Research On Dynamic Clustering Algorithm For Target Tracking In Wireless Sensor Networks

Wireless Sensor Network(WSN), which integrate the sensor, micro electro mechanical systems(MEMS), distributed information processing and wireless communication. As a result of its inherent advantages, such as self-organization, wide coverage, high fault-tolerance and high accuracy measurements, as well as the features of low cost, low power, flexible deployment etc, WSN are applied ranging from the military areas to the civil areas. Tracking and localization of moving target based on Wireless Sensor Network is a hotspot information technology. One of the research kernels is how to use the network energy efficiently while promising the tracking accuracy of moving target.This paper deeply studied the localization and target tracking problems in Wireless Sensor Network. The important research results are as follows:①Conventional DV-Hop localization algorithm considers the average hop distance as the expected distance, so it will lead to larger error in location. In response to this problem, an improved DV-Hop localization algorithm is proposed. In improved algorithm, anchor nodes correct the average hop distance by the total error between calculated distances and estimated distances. Improving the method of nodes coordinates computing, we will not adopt traditional triangulation algorithm but use 2-D hyperbolic location algorithm .Then ,we will get the ultimate node coordinates which includes error correction. It is closer to the actual coordinates. The simulation results show that the improved algorithm can improve the localization accuracy effectively without increasing hardware consumption in nodes comparing with conventional DV-Hop algorithm.②Based on cluster-based approach for collaborative single target tracking algorithm, a dynamic clustering target tracking(DCTT) algorithm is proposed which improves the accuracy of target tracking and prolong the network lifetime.In this approach, the sensor nodes collaboratively monitor and track the movement of the target. Tracking process includes three phases, this involves target detecting, dynamic clustering and localization of target. When the target moving into the monitoring area , detection of target is initiated by the phenomenon emitted by the target. Sensor nodes are dynamically clustered based on received signal strength. The function of leader sensor node is to compute the location of target, based on the information obtained from other sensor nodes in its cluster. A Weighted centroid target localization algorithm based on perceived probability is used to compute the location of target. This paper focuses on tracking error and energy consumption involved in tracking the target. And our simulation results show that our dynamic clustering target tracking algorithm can achieve more energy efficiency and better performance.