Research On The Three-dimensional Range-free Localization Algorithm For WSN

Wireless sensor network(WSN), which consists of a large number of simple and inexpensive sensor device, equipped with wireless communication interface, has been proposed for various applications such as environment monitoring, disaster relief, target tracking, medical care, and so on. In these applications, it is important for the sensor nodes to be aware of their own locations, because sensed data is always meaningless without relating to its physical position. Furthermore, many protocols of WSN, for example, the location aided routing protocols, will likely depend on the location aware sensors. Therefore, localization has become a critical problem in WSN and also attracted massive attentions from both industry and academy.A brief introduction of WSN is given to describe the structure and characteristics of WSN. Then the localization algorithm is explained and investigated particularly. Localization algorithms can be divided into two types:range-based methods and range-free methods. Because of the advantages on power and cost on sensor node, range-free methods are stressed in this paper. The most representative range-free methods such as Centroid algorithm, Convex optimization, DV Hop, APIT, ROCRSSI, are discussed, a brief summary of advantages and disadvantages is given, and also the existing problems are concluded. But these algorithms based on two-dimensional (2D) plane are difficult to be expanded to three-dimensional (3D) application. However, sensor networks may often be deployed over complex 3D terrains in real applications. For example, an environmental monitoring network deployed in the sea to detect the temperature and salinity of sea water, the ocean currents and tide, and so forth. In these networks, network topologies could be much more complex than 2D cases, which requires the localization schemes to be more robust to network irregularities.Our work is primarily focused on the 3D localization of WSN. The existing 3D localization algorithm can be classified into two sorts:one solution is based on information exchange with anchors, and the position is calculated by methods such as triangulation. The other is depended upon dynamic message broadcasting of infrastructure. For the latter, success of localization is entirely determined by the infrastructure, which leads to security threat. Thus we concentrate on the first solution. According to the shortage of algorithms proposed for 3D up to now, based on the grid-based representation of the sensing space and the voting process, we present a novel distributed range-free localization algorithm for 3D(3D-DRL). Based on the grid-based representation, the location area is divided into cubic grid and anchors will vote for each cubic cell.3D-DRL assumes the centroid of cells with highest votes as the estimated location of unknown nodes. The scheme avoids the interaction between the blind nodes, costs lower communication overhead, is independent of the anchor node density and robust to the network topology.Finally, a simulation model is created based on MATLAB, and a detailed performance analysis of 3D-DRL is made, and simulation results indicate that even under the irregular radio propagation model, the scheme also performs well in terms of low complexity, high location accuracy and 100 percent location ratio. Thus,3D-DRL provides an effective method to solve the 3D localization of WSN.