Student research opportunities
An FDI-inspired approach to diagnosis of DES
Project Code: CECS_675
This project is available at the following levels:
Honours, Masters, PhD
Keywords:
Diagnosis, DES
Supervisor:
Dr Alban GrastienOutline:
Diagnosis is the problem of detecting faults in a system and, if faults are detected, to isolate and identify precisely which fault occurred.
Diagnosis has been applied for a wide range of systems. In case of continuous systems, the system dynamics is describes in terms of equations on real-valued variables and sensors returns the real values of some of the system variables. The FDI (fault detection and isolation) community developed efficient diagnosis algorithms by using redundancy in the system and checking that the observed values satisfy the internal equations.
On the other hand, discrete event systems (DES) are dynamic systems whose variables have discrete values. DES are usually diagnosed by monitoring the states after each observation. This is extremely costly as it requires tracking all possible explanations of the observations, and the method does not scale up to very large systems.
This project is about exploring an approach of diagnosis for DES that is inspired by the FDI approach. In such an approach, the diagnosis is performed by a number of simple indicators (probably similar to chronicles).
Goals of this project
Define a framework for diagnosing DES with indicators
Explore how to build indicators for DES
