论文部分内容阅读
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。
The problem of data association in multisensor multi-target tracking is the difficulty and core in the field of target tracking. If the sensor is a passive sensor with only angle measurement, the correlation problem becomes more complicated. Aimed at the multi-target tracking problem of purely azimuth multi-passive sensor system, a dynamic multi-dimensional distribution method based on Gaussian-Hermitian filter is proposed. Firstly, a Gaussian-Hermitian filter model of multi-passive sensors in Cartesian coordinates is established. Based on this model, the multi-dimensional distribution problem is used to establish a candidate association of angle measurement and target angle prediction. And dynamically allocate them to improve the association efficiency. Simulation results show that this method can solve the problem of data association in the multi-passive sensor system in real time and efficiently, and it can track the multi-target steadily.