Cooperative fault detection and isolation in a surveillance sensor network: a case study

Abstract

This work focuses on Fault Detection and Isolation (FDI) among sensors of a surveillance network. A review of the main characteristics of faults in sensor networks and the associated diagnosis techniques is first proposed. An extensive study has then been performed on the case study of the persistent monitoring of an area by a sensor network which provides binary measurements of the occurrence of events to be detected (intrusions). The performance of a reference FDI method with and without simultaneous intrusions has been quantified through Monte Carlo simulations. The combination of static and mobile sensors has also been considered and shows a significant performance improvement for the detection of faults and intrusions in this context.