The engineering of vehicular systems is becoming increasingly difficult, mainly
due to the ongoing integration of cyber-physical systems (CPS) aiming to
automate difficult tasks or provide additional features to drivers. This
automation potential leads to increasing complexity when engineering the vehicle
itself or its subcomponents. In particular the development of a future-oriented
kind of mobility, namely, connected autonomous vehicles (CAVs), is accompanied
by new challenges, leading back to the different domains to be considered. To
cope with this complexity and enable the mutual engineering of vehicular
embedded systems, the Software Platform Embedded Systems (SPES) framework
provides viewpoints and hierarchy layers in the shape of a matrix. However, to
address all domains considered during the development of CAVs, the SPES
methodology lacks specifications of how to model such vehicles across multiple
domains, which impede its utilization in actual industrial projects. Thus, to
deal with the mentioned issues, this article introduces an architecture
framework enabling model-based systems engineering (MBSE) of CAVs across
multiple domains. This includes a domain-specific reference architecture
mapping, a particular development process, a detailed architecture definition,
and a corresponding modeling approach. Subsequently, the resulting methodology
is then evaluated for applicability using a real-world case study on the
charging process of an electric vehicle (EV). The outcome of this article will
thereby increase the usability of the SPES framework for industrial projects on
the one hand and counteract the increasing complexity when engineering vehicular
systems on the other hand.
