Cloud Services for Autonomous Driving
Together with Fraunhofer IESE, Robert Bosch GmbH has developed a safety concept for a cloud service for automated driving.
Last modified:
Together with Fraunhofer IESE, Robert Bosch GmbH has developed a safety concept for a cloud service for automated driving.
Last modified:
What it is all about
Fraunhofer IESE collaborates with Bosch on a safe cloud service to support highway pilots
The challenge:
Lack of safety standards for cloud services in the automotive sector
The support:
Fraunhofer IESE and Bosch develop a safety concept for the automotive sector
The result:
The approach provides the foundation for a safety standard for automated driving
Your benefit:
Transfer of current safety standards into the company process
In the future, automated driving will be increasingly supported by cloud services: A vehicle‘s sensor system only has limited range and is expensive, which is why in the future, vehicles will be supplied with information via the cloud. This information may originate from other vehicles, but also from other sources such as weather services.
This project with Bosch was also about a cloud service: This one informs a highway pilot about the weather conditions on the road and the road‘s static friction coefficient. Highway pilots are reliant on this information since their area of application and their driving behavior depend on the level of friction. Developing a safety concept for this cloud service − despite the lack of standards for safety issues in the automotive sector − was the challenge facing the experts of Fraunhofer IESE.
There is as yet no standard explaining how to develop cloud services for highway pilots. The safety standard ISO 26262 and the Safety-Of-The-Intended-Functionality (SOTIF) standard ISO PAS 21448 do refer to vehicles, but their application cannot be transferred easily to cloud services. IT security standards also address cloud services, but not the safety issues from the automotive sector.
However, just because there is no standard does not mean that such services can be developed any which way. One should still keep to the state of the art, even though it cannot be found in a single standard.
The researchers at Fraunhofer IESE are familiar with the state of the art as well as with current developments in the area of safety and autonomous driving. Based on this background knowledge, they compiled relevant requirements from standards in this project with Bosch and demonstrated how to implement these methodologically.
Furthermore, Fraunhofer IESE supported Bosch in the methodological implementation. In cooperation with the domain experts from Bosch, they developed a functional architecture that represents the entire information processing. Subsequently, the team analyzed the individual processing steps using component fault trees. They also conducted a safety analysis regarding the cloud platform on which the software developed by Bosch runs. Based on the analysis results, the Fraunhofer experts derived a safety concept and developed a safety argumentation using the Goal Structuring Notation (GSN). They modeled all artifacts with safeTbox, the tool framework developed by Fraunhofer IESE to support the development and certification phases of safety-critical systems. Due to their modular character, the modeled artifacts can be easily tailored to different customer requirements.
This is what Erik Lesser, Chief Product Owner, Robert Bosch GmbH, says:
As the result of the project, Bosch received a comprehensive safety concept for its special cloud service. However, the approach and the methods and tools employed can be easily transferred to other cloud services. Thus, the project also provides a model-based safety engineering approach for cloud services and lays the foundation for a hitherto missing standard.