Difference between revisions of "AUTO-CAAS"

From CERES
Jump to: navigation, search
 
Line 30: Line 30:
 
   
 
   
 
==  Publications ==
 
==  Publications ==
 +
* W. Mostowski, T. Arts, and J. Hughes. [[media:Mostowski_mars2017.pdf|Modelling of Autosar Libraries for Large Scale Testing]]. 2nd Workshop on Models for Formal Analysis of Real Systems ([http://www.mars-workshop.org/mars2017/index.html MARS 2017]), Open Publishing Association, 2017.
 +
 
*  S. Kunze, W. Mostowski, M.R. Mousavi, and M. Varshosaz. [[media:mousavi_wasa_2016.pdf|Generation of Failure Models through Automata Learning]]. Second International Workshop on Automotive Software Architectures ([http://www.win.tue.nl/wasa2016/ WASA 2016]), IEEE CS Press, 2016.  
 
*  S. Kunze, W. Mostowski, M.R. Mousavi, and M. Varshosaz. [[media:mousavi_wasa_2016.pdf|Generation of Failure Models through Automata Learning]]. Second International Workshop on Automotive Software Architectures ([http://www.win.tue.nl/wasa2016/ WASA 2016]), IEEE CS Press, 2016.  
  

Latest revision as of 15:40, 2 June 2017

Automated Consequence Analysis for Automotive Standards

Brief Description

The Automotive Open System Architecture (AUTOSAR) standard is gaining momentum with several automotive manufacturers and there is a growing trend towards new vehicle platforms based on the latest versions of this standard. The standard enables manufacturers to allow Tier-1 suppliers to contract arbitrary Tier-2 software developer for ECUs, as long as the developed software conforms to the specified behavior according to AUTOSAR. This is in clear contrast to earlier situation, in which a preferred Tier-2 developer was appointed to develop software for all Tier-1 hardware suppliers. This paradigm shift brings about economical and financial benefits (both for suppliers and manufacturers). However, it also introduces certain risks and challenges.

The AUTOSAR standard is complex and does leave room for interpretation and optimizations. In order to be competitive, Tier-2 developers strive after implementing several optimizations and utilizing room for interpretation of the standard to make their product out-perform the competition.

The goal of this project is to exploit the technology of model-based testing in order to detect deviations from the AUTOSAR standard and furthermore trace the consequences of such deviations into visible deviating behavior (failures). To this end, we will use and enhance the model-based testing framework developed at QuviQ to detect deviations from the AUTOSAR standard. This framework is, for example, used by SP (Technical Research Institute of Sweden) to certify software delivered to Volvo Car Corporation. As noted before, one of the major obstacles in using the current model-based framework is the different interpretations of the standard. Unless the consequences of these interpretations are properly analyzed, such variations are justified by the developers. This poses a major challenge for the widespread application of the standard as a model for certifying components, modules, ECUs and vehicle functions.


Project Team

  • Thomas Arts, Co-Founder and CTO of QuviQ AB
  • John Hughes, Co-Founder and CEO of QuviQ AB, Professor at Chalmers University of Technology
  • Wojciech Mostowski, Assistant Professor, AUTO-CAAS Project
  • Mohammad Mousavi, Group Leader, Principal Investigator for AUTO-CAAS Project
  • Michael Svenstam, ArcCore AB

Publications

Resources

Outreach and Dissemination

Contact

Mohammad Mousavi, Professor of Computer Systems Engineering