一种基于粒子滤波的分布式跟踪算法

A Decentralized Tracking Algorithm Based on Particle Filtering

在多平台协同作战中,由于通信时间的延迟,常常出现各传感器量测不能按照正常时序发送到数据融合中心处理的情形.为解决协同作战系统无序量测的滤波问题,同时提高其目标跟踪的性能和降低系统的算法存储量,提出了一种Unscented粒子滤波算法.该算法估计精度高、计算量和存储量较小、输出无延迟.仿真结果表明,该算法同其他滤波算法相比,提高了跟踪精度,较好地解决了时间延迟的问题.

In multiplatform cooperative engagement,sensor measurements cannot usually arrive at the fusion center according to detection time because of the delay for communicating time.To solve an outof-sequence measurements filtering problem,to improve the tracking performance and to reduce the computation cost,an unscented particle filtering algorithm was developed.The filter has higher estimation accuracy,fewer computation costs,fewer memory requirements,and no delayed output.Simulation results show that the algorithm has better tracking accuracy than other filter algorithms,which solves the problem to the time delaying.