Abstract
In this paper, we study and propose a simple algorithm for localization applications in Wireless Sensor Networks (WSNs) with distributed learning approach. The considered model consists of some sensor nodes (beacons), which know exactly their locations, and non-beacon nodes whose positions are needed to determine. The non-beacon nodes are represented by vectors including hop-counts information from themselves to all of the beacons. Based on the information, the beacons give local estimates on locations of non-beacon nodes by Kernel method, by which each non-beacon node determines its final position using majority vote rule on the estimates. Computer simulations demonstrate the effectiveness of the proposed method. © 2014 IEEE.