Fault Tolerant Sensor Network Design with Respect to Diagnosability Properties

Abstract

Interest for more reliable and safety systems is growing as control systems become increasingly complex and encounter various unexpected component failures. Indeed, fault-tolerant control, and fault detection and identification take into account, at the early design stage, unexpected system failures. For this reason, the ability of a system (sensors, actuators, components) to tolerate a fault must be taken from designing. This paper deals with the problem of fault tolerant sensor network design. A new approach based on residual structure analysis is proposed to provide a fault tolerant architecture. The proposed analysis is based on diagnosability properties in case of sensor loss situations; it evaluates the capacity of the system to be fault tolerant. A graph representation is used to study diagnosability, and evaluate fault tolerance capacity in different cases of sensors distribution.