Current automotive E/E architectures are comprised of hardware and software and are mostly designed in a monolithic approach, static over the lifetime of the vehicle. Design, implementation and updates are mostly performed on a per-component-basis, exchanging complete Electronic Control Units (ECUs) or their software image as a whole. With an increasing amount of functionality being realized in software, the benefits of software can be used increasingly. This includes modularization of components, which forms the basis for updates and addition of functions. Additionally, this modularization allows the consolidation of ECUs and supports a higher level of integration. Such modularization and dynamic behavior over the lifetime of a vehicle feet, as well as a single vehicle does, however, hold a lot of challenges for safety-critical systems. Safety-critical systems, such as cars, require their behavior to be deterministic. The design of such modular systems needs to consider and cope with uncertainties in modular architectures. This paper highlights some of the dimensions of uncertainty, which will exist in future E/E architectures and presents initial approaches on how to manage these.

The development of automotive electronics and software systems is often associated with high costs due to their multi-domain nature (including control engineering, electronics, hydraulics, mechanics, etc). The involvement of these different disciplines makes it difficult for control engineers to test their controllers with them being integrated in the whole system, in early development phases. By introducing the "XiL approach", ETAS wants to leverage virtualization techniques in order to bring embedded systems to the desk of every developer. The XiL approach implies fast development cycles due to easy integration of software, vehicle and plant components on the PC. XiL strives for seamless transition between X-in-the-loop setups, with X representing any control model (M), software (S), or hardware (H) under test. This paper will outline this methodology with an engine management system example.

Model-based Systems Engineering plays a key role in the automotive industry by reducing the development costs while designing more complex controllers. Thanks to the continuous improvement of models accuracy and their computational performance, introducing system simulation in form of "in-the-loop" simulations in early stages of the development is possible. However, this is often limited by the efforts required to integrate and co-simulate heterogeneous models originating from different domains (e.g. electronic, mechanic, hybrid, etc). This work presents a standard-based integration, simulation and analysis solution leveraging the AUTomotive Open System ARchitecture (AUTOSAR) and the Functional Mockup Interface (FMI) [1] standard, and enabling a seamless transition from Model- to Software-in-the-loop simulations. Moreover, we demonstrate an accurate powertrain co-simulation based on the above-mentioned solution.