AutoDevSafeOps – Integrated development and operation of safe automotive systems

Highly automated and autonomous driving functions are becoming increasingly important, which means that modern cars are becoming mobile computers. This is accompanied by a rapidly growing amount of data in the vehicle, which must be processed in real time or in near-real time.

The resulting high demands on computing power, flexibility and efficiency demand new approaches in computing and software architecture. In order to achieve these goals, the strategic projects of the “MANNHEIM” funding guideline, within the framework of the “Zukunftsfonds Automobilindustrie” fund, are researching high-performance computing platforms, novel vehicle architectures and sustainable software development processes and methods.

The MANNHEIM project AutoDevSafeOps, in which Fraunhofer IKS is also involved, is developing an important approach to meet the high demands of automated and networked vehicles on the already existing software architecture. The vehicle software must not only work efficiently and safely at all times, but at the same time allow flexible optimization of the functionalities via software updates. The AutoDevSafeOps project provides an important building block for this by developing a holistic DevOps approach that enables over-the-air updates for safety-critical driving functions.

The special feature: In the AutoDevSafeOps project, a hitherto unique holistic DevOps approach with integrated safety methods is developed. It is intended to enable modular updates of safety-critical driving functions, including the associated safety processes and procedures, across the system boundary between the vehicle and the back end – and this in a dynamically changing environment. The results and the specific benefit are evaluated and demonstrated by the experienced project partners from research and industry in three exemplary use cases.

Fraunhofer IKS manages the work package for developing the technological fundamentals and analyzes their application in the use cases of AutoDevSafeOps.

The starting point is the design of a contract language which maps the requirements of the planned use cases, such as the over-the-air update of the automated vehicle systems, on capabilities or guarantees of involved components. In the case of an update during vehicle operation, it must be ensured that the reliability of the overall system is not compromised. Contracts form an essential basis here, both during the architectural design as well as during operation or runtime. In order to verify strict adherence to the contracts, runtime monitors generated from contracts are used. These runtime monitors in turn are a prerequisite for mitigation strategies, for example if, in the event of deviations or violations of contracts, the intervention of a monitoring entity becomes necessary in order to maintain essential driving functions. The methods mentioned are tested both in a digital twin and in the real vehicle.

Through its research contribution, Fraunhofer IKS guarantees the specification of resilient and trustworthy cognitive systems with flexible and adaptive architectures. The architectures developed can manage and adapt themselves, so that they function smoothly even when they get into situations that were not foreseeable at the time of development.

The project is funded by the Federal Ministry of Education and Research.