Abstract
In this paper, we propose a hybrid approach to the wireless sensor network (WSN) localization problem. The proposed approach harnesses the strengths of two techniques: RF mapping and cooperative ranging, to overcome the potential weaknesses in one another. The idea is to first allow every node to obtain an initial estimate of its own position in a neighbor-independent way using a coarse-grained RF map acquired with minimal efforts. Then each node iteratively refines its own position through distance ranging to each of its neighbors, irregardless of their positions with respect to itself. Through simulation performance experiments, we show the potential of this hybrid approach as a practical localization system for WSN that can achieve reasonable localization accuracy without significant deployment efforts. ©2009 IEEE.