项目名称: 无线传感器网络节点定位的矩阵完成理论与方法
项目编号: No.61201273
项目类型: 青年科学基金项目
立项/批准年度: 2013
项目学科: 电子学与信息系统
项目作者: 陈章鑫
作者单位: 电子科技大学
项目金额: 27万元
中文摘要: 节点位置是移动无线传感器网络(WSN)的重要信息之一,对WSN信息获取与探测具有重要意义。理论上,多维标度(MDS)定位方法能利用WSN节点间的距离矩阵估计节点的位置,但是复杂的应用环境、受限的功率、信号带宽和成本等因素往往导致距离矩阵出现不完整性问题。这种由测量数据导致的信息损失不仅增加了节点位置估计问题的欠定程度,还增加了MDS定位算法的复杂度。将矩阵完成分析理论应用于WSN节点定位,开展矩阵元素的测度完整性分析、矩阵的显式完成分析、矩阵的隐式完成分析以及计算与定位精度折衷分析等方面的研究,克服因测量数据的信息损失和复杂应用环境的影响,提高无线传感器网络节点定位算法的稳健性和实用性。
中文关键词: 无线传感器网络;矩阵完成;低秩矩阵;欧氏距离矩阵;凸优化
英文摘要: Node positioning is one of the important information in wireless sensor networks (WSNs), which has great significance for information acquisition and target detection. Multidimensional scaling (MDS) algorithms can be used for node positioning by constructing a pair-wise squared distance matrix, however, some factors such as complex application conditions, limited power, bandwidth and cost could all lead to the imperfection of the distance matrix. This information loss from measurement not only augments the underdetermined degree for node position solution, but also increases the complexity of MDS algorithm. We can recast this problem into a matrix completion problem and then develop the node positioning methods via the analysises of measure completion, explicit matrix completion, implicit matrix completion and computation-accuracy tradeoff. These efforts would reduce the effects of information loss and complicated environments and improve the robustness and practicability of node positioning methods.
英文关键词: Wireless sensor networks;Matrix completion;Low-rank matrix;Euclidean distance matrix;Convex optimization