Semidefinite Optimization Approach to Estimate the Locations of Nodes in the Sensor Network with High Accuracy

该研究的目的是有效地估计N个传感器在r维空间中的位置，使用p个锚点的位置信息（通常p是很小的）和传感器和锚点的图的部分距离矩阵。我们提出了两种方法来划分传感器的集合：一种是使用每个传感器的度的方法，另一种是使用图中的团结构的方法。我们的数值经验报告，使用团结构的方法是更稳定的参数的变化。

The study aims to efficiently estimate the location of N censors in r-dimensional space using the information of the location of p anchors (usually p is very small) and the partial distance matrix of the graph of sensors and anchors. We propose two methods for dividing the set of sensors : a method using the degree of each sensor and a method using clique structures in the graph. Our numerical experiences report that the method using clique structures is more stable with respect to changes of parameters.