A Novel Wireless Sensor Network Node Localization Algorithm Based on BP Neural Network

Abstract

The accurate localization of wireless sensor network node is one of the supporting technologies of network application. A novel localization algorithm of wireless sensor network node based on BP neural network is put forward in the paper. This localization algorithm constructs the BP neutral network model in accordance with the number of the anchor node firstly, and then trains the network by the anchor node and estimates the location of the unknown node. Moreover, the virtual anchor node is introduced into this algorithm in order to realize its optimization, which increases the anchor node scale in the network and improves the localization accuracy of the node. The simulation experiment results in two different conditions show that compared with Centroid algorithm and DV-Hop algorithm, the localization algorithm of this paper estimates the location of the unknown node more precisely and improves the location accuracy more effectively. This algorithm demonstrates its merits greatly.