Hybrid Fault Diagnosis in Sensor Networks
Advisor Information
Azad Azadmanesh
Location
UNO Criss Library, Room 225
Presentation Type
Oral Presentation
Start Date
2-3-2018 1:45 PM
End Date
2-3-2018 2:00 PM
Abstract
Abstract – Wireless Sensor Networks (WSNs) are often prone to component failures due to factors such as energy depletion, environmental changes, and their deployment in hostile terrains. This research proposes a Distributed Fault Diagnosis (DFS) protocol to diagnose and remove faulty sensor nodes from further task participation. Most existing DFD algorithms consider simple and singular faults. Instead, the proposed algorithm takes advantage of hybrid fault modeling that allows for co-existence of various faults with different severities. Depending on node spatiality and application, the algorithm is able to dynamically tune the fault model and use one-hop or two-hop neighbors to adjust the accuracy of the fault diagnosis. The proposed approach will be simulated and compared against some benchmark algorithms.
Hybrid Fault Diagnosis in Sensor Networks
UNO Criss Library, Room 225
Abstract – Wireless Sensor Networks (WSNs) are often prone to component failures due to factors such as energy depletion, environmental changes, and their deployment in hostile terrains. This research proposes a Distributed Fault Diagnosis (DFS) protocol to diagnose and remove faulty sensor nodes from further task participation. Most existing DFD algorithms consider simple and singular faults. Instead, the proposed algorithm takes advantage of hybrid fault modeling that allows for co-existence of various faults with different severities. Depending on node spatiality and application, the algorithm is able to dynamically tune the fault model and use one-hop or two-hop neighbors to adjust the accuracy of the fault diagnosis. The proposed approach will be simulated and compared against some benchmark algorithms.