Developing predictable vehicular distributed embedded systems on multi-core

Abstract

In this paper we address the challenges related to supporting modeland component-based development of predictable software for vehicular distributed embedded systems, utilizing multi-core platforms. We present a research plan for the development of new methods and techniques to deal with these challenges. The techniques will support various aspects such as modeling of the software architecture; supporting multiple criticality levels; verifying predictability of the system using end-to-end timing analysis; code generation; and providing a predictable runtime support on multi-core platforms by virtualizing a certified single-core real-time operating system. As a proof of concept, we will implement the newly developed techniques in a commercial tool chain (Rubus-ICE). The efficacy of the newly developed techniques and the extended tool chain will be demonstrated on the industrial case studies.