Exploiting the distribution of distances between nodes to efficiently solve the localization problem in wireless sensor networks

The localization problem is a central one in the field of wireless sensor networks, but despite the ingenuity of previous attempts at solution, developing an efficient and accurate localization method has remained essentially an open problem. This is so, in our view, because there has been a lack of theoretical results to support more solid advances. In this paper, we build on the semi-analytical approach we developed elsewhere, by exploiting its results regarding the relationship between Euclidean distances and hop distances in wireless sensor networks to improve on one of the most robust localization methods we know of, the well-known DV-hop. Our improvements result in better accuracy, reduced communication costs and local processing, as well as time and energy savings. The results we present are based on simulations through which we compare the new method with the original DV-hop. In addition to the new, more efficient localization method, we also provide a new perspective on the distance relationships, and thereby also contribute to furthering the establishment of a better theoretical understanding of the subject.