Energy Efficient NARX Model for Target Tracking in Wireless Sensor Network

Abstract

Mobile target tracking is one of the most important applications of wireless sensor network. In general, sensor nodes equipped with a battery which is non-rechargeable and irreplaceable and target tracking requires continuous monitoring which consumes significant energy. Therefore, energy constraint is of paramount issue in wireless sensor network. In recent times, various techniques starting from Kalman based techniques to soft computing based are used to track the exact current position of target and to predict the next possible position of target. There is a trade-off between energy efficiency and tracking accuracy. This paper addresses an Energy Efficient Non-linear Autoregressive Neural Network with Exogenous Inputs (EENARX) target tracking algorithm which achieves better performance for energy efficiency as well as tracking accuracy. The energy efficiency is achieved by reducing the number of active nodes and by reducing the communication between sensor nodes and sink node as most of the energy is consumed in communication. EENARX also achieves tracking accuracy by using NARX model for prediction. For energy efficiency proposed algorithm is compared with Circle based target tracking scheme and Minimal Contour Tracking Algorithm (MCTA) and to prove tracking accuracy it is compared with Time Delay Neural Network (TDNN) approach.