Connectivity-Based Self-Localization in WSNs

Abstract

Efficient localization methods are among the major challenges in wireless sensor networks today. In this paper, we present our so-called connectivity based approach i.e, based on local connectivity information, to tackle this problem. At first the method fragments the network into larger groups labeled as packs. Based on the mutual connectivity relations with their surrounding packs, we identify border nodes as well as the central node. As this first approach requires some a-priori knowledge on the network topology, we also present a novel segment-based fragmentation method to estimate the central pack of the network as well as detecting so-called corner packs without any a-priori knowledge. Based on these detected points, the network is fragmented into a set of even larger elements, so-called segments built on top of the packs, supporting even more localization information as they all reach the central node.