NEURAL NETWORK BASED INDOOR LOCALIZATION FOR WIRELESS SENSOR NETWORK

摘要

In Wireless Sensor Network (WSN), location estimation (Localization) is important for routing efficiency and localization-aware services. This paper examines the WSN application which allows indoor localization based on the neural network and grid sensor training phase for accurate localization of sensors. The communication between the modules is monitored during the experiment whereby the received radio signal strength indicator (RSSI) values are recorded by a mobile sensor node. The received data used for training the feed-forward type of neural network such as Levenberg-Marquardt (LM) and Back Propagation Algorithm (BPA). This paper provides a solution to discover sensor nodes in WSN's using the past and present values obtained from neighbouring nodes based on neural networks. LM and BPA algorithms are used to find shortest distance between node and coverage of the node. This solution can be also a way to discover the malfunctioning nodes that are not a subject of an attack. Being localized on the base station level, this algorithm is suitable even for large-scale sensor networks.