Weighted localization for underwater sensor networks

Abstract

ELSN (Efficient Localization for large-scale underwater Sensor Networks) gets the node location by solving equations. However, sometimes the unique solution (no solution or multiple solutions) cannot be obtained from equations because of measurement error. In this case, the node cannot be located. On the basis of ELSN algorithm, a new localization algorithm is proposed. It transforms the process of solving equations into looking for the best point of intersection of the three circles in the plane, and regards the point as the potential location of the node. First, in this algorithm, the node is projected into a two-dimensional plane to reduce the computational complexity of the algorithm. Second, every three reference nodes are selected as a combination at random. That is to say, each triplet of nodes which represent a triplet of equations forms three tangent or intersecting circles. Based on the positional relationship of three circles, an optimal point of intersection is served as a potential location of the target node, and the residuals of each triplet are served as a potential weight. Finally, the weighted results of all potential locations are considered as the final position of the node. © 2014 Springer-Verlag Berlin Heidelberg.