基于分段动态时间序列的异步航迹匹配

Asynchronous Track Matching Based on Piecewise Dynamic Time Series

研究了分布式目标跟踪系统中的航迹匹配,主要解决的是各传感器产生的异步局部航迹之间高效、准确地匹配问题。目前,主要的解决方法是进行时间配准,然而该方法会增大估计误差,影响后续的航迹关联和融合效果。在此情况下,提出了采用基于数值-导数的动态时间规整(Value-Derivative Dynamic Time Warping,VDDTW)和分段线性近似(Piecewise linear approximation,PLA)相结合的方法进行异步航迹匹配。VDDTW方法允许不同长度的航迹序列的匹配,且鲁棒性好;PLA方法将航迹序列分割为一系列子序列,降低计算成本,支持在线计算。仿真结果表明,该方法能快速、有效地解决异步航迹关联问题。

In this paper,the track matching in distributed target tracking system is studied.The main objective is to solve the problem of efficient and accurate matching between the asynchronous local tracks produced by each sensor.At present,the main solution is to make time registration,but this method will increase the estimation error and influence the subsequent effect of association and fusion.In view of this situation,this paper proposes an asynchronous track matching method based on VDDTW and PLA.The VDDTW allows to match the track sequences with different lengths,and it has good robustness.