The Research On Range-free Localization Algorithms For Wireless Sensor Networks

Wireless sensor networks (WSN) have been applied to many areas such as military affairs, environmental monitoring, traffic management, anti-disaster and industrial field etc. As one of the key backstop technologies and research hotspots, the nodes localization has great significances due to its direct correlation with theoretical study and practical application. Because of constraints in power, and cost of sensor nodes, the localization algorithms should be low energy consumption, low complexity and can achieve high accuracy.With regard to the mechanisms used for estimating location, these approaches can be divided into two categories: Range-based and Range-free approaches. Range-based schemes typically can achieve high location accuracy, but require additional hardware on sensor nodes and high energy consumption, which are impractical for resource-limited WSNs. Range-free schemes do not need the distance or angle information for localization. They just utilize network connections to calculate the approximate position of nodes. Compared to Range-based schemes, Range-free schemes have significant advantages in energy consumption and cost and received extensive attention in WSNs.This thesis first briefly reviewed the typical localization algorithms of Range-free schemes. And we mainly research on the range-free localization algorithm with mobile anchor nodes. Due to the advantages in cost, flexibility and applications in complex industrial environment, mobile anchor-based localization is a hot research area now.First of all, we propose a RSS-based localization algorithm with mobile anchor nodes, which not only reduces the number of anchor nodes but also improves the localization accuracy.Secondly,the thesis mainly analyzes the novel Range-free localization algorithm APIT which is based on lapped area and DV-Loc algorithm which is based on Voronoi Diagram and their drawbacks, and put forward a new algorithm VBL. It utilizes mobile anchors to build Voronoi Diagram and fix the unknown nodes in a certain Voronoi polygon, then it narrows down the area with more Voronoi polygons to localize the unknown nodes. We do some simulations, which show that it has less communication overhead, compute time and high localization accuracy.Mobile sensor network is another hot research area now. We analyze a proposed localization algorithm EMAP, which is suitable for a special mobile sensor network. Based on the analyses of its drawbacks in applications, we propose a novel localization algorithm based on geometric constraints. This algorithm does not require complex hardware and computation. Via extensive simulations, we find it can reduce the energy consumption and improves the localization accuracy and is more suitable for practical application than EMAP.