Controlling Concurrent Change - A Multiview Approach Toward Updatable Vehicle Automation Systems

Abstract

The development of SAE Level 3+ vehicles [{SAE}, 2014] poses new challenges not only for the functional development, but also for design and development processes. Such systems consist of a growing number of interconnected functional, as well as hardware and software components, making safety design increasingly difficult. In order to cope with emergent behavior at the vehicle level, thorough systems engineering becomes a key requirement, which enables traceability between different design viewpoints. Ensuring traceability is a key factor towards an efficient validation and verification of such systems. Formal models can in turn assist in keeping track of how the different viewpoints relate to each other and how the interplay of components affects the overall system behavior. Based on experience from the project Controlling Concurrent Change, this paper presents an approach towards model-based integration and verification of a cause effect chain for a component-based vehicle automation system. It reasons on a cross-layer model of the resulting system, which covers necessary aspects of a design in individual architectural views, e.g. safety and timing. In the synthesis stage of integration, our approach is capable of inserting enforcement mechanisms into the design to ensure adherence to the model. We present a use case description for an environment perception system, starting with a functional architecture, which is the basis for componentization of the cause effect chain. By tying the vehicle architecture to the cross-layer integration model, we are able to map the reasoning done during verification to vehicle behavior.