A flexible automotive systems architecture for next generation ADAS

Abstract

The evolution of Advanced Driver Assistance Systems (ADASs) increasingly paves the way towards autonomous driving. The increase in sensor data to be processed as well as the complexity of processing algorithms demands more powerful processing platforms. Additionally, aspects like functional safety and platform security have to be taken into account. Algorithms should be re-usable to facilitate rapid development. Current solutions are very much focused on single sub-systems and incompatible with next generation processing platforms available for ADASs. In this paper we therefore propose a novel architecture for future ADAS development, taking into account the aforementioned aspects. We also outline tools to aid developers in dealing with the real-time and functional safety aspects of ADASs. The utilization of the presented architecture in the EU-funded project ADAS&ME is also described.