Localization in Wireless Sensor Networks with Cognitive Small World Characteristics

In this paper, a method for node localization in a novel small world - wireless sensor network (SW-WSN) is proposed. A SW-WSN is created by introducing small world characteristics into a conventional wireless sensor network (WSN). Introduction of long-ranged links between selected nodes in a WSN reduces average path length (APL) of the network without affecting the clustering coefficient (CC) significantly. Localization performance and anchor requirement is made more efficient by the introduction of newer links with an optimal probability of link addition (PLA). In this work a cognitive approach for computing the optimal PLA according to bandwidth availability is also developed. This helps in minimizing energy consumption and bandwidth requirement in the network during localization. Extensive localization experiments are conducted on U. of Michigan WSN repository. The proposed localization indicates significant improvements over conventional methods in terms of localization error and bandwidth utilization.