Data mining for fault diagnostics in cyberphysical systems

From ISLAB/CAISR
Title Data mining for fault diagnostics in cyberphysical systems
Summary Data mining for fault diagnostics in cyberphysical systems
Keywords Machine learning, research preparatory, supervised learning, text data mining.
TimeFrame The project should be done during the spring of 2014 and finished in late May. Ideally, the student(s) should start before the end of 2013.
References [[References::[1] Sankavaram, Kodali, Martinez Ayala, Pattipati, Singh, Bandyopadhyay, "Event-driven Data Mining Techniques for Automotive Fault Diagnosis", 21st International Workshop on Principles of Diagnosis, (2010), http://www.phmsociety.org/node/451

[2] Saxena, Wu, Vachtsevanos, "A Hybrid Reasoning Architecture for Fleet Vehicle Maintenance", IEEE Instrumentation & Measurement Magazine, pp 29-36, (2006) [3] Kargupta, Gilligan, Puttagunta, Sarkar, Klein, Lenzi, Johnson, "MineFleet ®: The Vehicle Data Stream Mining System for Ubiquitous Environments", in (May and Saitta, Eds) Ubiquitous Knowledge Discovery, Lecture Notes in Artificial Intelligence 6202, Springer-Verlag, pp. 235–254, (2010)]]

Prerequisites Learning systems, multivariate analysis, programming skills
Author
Supervisor Thorsteinn Rögnvaldsson, Stefan Byttner
Level Master
Status Open

Generate PDF template

The problem is learning to link specific characteristics/features with observed historical faults in mobile cyberphysical systems (city buses). The available data bases are: (1) a large database with on-board data on a fleet of city buses over the period Aug 2011 – Dec 2013; (2) the service records for the same buses over the same period. The project is suitable for 2 students. The students must have experience with matlab, be skilled in mathematics, be interested in text data mining and be able to work in an organized way. Deliverables: (a) A definition of the problem (b) A state-of-the-art description of diagnostics methods, specifically data-mining based (c) A set of suitable diagnostic cases from the data bases (d) A list of suitable methods to use (e) Results of testing with the set of suitable methods on the set of suitable cases (f) A report The work is well suited for writing a short scientific paper in the end and submit it to a conference. The project is suitable for 2 students with high work capacity and the ambition to show abilities for scientific or high-level development work.

Work packages: (a) Defining the problem (b) State-of-the-art definition (c) Extracting cases from the data bases (d) Defining the set of methods to test (e) Running the tests (f) Analysis, conclusions and writing the report