Distributed Sensor Diagnosis in Complex Wired Networks for Soft Fault Detection Using Reflectometry and Neural Network

• Published in: 2019 IEEE AUTOTESTCON pp. 1 - 6
• Main Authors: Osman, Ousama, Sallem, Soumaya, Sommervogel, Laurent, Olivas, Marc, Peltier, Arnaud, Bonnet, Pierre, Paladian, Francoise
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01-08-2019
• Subjects: complex wired network; detection; distributed diagnosis; localization; MCTDR; multisensor data fusion; neural network; Time domain reflectometry
• ISSN: 1558-4550
• DOI: 10.1109/AUTOTESTCON43700.2019.8961069

This paper presents a neural network (NN) approach to detect and locate automatically multiple soft faults in complex wired networks using multi-sensor information fusion. The location process is based on monitoring the wired network topology by several sensors (reflectometers). The soft fault detection and location are achieved by Multi-Carrier Time Domain Reflectometry (MCTDR) combined with feedforward Multi-Layer Perceptron (MLP) neural network, trained by backpropagation algorithm. The NN ensures the data fusion between different reflectometers. The required datasets for training and testing the NN are generated by simulation of faults for various soft faults scenarios (fault locations and fault impedance). The effectiveness of the proposed approach is demonstrated by simulation for locating multiple soft faults in branched network.