Project

Full title: Verification Strategies and Safety Assurance of Continuously Evolving Automated Driving Systems

VerSACE develops new ways to verify and assure safety for automated driving systems that are updated frequently, without slowing down software delivery. The project combines agile DevOps with safety assurance so that verification becomes continuous and data-driven: safety analyses and requirements (Dev), pre-deployment testing (simulation, closed-loop and real-world with safety drivers), and post-deployment data collection (Ops) are tightly connected. By treating operational data as a “first-class citizen”, VerSACE enables targeted evidence generation that supports safety claims even as the system evolves.The expected results include modular, verification-friendly product architectures for complex software and AI/ML functions, a contract-based framework that aligns requirements, verification activities and safety arguments across system levels, and forward-planned verification strategies where acceptance criteria and methods are defined for several future releases in advance.  VerSACE also integrates falsification techniques to actively challenge assumptions and uncover hidden risks early. Finally, the project delivers a roadmap for tool support that can partially automate building and maintaining a safety case within a continuous integration and deployment environment, together with practical methods for industrial adoption.

Externa partners: Adamai, RemotiveLabs, KTH

Traffic safety benefit: VerSACE improves road safety by ensuring frequent software updates in automated   vehicles are continuously tested and monitored to help ensure safety in transport.

Key words: Continuous safety assurance, DevOps, Automated driving