角度信息辅助的集中式多传感器多假设跟踪算法

对于密集杂波环境中的多目标跟踪，传统集中式多传感器多假设跟踪(CMS-MHT)算法在每一时刻的航迹-量测关联假设数量大大增加，导致数据关联不确定性增大，以至很难由常规航迹得分给出正确关联，表现为高的漏情率以及航迹分裂现象。基于传感器测量误差较小时虚警与目标量测的空间分布特点，针对多个相同类型传感器进行目标跟踪，该文提出一种角度信息辅助的CMS-MHT算法，设计了新的角度信息辅助的航迹得分计算方法，可以降低航迹-量测关联的不确定性，从而得到比传统CMS-MHT更优的关联假设。仿真实验结果表明，在密集杂波环境中，该算法能有效降低漏情率，并有更好的航迹完整性。

Angle Aided Centralized Multi-sensor MultipleHypothesis Tracking Method

For multi-target tracking in heavily cluttered environment, the number of measurement-to-track association hypotheses in each scan grows rapidly in traditional Centralized Multi-Sensor Multiple Hypothesis Tracking (CMS-MHT) method, which leads that the uncertainty of data association greatly increases such that correct association can hardly be given using traditional track score resulting in high leakage rate and effects of track splitting. Based on the space distribution characteristics of false alarm and target measurement when the sensor measurement error is small, for target tracking using multiple sensors of same type this paper proposes a new angle aided CMS-MHT method, which designs angle aided track score computation to reduce the uncertainty of measurement-to-track association. In such a way, the proposed angle aided CMS-MHT can provide better association hypotheses compared with traditional CMS-MHT. The experimental results illustrate that angle aided CMS-MHT reduces leakage rate and has better track integrity in heavily cluttered environment.