Hidden Gaussian Markov model for distributed fault detection in wireless sensor networks
- 15 March 2017
- journal article
- research article
- Published by SAGE Publications in Transactions of the Institute of Measurement and Control
- Vol. 40 (6) , 1788-1798
Abstract
Wireless sensor networks are based on a large number of sensor nodes used to measure information like temperature, acceleration, displacement, or pressure. The measurements are used to estimate the state of the monitored system or area. However, the quality of the measurements must be guaranteed to ensure the reliability of the estimated state of the system. Actually, sensors can be used in a hostile environment such as, on a battle field in the presence of fires, floods, earthquakes. In these environments as well as in normal operation, sensors can fail. The failure of sensor nodes can also be caused by other factors like: the failure of a module (such as the sensing module) due to the fabrication process models, loss of battery power and so on. A wireless sensor network must be able to identify faulty nodes. Therefore, we propose a probabilistic approach based on the Hidden Markov Model to identify faulty sensor nodes. Our proposed approach predicts the future state of each node from its actual state, so the fault could be detected before it occurs. We use an aided judgment of neighbour sensor nodes in the network. The algorithm analyses the correlation of the sensors’ data with respect to its neighbourhood. A systematic approach to divide a network on cliques is proposed to fully draw the neighbourhood of each node in the network. After drawing the neighbourhood of each node (cliques), damaged cliques are identified using the Gaussian distribution theorem. Finally, we use the Hidden Markov Model to identify faulty nodes in the identified damaged cliques by calculating the probability of each node to stay in its normal state. Simulation results demonstrate our algorithm is efficient even for a huge wireless sensor network unlike previous approaches.Keywords
This publication has 18 references indexed in Scilit:
- Distributed Event Detection in Wireless Sensor Networks for Forest Fires2013
- Decentralized fault detection in Wireless Sensor Network based on function error2013
- Fault Detection and Recovery in Wireless Sensor Network Using ClusteringInternational Journal of Wireless & Mobile Networks, 2011
- Sensor faultsACM Transactions on Sensor Networks, 2010
- A New Method for Node Fault Detection in Wireless Sensor NetworksSensors, 2009
- A Survey of Fault Management in Wireless Sensor NetworksJournal of Network and Systems Management, 2007
- Weighted-Median Based Distributed Fault Detection for Wireless Sensor NetworksJournal of Software, 2007
- Distributed fault detection of wireless sensor networks2006
- Fault-tolerant clustering of wireless sensor networks2004
- Coverage in wireless ad hoc sensor networksIEEE Transactions on Computers, 2003