| Forschungsprojekt: | "TrustMe" |  |
| Certifiable AI Applications in Aerospace Production Systems | ||
| Research Area: | Artificial Intelligence, Transfer Learning, Imitation Learning, Process Certification, Manufacturing Process Optimization | |
| Funded by: | Federal Ministry for Economic Affairs and Energy | |
| (LuFo VII-1) | ||
| In cooperation with: | Airbus Operations GmbH (Airbus), Airbus Aerostructures GmbH (ASA) Fraunhofer Society (FhG) German Aerospace Center (DLR) Helmut Schmidt University/University of the Federal Armed Forces Hamburg (HSU) University of Augsburg (UA) | |
| Consortium Leader: | Airbus Operations GmbH | |
| Project Start: | December 2025 | |
| Project End: | May 2029 | |
| Contact Person at the Institute: | M.Sc.Ehssan Roshankar |
Description:
TrustMe - Certifiable AI Applications in Aviation Production Systems
The overarching project goal of TrustME is to establish a foundation for the secure and certifiable use of AI in aviation production.
The central goal of TrustME is the development and validation of architectures, methods, and procedures that enable AI applications in the context of aviation production to be
certifiable, transparent, and verifiable. By creating a trustworthiness framework that ensures the traceability of AI decisions, the
main barrier to the industrial use of this technology will be overcome. The ultimate goal is to enable more efficient, higher-quality, and more sustainable
manufacturing through the safe and trustworthy use of artificial intelligence.
The Hamburg University of Technology (TUHH), represented by the IPMT Institute and the IFPT Institute, is focusing on developing certifiable AI-supported approaches for quality prediction and optimization of drilling processes for rivet holes in aircraft component assembly. The particular focus is on developing AI models that, through transfer learning and imitation learning, enable rapid and data-efficient adaptations to new manufacturing processes. Transfer learning is used to transfer drilling quality predictions to new process variants with minimal effort. Simultaneously, imitation learning allows manual process characteristics to be transferred to robotic drilling applications, enabling high-quality rivet holes to be achieved with flexible robotic systems. Another key aspect of the project is the development and application of methods for certifying the AI models and the robotic drilling process. The existing certification of the manual drilling process is used as a basis to simplify the certification of the automated, robotic process, thus bridging the gap between theoretical research and practical, aerospace-specific validation on a laboratory scale. Through this holistic approach, the "TrustMe" project makes a significant contribution to increasing process quality, efficiency, and safety in aerospace manufacturing.
Contact person at the institute: M.Sc. Ehssan Roshankar