摘要
稀疏无线传感器网络由于缺乏足够的距离和连通性信息,导致大多数定位算法无法有效工作。提出了一种非度量多维标度移动节点辅助定位算法——NMDS-LRA(M)。该算法对移动节点运动轨迹抽样,添加拓扑约束关系,然后利用奇异值分解计算节点相异性矩阵的逼近阵,从而有效解决了移动辅助定位问题,并且避免了以往移动定位算法中虚拟节点间距离误差较大对定位精度的影响。仿真分析表明,与以往算法相比,提出的算法有更好的定位精度,而且在较低网络连通度和不规则网络分布的条件下表现出更好的可靠性。
In sparse sensor networks,most existing localization algorithms cannot work properly due to the lack of distance and connectivity information to uniquely localize sensors. A new nonmetric MDS mobile assisted location algorithm, NMDS-LRA(M), was presented to solve the problem. The algorithm samples the tracks of mobile nodes, and then computes the approximate matrix of dissimilarity matrix by use of singular value decomposition. In other algo- rithms, the high error of distance among virtue nodes deteriorate the localizatiton precision. NMDS-LRA(M) avoids the problem effectively. Simulation results demonstrate that the new algorithm can promote localization precision, and the most important, performs well on range error and anisotropie topology.
出处
《计算机科学》
CSCD
北大核心
2008年第10期219-222,235,共5页
Computer Science
基金
国家自然科学基金资助项目(60673061)
高等学校博士学科点专项科研基金资助项目(20060532024)
湖南省自然科学基金资助项目(06JJ50111
06JJ50113)
长沙市科技攻关项目(K069015-12)
关键词
无线传感器网络
定位
多维标度
矩阵近似
Wireless sensor networks, Localization, Multidimensional scaling, Matrix approximation