Optimization for DV-Hop type of localization scheme in wireless sensor networks

Abstract

Distance Vector Hop (DV-Hop) is a range-free scheme used for node localization in wireless sensor networks (WSNs). The original DV-Hop scheme localizes the unknown nodes depending on a number of anchor nodes’ known position information and the multi-hop relationship among nodes. It is very popular and can meet most application requirements as the network is isotropous. However, it becomes powerless while the network is anisotropic due to the natural defects of its ranging strategy. In view of such problems of the traditional DV-Hop, we provide a scheme aiming to improve the original DV-Hop. In our scheme, an improved cosine similarity parameter is used to measure the similarity between path pairs, and the anchor–anchor path which is most like the path from the unknown node to the target anchor is selected to compute the average hop distance of the node-anchor path independently. Then, an improved particle swarm optimization and simulated annealing hybrid algorithm is adapted to improve the position accuracy of the initial position of an unknown node, which has been derived by the trilateration algorithm used in the original DV-Hop scheme. Based on the simulation result, in comparison with the original DV-Hop scheme and another two existed improved schemes, our proposed scheme can perform much better both on the distance estimation accuracy and on the final node localization accuracy. Thereby, our proposed scheme is a feasible and optimized choice for node localization in WSNs.