Distributed interacted multisensor joint probabilistic data association algorithm based on D-S theory

In order to resolve the multisensor multiplied maneuvering target tracking problem, this paper presents a distributed interacted multiple model multisensor joint probabilistic data association algorithm （DIMM-MSJPDA）. First of all, the interacted multiple model joint probabilistic data association algorithm is applied to each sensor, and then the state estimation, estimation covariance, model probability, combined innovation, innovation covariance are delivered to the fusion center. Then, the tracks from each sensor are correlated and the D-S evidence theory is used to gain the model probability of an identical target. Finally, the ultimate state estimation of each target is calculated according to the new model probability, and the state estimation is transmitted to each sensor. Simulations are designed to test the tracking performance of DIMM-MSJPDA algorithm. The results show that the use of DIMM-MSJPDA algorithm enables the distributed multisensor system to track multiplied maneuvering targets and its tracking performance is much better than that of IMMJPDA algorithm.

集中交互式多传感器模糊联合概率数据互联算法

为了解决杂波环境下多传感器多机动目标跟踪问题,本文提出了一种集中交互式多传感器模糊联合概率数据互联算法.该算法首先设定多个跟踪模型并计算每个模型中测量点迹与航迹测量预测之间的模糊综合相似度;然后基于阈值判别及经验概率法则计算模糊联合互联概率,并基于此概率对各航迹进行状态估计及状态估计协方差的更新;最后计算各模型概率,并据此概率对各模型所获得状态估计进行加权得出各航迹在融合中心最终的状态估计.对该算法与集中交互式多传感器联合概率数据互联算法进行仿真比较,仿真结果显示该算法的跟踪精度较后者有显著提高,同时有效地降低了周期耗时,综合性能更优越.

集中交互式多传感器联合概率数据互联算法

为了解决杂波环境下多传感器多机动目标跟踪问题,本文提出了一种集中交互式多传感器联合概率数据互联算法。本文提出的算法首先应用广义S-D分配的规则对每个传感器送来的观测数据进行排列组合,并对所有的测量组合进行有效性判断,然后应用数据压缩的方法将每个有效量测组合压缩成一个等效量测点并根据每个等效量测点的联合似然函数计算其联合互联概率,最后在此基础上应用交互式多模型算法的思想以处理目标出现机动的问题。本文最后给出了该算法的分析,仿真结果表明,本文算法能够很好地解决杂波环境下多传感器多机动目标的跟踪问题。

多传感器多机动目标跟踪方法研究进展

讨论了多传感器多机动目标的研究现状,分析了几种跟踪滤波器的性能,着重探讨了交互多模型(IMM)算法的发展状况和应用。同时还研究了多传感器跟踪多机动目标过程中的数据关联方法。最后指出,寻求采用自适应变结构的IMM算法,研究模型切换过程中算法的过渡性能,以及在数据关联中引入混沌、遗传算法等智能控制方法,将是未来多传感器多机动目标研究的方向。