无线传感器网络动态轨迹多目标跟踪算法  

Multi-Object Dynamic Trajectory Tracking in Wireless Sensor Network

作　　者：王瑞锦[1] 郭祥[1] 王佳昊[1] 秦志光[1] 

机构地区：[1]电子科技大学信息与软件工程学院,成都610054

出　　处：《电子科技大学学报》2016年第2期233-239,共7页Journal of University of Electronic Science and Technology of China

基　　金：国家自然科学基金(60903157);四川省科技厅计划项目(2015JY0178);中央高校基本科研业务费专项资金(ZYGX2014J051;ZYGX2011J066)

摘　　要：为以较低的能耗高效地解决多目标区域轨迹跟踪问题,并有效地延长网络生存期,该文提出了一种新的基于边权值动态轨迹的多目标区域跟踪算法EWDT。该算法首先通过对任意指定的一块区域进行边权值计算,对该区域中目标的个数、身份及状态等关键信息进行分析和确认;其次设计了节点状态转换机制,并利用动态跟踪簇以及卡尔曼滤波技术对目标进行轨迹跟踪,以进一步降低系统能耗与时延。实验结果表明,EWDT算法与目前的Forms和LLS方法相比,明显降低了系统能耗与时延,具有较强的实用性。For solving multi-objective trajectory tracking problem by lower energy consumption costs and extending the network lifetime in wireless sensor networks, a new multi-target area tracking algorithm（EWDT） is presented based on edge weights dynamic trajectory. EWDT analyzes and confirms the number, identity and status of the target by calculating the value of edge weights in the any given area. To further reduce the energy consumption and latency, EWDT designs the node state transition mechanism and utilizes dynamic tracking clustering and Kaman filtering techniques to track the target trajectory. Experimental results show that EWDT can reduces the system power consumption and latency significantly comparing with the current Forms and locality aware location service（LLS） algorithm.

关 键 词：动态轨迹 边权值 目标跟踪 无线传感网络 

分 类 号：TP309.3[自动化与计算机技术—计算机系统结构]