AIICS

Artificial Intelligence & Integrated Computer Systems

LiU » IDA » AIICS

Troubleshooting of Automotive Systems

Heavy trucks is an application domain for troubleshooting.

An overview of a typical troubleshooting procedure.

In the Troubleshooting of Automotive Systems project, methods for automatic and efficient troubleshooting of automotive systems are developed. This project is run in collaboration with the truck manufacturer Scania CV AB.

The trend in the automotive industry is that the complexity of the systems to repair increases which complicates the task of manual troubleshooting. This is an important problem in the automotive industry: efficient troubleshooting leads to improved uptime for the operator and saved money for the workshop.

Troubleshooting software should guide the mechanic in the workshop in the process of finding and repairing faults on a vehicle, by recommending actions to perform where the aim is to minimize the expected cost of repair. A recommended action is the first action in a plan of actions that has a minimal expected cost.

The vehicle is a partially observable system where the status of the components is only indirectly observable . Actions that can be performed on the system can affect the status of components, change the configuration of the system and make observations. Depending on the configuration of the system not all actions can be performed.

Issues addressed in this project include:

  • Methods for modelling of the diagnostic aspect of the system.
  • Methods for inferring possible diagnoses.
  • Methods for probabilistic contingent planning.
  • Scalability.

Publications

[4] Håkan Warnquist, Anna Pernestål, and Petter Säby. Anytime Near-Optimal Troubleshooting Applied to a Auxiliary Truck Braking System. In Proceedings of the 7th IFAC Symposium on Fault Detection, Supervision and Safety of Technical Processes (SafeProcess), Barcelona, Spain, June 2009. Accepted for publication. [ Conference ]
[3] Håkan Warnquist and Mattias Nyberg. A Heuristic for Near-Optimal Troubleshooting Using AO*. In Alban Grastien, Wolfgang Mayer, and Markus Stumptner, editors, Proceedings of the 19th International Workshop on Principles of Diagnosis (DX), Blue Mountains, Australia, September 2008. [ Conference | .pdf ]
[2] Håkan Warnquist, Mattias Nyberg, and Petter Säby. Troubleshooting when Action Costs are Dependent with Application to a Truck Engine. In Proceedings of the 10th Scandinavian Conference on Artificial Intelligence (SCAI), volume 173 of Frontiers in Artificial Intelligence and Applications, Stockholm, May 2008. IOS Press, Amsterdam, The Netherlands. [ Conference | .pdf ]
[1] Håkan Warnquist and Petter Säby. Conditional Planning for Troubleshooting and Repair in a Partially Observable Environment. Master's thesis, Linköping University, April 2008. LIU-IDA/LITH-EX-A-08/013-SE. [ E-Press | .pdf ]