Variant Management for Technical Architecture of Highly-Automated Driving Systems

Abstract

As automotive systems become more and more complex and customized, the variability of and within technical architectures also significantly rises. The rising adoption of highly-automated driving features further adds complexity to the type and number of variants to handle. In addressing this issue, effective variant handling via well-structured feature models plays a crucial role. Although many approaches deal with the variability related challenges in the software level, there exists no proven variant management approach for the technical architecture level. The variant management method for technical architecture shall adequately handle different levels of variants as well as cover the needs and requirements of the users in the model-based system engineering (MBSE) environment. This includes not only modelling the variability existing in the system and visualizing architectural variants and dependencies between features but also ensuring traceability and enabling efficient collaboration between different design levels (i.e. system design, functional architecture design, technical architecture design, subsystem design) throughout the V-Model via a consistent a feature modelling concept and smooth exchange of technical information. In this thesis, we present a variant management method for the technical architecture, which addresses the identified variability related challenges and the needs from the user side.