一种耦合检测和JPDA滤波的多目标跟踪算法

传统雷达信号处理中对目标的检测和跟踪是割裂处理的,通常为先检测后跟踪(DBT);当目标的信噪比较低时,检测过程中将出现大量的虚警及漏检,使得后继的跟踪算法失效。针对这一问题,在联合处理检测和跟踪方法的基础上,提出了一种耦合贝叶斯检测和联合概率数据关联(JPDA)滤波的多目标跟踪算法(JPDAF-BD)。JPDA滤波器将目标的位置分布信息反馈到贝叶斯检测器,继而贝叶斯检测器将该反馈作为先验信息用于检测判决。仿真结果表明,所提出的JPDAF-BD算法较之传统DBT体制下的多目标跟踪算法(JPDAF-NP)有显著的性能提升,可以实现更低信噪比下的多目标检测和跟踪。

Integration of Detection with JPDAF for Multi-Target Tracking

Target detection and tracking in conventional radar signal processing are treated separately, namely detection before tracking（DBT）.When the signal to noise ratio is low,the tracking algorithm will not work because of too many false alarms and missed alarms in the detection.In view of this problem,the method of coupling detection and tracking is studied,and an integration of Bayesian detection with JPDAF （JPDAF-BD）algorithm is proposed.JPDA filter provides the posterior distribution of the targets location to the detector,then the Bayesian detector utilizes the feed-back as prior information.Simulation results show that the proposed JPDAF-BD can detect and track multi-target successfully in lower SNR in comparison with the traditional multi-target tracking algorithm（JPDAF-NP）,which obtains considerable performance improvement.