A Fletcher-Reeves Conjugate Gradient Neural-Network-Based Localization Algorithm for Wireless Sensor Networks

• Published in: IEEE transactions on vehicular technology Vol. 59; no. 2; pp. 823 - 830
• Main Author: Chatterjee, A.
• Format: Journal Article
• Language: English
• Published: New York, NY IEEE 01.02.2010; Institute of Electrical and Electronics Engineers; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Algorithms; Applied sciences; Artificial intelligence; Backpropagation algorithms; Character generation; Cities and towns; Classification; Computer science; control theory; systems; Conjugate gradient (CG) algorithm; Conjugate gradients; Connectionism. Neural networks; Electric, optical and optoelectronic circuits; Electronics; Exact sciences and technology; Feedforward neural networks; Fletcher-Reeves update; Localization; Monitoring; Multi-layer neural network; Networks; neural network based classification; Neural networks; Position (location); Sensor systems; Sensors; Services and terminals of telecommunications; Spread spectrum communication; Switching and signalling; Systems, networks and services of telecommunications; Telecommunications; Telecommunications and information theory; Telemetry. Remote supervision. Telewarning. Remote control; Wireless sensor networks; wireless sensor networks (WSNs); Performance evaluation; Wireless telecommunication; Updating; Remote sensing; Networked system; Feedforward neural nets; Classification; Connectedness; Localization; Gradient method; Signal detection; Conjugate gradient method; Packet switching; wireless sensor networks (WSNs); Measurement sensor; neural network based classification; Neural network; Algorithm; Case study; Multihop network; Beacon; Wireless network; Fletcher― Reeves update; Conjugate gradient (CG) algorithm; Sensor array; Multilayer network
• ISSN: 0018-9545; 1939-9359
• DOI: 10.1109/TVT.2009.2035132

Multihop connectivity-based algorithms have been receiving increased attention in recent times for localization in wireless sensor networks (WSNs). This paper proposes the development of a Fletcher-Reeves update-based conjugate gradient (CG) multilayered feedforward neural network for multihop connectivity-based localization of a large number of sensor nodes in a 2-D sensor network on the basis of information gathered from beacon nodes. The neural-network-based system employs a classification scheme where the location of a sensor is simultaneously estimated in both the x - and y -directions. The usefulness of the proposed scheme is demonstrated by employing the scheme for three case studies, with varied environments, where it could consistently show better performance than two popular recently proposed schemes.