Summary: Toward Distributed Diagnosis of Complex Physical
Gautam Biswas Sherif Abdelwahed Xenofon Koutsoukos Jyoti Gandhe Eric Manders
Department of Electrical Engineering and Computer Science,
and the Institute for Software Integrated Systems, Vanderbilt University.
The size and complexity of present day systems motivates the need for de-
veloping distributed fault diagnosis algorithms. This paper uses the Tran-
scend approach to fault diagnosis to develop a methodology that partitions the
set of possible fault candidates in a physical system to independent sets of faults
given a set of measurements. Separate diagnosers that do not interact with each
other can be constructed for each independent fault set while maintaining complete
diagnosability of the system. The implication of this approach is that a compu-
tationally expensive diagnosis task is decomposed into a set of computationally
simpler tasks that can be performed independently.
In the model based approach to fault diagnosis , an important entity is the system
model that provides the basis for analyzing observed discrepancies in system behavior for
isolating faulty components in the system. For large systems, building accurate models