Berlin, April 22, 2025. Artificial intelligence (AI) has the potential to optimize rail transport - the possible applications range from better obstacle detection and more energy-efficient operation to self-driving trains. This is where the safe.trAIn research project, funded by the Federal Ministry for Economic Affairs and Climate Protection (BMWK) and the EU, comes in. It aims to make AI safe and reliable for use in fully automated rail transportation. Two new standards from the German Institute for Standardization (DIN) and the German Commission for Electrical, Electronic & Information Technologies (DKE) are important milestones on the way to fully automated train operation.
DIN DKE SPEC 99002 defines key terms relating to AI in the rail environment. DIN DKE SPEC 99004 describes how the area of application of AI systems in rail transportation can be precisely defined. The two standards are therefore important milestones on the way to fully automated train operation. The results obtained will also be incorporated into international standardization processes, thus strengthening Germany's competitiveness as a location.
Closely interlinked in standardization projects
The two DIN DKE SPECS were developed by experts from the rail sector, AI, industry and research; DIN and DKE supported the standardization projects. "Norms and standards help to bring innovative AI technologies onto the rails quickly and safely," explains Syad Akkoub, Project Manager for Research and Transfer at DIN. "The standards developed via safe.trAIn show how standardization provides the necessary foundations to drive forward complex technological challenges. By closely interlinking research, industry and standardization, we are paving the way for a standardized safety architecture in fully automated rail transport."
Marko Kesic, Mobility Project Manager at VDE, adds: "The more complex the traffic situation, the more challenging the use of driverless trains - for example autonomous trains in dynamic urban traffic or regional trains running on open routes. Therefore, norms and standards are of central importance in order to ensure safe and reliable operation."
Active for attractive rail transport
safe.trAIn is a lighthouse project of the German Standardization Roadmap AI published by DIN and DKE. The standardization roadmap promotes standards for the development and application of AI technologies and thus contributes directly to the AI Act. safe.trAIn is supported by a broad consortium of the rail and AI industry, technology suppliers, research institutions and standardization and testing organizations. One focus of the lighthouse project is also on the development of standardized testing methods and tools to ensure the safe use of fully automated trains. DIN is leading the standardization and dissemination work package for this - with the main aim of transferring the relevant project results into standardization.
Uniform understanding facilitates cooperation
DIN DKE SPEC 99002 "Terminology - AI in railway applications" defines terms in the cross-cutting topics of AI and railroads. These uniform terms are essential to improve communication and cooperation between all stakeholders in this environment - for example between train manufacturers, suppliers, transport companies, operators, certifiers and approval bodies as well as scientific institutions. The second standard DIN DKE SPEC 99004 " Specification of ODD in Rail" describes how the area of application of AI systems in rail transport can be precisely defined. "ODD" stands for Operational Design Domain and defines the conditions under which an autonomous vehicle can operate safely and effectively. This provides companies and test centers with a reliable basis for safely defining the operational limits of AI applications in this environment.
About the consortium
The consortium leader of the safe.trAIn project is Siemens AG. In addition to DIN and VDE, the members of the consortium are: Siemens Mobility GmbH, BIT Technology, Solutions GmbH, Bridgefield GmbH, Edge Case Research GmbH, ITQ GmbH, Merantix Labs GmbH, SETLabs Research GmbH, TÜV NORD Systems GmbH & Co. KG, TÜV Rheinland InterTraffic GmbH, TÜV SÜD Rail GmbH, Fraunhofer IAIS, Fraunhofer IKS, Düsseldorf University of Applied Sciences and Otto von Guericke University Magdeburg.
Further information on safe.trAIn and the DIN DKE SPECs can be found at www.safetrain-projekt.de and at din.de and dke.de. The DIN DKE SPECS are available at www.dinmedia.de.
