An observer based fault tolerant scheme for distributed networked control systems

This paper deals with the development of an observer based fault tolerant scheme for distributed networked control systems (NCS). The NCS under consideration consists of a number of nodes and is hierarchically structured with a central station at the higher level and subsystems distributed at the process level. The system nodes are connected by a communication network like Ethernet. The core of the scheme is to construct a distributed observer that provides the central station and local subsystems with process informations needed for control, fault diagnosis and system reconfiguration purposes. Making use of the distributed computation power available at the subsystems and taking into account the packet-oriented data transmission, an optimal trade-off strategy between the network QoS (Quality of Service) and the system QoP (Quality of Performance) is integrated into the fault tolerant scheme.