Diagnosis is the process of identifying or determining the nature and root cause of a failure, problem, or disease from the symptoms arising from selected observations, checks or tests. Diagnosis is crucial for systems safety, dependability, and maintainability. It’s a key ingredient of autonomous systems. The DISCO research team conducts a rich and varied methodological research in the field of automated diagnosis along two lines:

model-based diagnosis data-driven diagnosis

Dynamic systems exhibiting complex behaviors are analysed from a qualitative behavioral perspective able to provide informative and discriminant symptoms. Proposing relevant abstractions to interpret the available data is hence key to its research. The DISCO research team relies on formalisms borrowed to both the field of Automatic Control and the field of Artificial Intelligence and develops expertise at the intersection of these two fields. Model-based methods as well as machine learning and data mining methods are investigated to provide original diagnosis solutions. The pluridisciplinary approach and the wide spectrum of classes of systems that are dealt with are the trademarks of the team, whose results are recognized in both fields.

The research topics cover the analysis of the impact of faults on a system throughout its entire life cycle, with special emphasis on the following items:

• Diagnosis in cooperative contexts and networked systems
• Diagnosis in fault tolerant and autonomous architectures
• Diagnosability analysis and diagnosis aware design
• From diagnosis towards maintenance and prognosis
• Data mining for diagnosis and prognosis