Target Tracking in WSN using NARX model
Jayesh Munjani, Moxanki A. Bhavsar, Maulin M. Joshi · 2017
Wireless Sensor Networks (WSN) are increasingly being envisioned for event detection and collection of physical or environmental data. Tracking applications require ensuring continuous monitoring which is a much difficult task that mere detection of an event. Energy-saving tracking is a difficult task of resource constrained wireless sensor network. A network lifetime can be enhanced by incorporating prediction based scheme, which saves energy of sensors by limiting communication. While basic Kalman filter is successfully implemented for linear applications, is unable to perform best of its capability in case of maneuvering target. Considering target movement as time series, neural network based approach is a good alternative as it is a model free estimator. In this paper, Nonlinear Autoregressive Network with Exogenous Inputs (NARX) neural network is proposed for tracking of the non-cooperative moving target. Simulation results show that proposed NARX based approach gives better accuracy as compared to Kalman filter.