Summary: System Diagnosis using Hybrid Failure
Sherif Abdelwahed and Gabor Karsai and Gautam Biswas 2
Abstract. This paper presents an approach for robust diag-
nosis of a general class of dynamic systems based on a tempo-
ral failure propagation model. The proposed approach can be
applied to a general class of systems with both time and event
driven dynamics such as hybrid and discrete event systems.
The paper presents the syntax and semantics of the proposed
model and introduces the diagnosis approach.
Diagnostic algorithms detect, isolate and estimate system fail-
ures using observed signals and measurements from the sys-
tem sensors and actuators, and comparing them against a
model that captures nominal and/or faulty behavior. The ob-
served behavior is explained by a set of hypotheses that cap-
ture parameterized changes in the system components.
Two kinds of modeling paradigms have been commonly
used to describe the behavior of engineering systems: analyti-
cal models and associative models. Analytical models such as