kumoh national institute of technology
Networked Systems Lab.

Muhammad Royyan, Jae-Min Lee and Dong-Seong Kim, "Stochastic-Based Faulty Node Detection Scheme in Industrial Sensor Network", IEEE Transactions on Industrial Informatics (IF 5.430, ISSN: 1551-3203), 2017. (Pr)
By :
Date : 2017-12-04
Views : 452

In this paper, a faulty node detection and prediction scheme with a Markov chain model are proposed for the industrial sensor network. Mostly industrial networks are heavily noised large-scale systems, and the system workload is unfairly distributed among the master node and sensor nodes. Therefore, the master node may not detect a faulty node in real-time. A faulty node could lead to severe damage to the system. In this paper, an enhanced faulty node detection scheme using a Markov chain model is investigated. Each sensor node's condition is divided into three states by probability calculation: \textbf{Good-}, \textbf{Warning-}, and \textbf{Bad-state}. The master node can predict which sensor node and when an error frequently occurs. The master node analyzes the faulty node during communication between the master node and several sensor nodes. This scheme can be used for detecting and preventing severe damage caused by sensor node failure of an industrial sensor network. Simulation results show that the proposed method can improve the reliability and accuracy of faulty node detection and the detection accuracy rate.

Submission date: 2017-12-04
Rejection date: 2018-06-18