Measuring the Evolution of Automotive Software Models and Meta-Models to Support Faster Adoption of New Architectural Features
Licentiate thesis, 2015

Background: The ever-increasing amount of software in cars today combined with high market competition demands fast adoption of new software solutions in car development projects. One challenge in enabling such a fast adoption is to develop the architecture and models of the automotive software systems in a structured and controlled way. Objective: The main objective of this thesis was to enable the fast utilization of new architectural features in automotive software models. This was achieved by developing methods and tools to analyze the evolution of the domain-specific meta-models that are used to define the language of software models and their features. In particular, we wanted to identify the underlying changes caused by meta-model evolution related to a specific set of architectural features and assess their impact on both the architectural models and modeling tools used by different roles (e.g., the Original Equipment Manufacturers, OEMs, and their suppliers) in the development process. Method: We achieved our objective by conducting an action research project in close collaboration with the Volvo Car Group (Volvo Cars) and the consortium of the AUTOSAR standard, which aims to standardize the architecture of automotive software systems. This collaboration facilitates fast feedback from experts in the field on the problems, ideas and methods we developed in the course of this research, thereby enabling the validation of the research results and proposed methods in on-going development projects, i.e., their direct application in the industry. Results: We identified the most suitable software measures for measuring the evolution of both the automotive software models and meta-models. The calculation and presentation of the measurement results were done with the support of two, newly-developed tools. We also developed a method for the automated identification of an optimal set of new architectural features that should be adopted in development projects to facilitate the decision-making process concerning the selection of which of these new features would be adopted. Conclusion: We applied the developed methods and tools to the architectural models and meta-models used at Volvo Cars and concluded that they provide valuable input for the decision-making process concerning which new versions of the standardized meta-model should be used in different projects. We also concluded that these methods and tools can facilitate the assessment of the impact of adopting new architectural features on the different roles involved in the development process.

software measurement

Modeling and meta-modeling

automotive software architectures

Jupiter room 243, Hörselgången 11, Chalmers Lindholmen
Opponent: Assoc Prof. Ekkart Kindler, Department of Applied Mathematics and Computer Science, Technical University of Denmark

Author

Darko Durisic

University of Gothenburg

ARCA - Automated Analysis of AUTOSAR Meta-Model Changes

2015 IEEE/ACM 7th International Workshop on Modeling in Software Engineering,; (2015)p. 30-35

Paper in proceeding

Measuring the Impact of Changes to the Complexity and Coupling Properties of Automotive Software Systems

Journal of Systems and Software,; Vol. 86(2013)p. 1275-1293

Journal article

Identifying Optimal Sets of Standardized Architectural Features - A Method and its Automotive Application

International Conference on Quality is Software Architectures,; (2015)

Journal article

Evolution of Long-Term Industrial Meta-Models – A Case Study

39th International conference on Software Engineering and Advanced Applications,; (2014)

Other conference contribution

Areas of Advance

Information and Communication Technology

Subject Categories

Information Science

Jupiter room 243, Hörselgången 11, Chalmers Lindholmen

Opponent: Assoc Prof. Ekkart Kindler, Department of Applied Mathematics and Computer Science, Technical University of Denmark

More information

Created

10/10/2017