Location Estimation based on Edge Weights in Wireless Sensor Networks

Localization in wireless sensor networks is to determine the positions of all nodes based on the Down positions of several nodes. Much previous work for localization use multilateration or triangulation based on measurement of angles or distances to the fixed nodes. In this paper, we propose a new centralized algorithm for localization using weights of adjacent nodes. The algorithm, having the advantage of simplicity, shows that the localization problem can be formulated to a linear matrix equalities. We mathematically show that the equalities have a unique solution. The unique solution indicates the locations of unknown nodes are capable of being uniquely determined. Three kinds of weights proposed for practical use are compared in simulation analysis.