Research On Localization Algorithm In Wireless Sensor Networks

The positioning technologies existed for the WSNs can be divided into two categories:the range-based and range-free. The APS is a series of distributed positioning algorithm and it is attracting lots of attention in the range-free, which is proposed under distance vector routing by Niculescu. In APS, Euclidean presents the accurate method for the calculation of the position of the unknown node which is two hops from the anchor node. Euclidean makes the breakthrough at the limitation that the unknown node can be located when it is adjacent with the anchor node. So Euclidean is more accurate than DV-Hop which need average distance estimation of every hop. But the deficiency existing in Euclidean:the theory of the plane geometry in Euclidean cannot directly extent to three dimensions so that it doesn’t apply to the node location of three dimensions. So a3D position algorithm based on Euclidean for Wireless Sensor Networks is presented. The problem of the distance calculation between an unknown node and an anchor node is transformed into a solution of the distance between the vertices in Hexahedral. And the distance can be solved by the method of coordinates presented. At last the paper adopts the quadrilateral localization method to get the coordinates. The distance between the unknown node and the anchor node can be solved exactly by the three dimensional model and the method of coordinates to improve the positioning accuracy and reduce the calculated amount.Moreover, the grade of membership functions of the fuzzy theory is introduced to the positioning field to deal with the uncertainty of Wireless Sensor Networks. The grade of membership functions indicates the degree between the reference nodes (or sample nodes) and the unknown nodes. On the basis, the information fusion is introduced to the positioning algorithm to further improve the accuracy of position. The judgment information by multiple anchor nodes is fused to obtain reference nodes with the largest degree of proximity. And the erroneous reference node is removed using clustering method to improve the positioning accuracy. The position coordinate can be deduced by repeated iterative operations. Simulation results also verify the feasibility and effectiveness of the presented algorithm.