Target Tracking in WSN Using Dynamic Neural Network Techniques Moxanki A. Bhavsar 1 , Jayesh H. Munjani 2(&) , and Maulin Joshi 1 1 Sarvajanik College of Engineering and Technology, Surat, Gujarat, India moxmahi@ymail.com, maulin.joshi@scet.ac.in 2 Chhotubhai Gopalbhai Patel Institute of Technology, Surat, Gujarat, India jayeshec12@gmail.com Abstract. Wireless Sensor Networks (WSN) are increasingly being envisioned for the collection of data, such as physical or environmental properties. Unlike detection of an event, tracking requires ensuring continuous monitoring. Resource constrained nature of wireless sensor networks makes energy ef ficient tracking a challenging task. Prediction based approaches try to save energy by reducing an avoidable communication. A Kalman-based approach has been widely used for target tracking but is inaccurate in the case of maneuvering target due to its inability to incorporate nonlinearity. In this paper, dynamic neural network based approaches called Time Delay Neural Network (TDNN) and Nonlinear Autoregressive network with Exogenous inputs (NARX) are proposed for non-cooperative target tracking application. The performance of NARX is compared with Kalman based approach and TDNN in terms of tracking accuracy. Simulation results show that NARX outperforms both Kal- man approach and TDNN for target tracking applications. Keywords: Wireless Sensor Network  Target tracking  Neural networks Nonlinear autoregressive network with exogenous inputs (NARX) Time Delay neural network (TDNN)  Kalman filtering  Prediction mechanism 1 Introduction The smart cities without sensors can’t be predicted in near future. An auto configurable sensor network made up of small but powerful sensors, which can work without any human intervention, can be utilized effectively to save human efforts and time. Wireless Sensor Network has been widely utilized in civic and military industries, especially for tracking targets in critical areas of some fields such as intruder detection, vehicle location tracking, logistics management and anti-terrorism etc. In literature, many static and dynamic duty cycle based methods are proposed that try to adjust sleep time of sensor motes when the target is not likely to be nearby [1]. To save power, sensor nodes can be put into sleep mode but it increases chances to miss the target. The sleep mode power consumption of sensor mote still increases as the product of time and per second sleep mode power consumption becomes larger. Deep © Springer Nature Singapore Pte Ltd. 2018 P. Bhattacharyya et al. (Eds.): NGCT 2017, CCIS 828, pp. 771–789, 2018. https://doi.org/10.1007/978-981-10-8660-1_58