综合约束信息下的被动多传感器多目标跟踪  

Passive Multi-Sensor Multi-target Tracking Based on Comprehensive Constraint Information

作　　者：陈昊[1] 侯睿 林志祥 程悦 Chen Hao;Hou Rui;Lin Zhixiang;Cheng Yue(The 28th Research Institute of China Electronics Technology Group Corporation,Nanjing 210007,China;One Center of the Equipment Development Department,Beijing 100032,China)

机构地区：[1]中国电子科技集团公司第28研究所,南京210007 [2]军委装备发展部某中心,北京100032

出　　处：《信息化研究》2022年第4期10-18,共9页INFORMATIZATION RESEARCH

摘　　要：地理信息(航路、海路、道路、空中走廊、禁飞区、禁航区、机场等)以及目标编队的空间相对关系表征了目标运动的等式和不等式约束。建立约束信息与传感器信息的关联和融合是提升目标检测与跟踪性能的重要途径。本文提出了综合利用约束信息和被动多传感器信息的概率假设密度(PHD)目标跟踪框架,给出了基于边界约束的被动多传感器箱粒子PHD算法实现。该方法利用先验已知的各类约束缩小新生目标的搜索范围和撒点区域,有利于减少无效探测和计算;跟踪维持过程中利用约束信息的投影可以进一步提升跟踪精度。仿真验证该算法在跟踪性能相当的情况下显著降低了计算量。The geographic information(air route, sea route, air corridor, prohibited area, airport and etc.) and the spatial relative relation of the aircraft formation represent the equality and inequality constraints of the targets. The establishment of association and fusion between constraint information and passive multi-sensor is an important approach to improve the performance of target detection and tracking. Firstly, the probability hypothesis density filter(PHD) target tracking based on the information of the constraints and passive multi-sensor is presented, then the algorithm implementation of the passive multi-sensor box particle PHD based on the boundary constraint is proposed. This algorithm utilizes the priori known constraints to narrow the birth targets searching and sampling region, which in favor of reducing invalid detections and calculations. The utilization of constraints information projection can further improve the tracking performance. The simulation results show that the proposed algorithm remarkably reduce the calculation with comparative tracking performance.

关 键 词：约束信息 概率假设密度 多目标跟踪 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]