无迹卡尔曼滤波在无线传感器网络节点定位中的应用  

Application of Unscented Kalman Filter Algorithm to Node Localization in Wireless Sensor Networks

作　　者：杜娟娟[1] 

机构地区：[1]南京邮电大学宽带无线通信和传感网技术教育部重点实验室,江苏南京210003

出　　处：《南京邮电大学学报（自然科学版）》2013年第1期84-90,共7页Journal of Nanjing University of Posts and Telecommunications：Natural Science Edition

基　　金：国家重点基础研究发展计划(973计划)(2011CB302903)资助项目

摘　　要：传统的基于距离的无线传感器网络节点定位技术,由于测距过程产生较大的误差,从而使定位精度不高。文中在基于接收信号强度(RSSI)测距、三边测量法初始定位的基础上,提出以接收信号强度为观测量,将无迹卡尔曼滤波(UKF)算法应用到节点精确定位中。通过仿真验证使用该方法后,相比以距离为观测量的UKF定位方法,节点的定位精度有一定的提高,并进一步定量的分析比较了两种实现模型下节点定位算法的误差概率分布。在此算法的基础上,通过权衡平均定位误差与算法运算复杂度之间的关系,给出最佳定位锚节点数量,并模拟具体环境,验证了文中节点定位算法的实用性。Because of the large errors associated with the process range-based node localization technique is with low precision in the received signal strength （RSSI） to measure the distance and of distance measurement, the traditional wireless sensor networks （WSN）. using three-edge measurement to provide the initial localization, the paper proposes to apply Unscented Kalman Filter （UKF） algorithm to the precise node locating with RSSI as the observed quantity. The emulation has shown that the node localization ac- curacy is improved compared to the UKF localization method with distance as observed quantity. A quanti- tative comparative analysis is made of the error probability distribution between these two localization algo- rithms. By a tradeoff between the average localization error and computational complexity, the optimal number of the anchor nodes for localization is determined. Finally the simulation under concrete environ- ment has verified the practicality of the proposed localization algorithm.

关 键 词：无线传感器网络 RSSI UKF 节点定位 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]