Topology Optimization for Target Localization in Wireless Sensor Networks

Target localization is an important application in wireless sensor networks. However, because of the limitation of the capability of the individual nodes and the noises in the environment, a single sensor usually cannot localize a target accurately and observations from multiple sensors need to be combined to improve the localization accuracy. In this paper, the Centroid method, a widely used node localization algorithm is applied to target localization. We first identify two factors that affect the accuracy of the Centroid method: node density and uniformity. Then we propose a classifying procedure by which a cluster-based network is formed. In this cluster-based network, only a part of sensors are active and the node density and uniformity is adjusted to improve the localization accuracy. Experimental results show that by using our classifying method, the accuracy of the localization result can be improved with fewer active nodes.