采用目标假设的三被动式传感器数据关联算法

被动式传感器目标跟踪是多传感器多目标跟踪领域的一个重要研究方向。针对三被动式传感器多目标定位系统全局最优数据关联的三维分配问题，在允许传感器漏检和虚警的情况下，通过分析拉格朗日松弛算法，提出一种假定真实目标的快速收敛算法。该算法通过粗关联假定真实目标并重新修改代价矩阵，然后进行细关联，使得拉格朗日松弛算法在保证关联精度的前提下有效地提高了收敛速度。理论分析和实验结果表明，该算法提高了数据关联的速度,并在一定程度上提高了关联准确率。

A data association algorithm for three-passive-sensor based on target assumption

Passive sensor target tracking is one of the important research directions in multiple sensor multi-target tracking field.Aiming at the three-dimension assignment problem of data association for three-passive-sensor and multi-target location system,an improved Lagrangian relaxation algorithm is presented which can be rapidly converged on the condition that missed detection and false alarm are allowed. On the assumption that some targets are real ones,the cost matrix is modified,which contributes to the fast convergence of Lagrangian relaxation. Theoretical analysis and simulation results show that the modified algorithm improves the speed and accuracy of data association,especially in multiple target and/or heavy noise condition.