The Austrian project FATE will develop a fault-based human-centric methodology for reasoning about cyber-physical systems (CPS) during different stages of their design.
Because CPS combine digital and analog behavior, they are complex to analyze and verify. Their verification relies on extensive expert knowledge and manual effort, and verification and testing tools are notoriously hard to use and results hard to interpret. For instance, if a specification violation is detected during simulation and testing, how to associate it to a specific fault? Or, when a design‘s parameters are configured to optimize certain functionality, how to know that it is robust to configuration perturbations (corresponding to production faults)?
As a follow-up of the project ADVANCED (https://projekte.ffg.at/projekt/3323906), which focused on checking robustness with respect to inputs and reliability by detecting anomalies, FATE will extend those results by addressing the questions
- Can we detect specification violations and associate them to specific faults?
- Can we extend our ML*+search-based testing approach from ADVANCED to explore the design space instead of the input space?
- Can we make tools more human-centric?
* Machine Learning
Facts
- Project duration: November 2022 – October 2025
- Coordination: AIT
- Budget: € 1,27 Millionen
- Förderung: € 0,96 Millionen
- Partner:
- Industrie: Infineon Technologies Austria
- Forschungseinrichtungen: Technische Universität Graz
Link