Collaborative project of Fraunhofer IESE, Fraunhofer IKS, and the University of York aims to accelerate time to market with new safety architecture
Autonomous systems require a paradigm shift in safety engineering
In the new ICON research project “LOPAAS” (Layers of Protection Architecture for Autonomous Systems), Fraunhofer IESE is collaborating with Fraunhofer IKS and the University of York in England in the Assuring Autonomy International Programme (AAIP) to develop an approach to assuring the safety of autonomous systems and, in particular, autonomous vehicles. The internal Fraunhofer funding program “ICON – International Cooperation and Networking” supports bilateral cooperation projects with internationally renowned universities and non-university research institutions. Over the next three years, 3 million euros will be available for the collaboration to bring about the necessary paradigm shift in safety engineering of autonomous systems.
The market introduction of autonomous systems and automated driving, in particular, poses many challenges. One key challenge is how to assure safety without operating with such severe restrictions regarding speed and other factors that the added value disappears. This is precisely the challenge that has been underestimated in many forecasts to date, and pilot studies conducted by automotive manufacturers also confirm passengers’ perception that autonomous vehicles are mostly slow and hesitant.
The Fraunhofer Institute for Experimental Software Engineering IESE, the Fraunhofer Institute for Cognitive Systems IKS, and the University of York have been researching this topic for years and are advancing the state of the art and the state of the practice. With the help of the ICON project, the partners are now bundling their research competencies to develop a reference safety architecture and safety argumentation for automated driving and autonomous systems. This includes, in particular, “dynamic risk management”, which gives the vehicle a better “understanding” of the current risks of a driving situation. Dynamic risk management takes into account not only external influencing factors such as the driving behavior of other road users, but also internal factors such as uncertainties in the perception of the current situation. Taking uncertainties in the safety architecture into account offers new opportunities to use Artificial Intelligence (AI) for safety-critical tasks. It describes how external and internal influencing factors can be used to automatically determine and control risks during ongoing operation.
The long-term goal of the project is to enable autonomous vehicles to drive faster and less hesitantly, as they can assess risks more safely and accurately. To achieve technology transfer, the results of the project are subsequently to be incorporated into standardization and norming.
The project partners – who contributes what expertise?
Fraunhofer IESE and Fraunhofer IKS are the two leading Fraunhofer Institutes in the area of safety engineering of autonomous systems. Together with the University of York, which is internationally regarded as the central research institution in field of safety, and the Assuring Autonomy International Programme, a consortium is being formed that can have a formative influence on future autonomous driving.
“For IESE, the topic of “Autonomous Systems” is of particular relevance, since systems and software engineering is the decisive success factor for the timely introduction and dependable operation of autonomous systems. In this ICON project, we particularly contribute our expertise in dynamic risk management, which enables autonomous systems to assess and control the risk of their options for action in a situation-specific manner”, says Dr. Rasmus Adler, Program Manager “Autonomous Systems” at Fraunhofer IESE and project manager of ICON “LOPAAS”.
The focus of the ICON project also fits the strategic orientation of Fraunhofer IKS. In this research project, IKS will focus on trustworthy AI-based situation recognition as well as runtime monitoring of the associated uncertainties.
ICON “LOPAAS” will bring about a major change in safety engineering
Safety assurance of autonomous systems requires a paradigm shift in safety engineering. Many new aspects need to be considered, such as the use of AI-based algorithms in safety-relevant applications and automated risk assessments. The expertise of the University of York, which has been conducting research on safety assurance of software-based systems for more than 30 years and has been instrumental in shaping today’s standard in safety engineering, is crucial for creating a new scientific consensus and presenting it in a convincing, structured safety case.