基于粒子滤波的无线传感器网络目标跟踪算法  

Tracking algorithms based on particle filter for wireless sensor networks

作　　者：黄艳[1,2,3] 梁韦华[1,2] 于海斌[1,2] 

机构地区：[1]中国科学院沈阳自动化研究所 [2]中国科学院工业信息学重点实验室,沈阳110016 [3]中国科学院研究生院,北京100049

出　　处：《控制与决策》2008年第12期1389-1394,共6页Control and Decision

基　　金：国家自然科学基金项目(60434030,60704046,60725312);辽宁省青年人才基金项目(3040004)

摘　　要：传感器节点的组织和路由对无线传感器网络(WSN)目标跟踪算法的性能有重大影响.为此,针对具有簇-树型网络拓扑结构的WSN,首先给出集中式粒子滤波跟踪算法(CPFTA)实现的具体步骤;然后提出一种分布式粒子滤波跟踪算法(DPFTA),构建性能评价体系,通过仿真实验给出两种跟踪算法的定量比较,结果表明DPFTA的跟踪精度稍低于CPFTA,但能大幅度减少通信开销,而且具有更小的跟踪反应时间;最后仿真分析了传感器覆盖密度和检测阈值对跟踪算法性能的影响.Performance of the tracking algorithm for wireless sensor networks （WSN） is significantly affected by the organization of sensor nodes and routing. For a WSN with cluster-tree topology, this paper presents the implementation details of centralized particle filter based tracking algorithm （CPFTA）, and then proposes a distributed particle filter based tracking algorithm （DPFTA）. A performance evaluation system is established： Two tracking algorithms are quantitatively compared by simulation. The results indicate that DPFTA can reduce the communication cost and tracking response time significantly at only a slight expense of tracking performance. Finally, the relationship between node density, detection threshold and the performance of tracking algorithms is studied by simulation.

关 键 词：无线传感器网络 粒子滤波 目标跟踪 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]