Self Recurrent Neural Network Based Target Tracking in Wireless Sensor Network using State Observer

Abstract

Background & Objective: Mobile target tracking based on data from wireless sensor networks (WSN) is a hot topic that has been investigated both from a theoretical and practical point of view in the literature. Tracking the position and velocity of a target moving in WSN (especially in the context of uncertain noisy measurement channel) is a very challenging task. To deal with the uncertainty in system dynamics as well as uncertainty in target states, an Observer Based Self Recurrent Neural Network (OBSRNN) is proposed in this paper. Methods: The proposed algorithm employs a state observer based tracking control strategy and thereby allows for accurate estimation of mobile target moving along a predefined route in WSN. The Self Recurrent Neural Network (SRNN) framework is used to approximate the uncertainty in the system dynamics, while a full-order state observer is used to estimate the unknown state vector. Conclusion: The simulation analysis is performed to evaluate the efficacy of the proposed work.