纯方位被动多传感器多目标跟踪算法  

Algorithm of bearings-only passive multi-sensor and multi-target tracking

作　　者：李彬彬[1] 冯新喜[1] 李鸿艳[1] 宁宣杰[2] 

机构地区：[1]空军工程大学电讯工程学院,陕西西安710077 [2]东北大学信息科学与工程学院,辽宁沈阳110004

出　　处：《红外与激光工程》2012年第5期1374-1378,共5页Infrared and Laser Engineering

基　　金：信息综合控制国家重点实验室基金项目(2010科技字第46号)

摘　　要：多传感器多目标跟踪中的数据关联问题是目标跟踪领域中的难点及核心。若传感器是只有角度量测的被动传感器,关联问题则变得更为复杂。针对纯方位多被动传感器系统的多目标跟踪问题,提出了一种基于高斯-厄密特滤波的动态多维分配方法。首先建立了直角坐标系下多被动传感器的高斯-厄密特滤波模型;在该模型的基础上,采用多维分配问题的思想,直接建立各传感器角度量测与目标角度预测值的候选关联组合,并将其进行动态地分配,提高了关联效率。仿真实验表明,该方法可以实时、高效地解决多被动传感器系统中的数据关联问题,并且能够对多目标进行稳定的跟踪。Data association for multi-sensor and multi-target tracking is a difficult and key issue in target tracking. If the sensor can only obtain the angle measurements from target, data association becomes more complex. For the problem of multi-target tracking in the system of beatings-only multi-sensors, a dynamic multi-dimensional assignment algorithm based on Gauss-Hermite filter was presented. Firstly, the model of Gauss-Hermite filter for passive multi-sensor in Cartesian coordinate was established. On the basis of the model, candidate association with angle measurements from each sensor and the predicted angle of target was directly established adopting the multi-dimensional assignment algorithm, and then the candidate association was dynamically assigned so that the efficiency show that the presented method can solve the problem of effectively, and also track multi-target steadily. of association was improved. Simulation results data association in passive multi-sensor system

关 键 词：被动多传感器多目标跟踪 高斯-厄密特滤波 多维分配 

分 类 号：TN953[电子电信—信号与信息处理]