Verification and Validation Framework for Safe Railway AI Systems Across Their Lifecycle (VRAIL)
The VRAIL framework addresses challenges of AI-based safety-critical signaling systems with the goal to enable their certification
Description
The VRAIL framework addresses both regulatory and technological challenges, providing methods and techniques for AI-based safety-critical rail systems. It serves as an enabling technology for developing a certifiable AI-based signaling system, which is essential for fully automated train operations at Stadler.
Key data
Projectlead
Deputy Projectlead
Project team
Martin Rejzek, Stefan Brunner, Preami Uthayavathanan, Carmen Frischknecht-Gruber
Project partners
Stadler Signalling AG
Project status
ongoing, started 11/2024
Institute/Centre
Institute of Applied Mathematics and Physics (IAMP)
Funding partner
Innosuisse - Innovationsprojekt
Project budget
704'456 CHF
Publications
-
Von Black Box zu Safety Evidence : Framework zur Sicherheitsargumentation für KI in Bahnsystemen
2026 Reif, Monika; Fernández Moguel, Leticia; Shin, Jiwon
-
Assurance framework for safe and trustworthy AI in railway systems
2026 Müller, Manuel; Brunner, Stefan; Fernández Moguel, Leticia; Reif, Monika; Kälin, Tommy Lee
-
From ODDs to simulation : automatic railway scenario generation
2026 Fernández Moguel, Leticia; Kälin, Tommy Lee; Reif, Monika; Shin, Jiwon