Consistency of Data-Defined Models (A05)

• Subject Area: Software Engineering and Programming Languages
• Term: since 2023
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 501798263

Many new CPS enclose components obtained from machine learning, which are often hard to model, verify, and predict. We address this by solving consistency problems between specified behaviour and the actual behaviour at run-time. Therefore, we investigate how to formulate specifications that an artificial intelligence (AI) system must adhere to and how to systematically construct a safety shield to ensure safety at all times. We will answer the research question of how we can design a safety shield ensuring consistency of data-defined models with safety specifications to include the AI system in a V-SUM.

• DFG Programme: Collaborative Research Centres
• Subproject of: SFB 1608: Consistency in the View-Based Development of Cyber-Physical Systems
• International Connection: USA
• Applicant Institution: Karlsruher Institut für Technologie
• Project Heads: Professor Dr.-Ing. Matthias Althoff; Professor Dr. André Platzer