Distributed Diagnosis of Networked, Embedded Systems

Networked embedded systems are composed of a large number of physically distributed nodes that interact with the physical world via a set of sensors and actuators, have their own computational capabilities, and communicate with each other via a wired or wireless network. Monitoring and diagnosis for such systems must address several challenges caused by the distribution of resources, communication limitations, and node and link failures. This paper presents a distributed diagnosis framework that exploits the topology of a physical system to be diagnosed to limit inter-diagnoser communication and compute diagnoses in an anytime and any information manner, making it robust to communication and processor failures. The framework adopts the consistency-based diagnosis formalism and develops a distributed constraint satisfaction realization of the diagnosis algorithm. Each local diagnoser first computes locally consistent diagnoses, taking into account local sensing information only. The local diagnosis sets are reduced to globally consistent diagnoses through pairwise communications between local diagnosers. The algorithm has been successfully demonstrated for the diagnosis of paper path faults for the Xerox DC265 printer.