Publication

This chapter provides an overview of architecture and systems engineering for autonomous driving system, through a set of complementary perspectives. For practitioners, a short term perspective uses the state of the art to define a three layered functional architecture for autonomous driving, consisting of a vehicle platform, a cognitive driving intelligence, and off- board supervisory and monitoring services. The architecture is placed within a broader context of model based systems engineering (MBSE), for which we define four classes of models: Concept of Operations, Logical Architecture, Application Software Components, and Platform Components. These classes aid an immediate or subsequent MBSE methodology for concrete projects. Also for concrete projects, we propose an implementation setup and technologies that combine simulation and implementation for rapid testing of autonomous driving functionality in physical and virtual environments. Future evolution of autonomous driving systems is explored with a long term perspective looking at stronger concepts of autonomy like machine consciousness and self-awareness. Contrasting these concepts with current engineering practices shows that scaling to more complex systems may require incorporating elements of so-called constructivist architectures. The impact of autonomy on systems engineering is expected to be mainly around testing and verification, while implementations shall continue experiencing an influx of technologies from non-automotive domains.