分布式多维标度定位算法的研究  

Research on Distribution Multidimensional Scaling Localization Algorithm

作　　者：王新生[1] 胡玉兰[1] 

机构地区：[1]燕山大学信息科学与工程学院,秦皇岛066004

出　　处：《计算机科学》2012年第2期80-83,87,共5页Computer Science

基　　金：信息系统协同性随机模型与可生存性复杂分析(973项目)子课题(2005CB321902)资助

摘　　要：针对经典MDS-MAP定位算法在定位精度和算法复杂度方面的不足,提出一种分布式多维标度定位算法。改进后的算法加入了分簇的思想,将大规模网络分成多个具有簇首的局部网络。局部定位时,引入Hop-Euclidean算法,计算簇内节点间距离,再用局部网络融合算法将局部相对坐标图合并成全局相对坐标图。仿真分析表明,提出的算法在各向同性和各向异性网络中都有很好的定位精度,而且在定位精度提高的情况下可用于不规则网络,有利于网络的扩展,更适用于大规模密集型网络。A distributional localization algorithm based on multidimensional sealing（MDS）technique was proposed aiming at the shortages of the classic MDS-MAP algorithm in localization precision and complication of computing.The improved clustering method was used to build different clusters for local positioning,and in the local positioning,Hop-Euclidean algorithm was used to calculate the distance matrix.Then the local maps were combined to a global relative coordinate map based on the aggregation algorithm from local network to global network.Simulation results demonstrate that the new algorithm can promote localization precision and perform well under anisotropic topology or large-scale intensive wireless sensor network.

关 键 词：无线传感器网络 多维标度 Hop-Euclidean算法 分布式定位算法 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]