 4 sensors 2 target trajectories correlation is given to explain the application of the extended Dempster rule in track correlation and computer simulation results show that the extended Despmster rule performs better than the classical Despmster rule does.