Use of Artificial Intelligence (AI) in Automated Aircraft Production – Project “TrustME”

As part of the joint project “TrustME” – “TRUSTworthy AI for Manufacturing and Engineering Processes” – the Fraunhofer Institute for Manufacturing Technology and Advanced Materials IFAM in Stade, Germany, together with project partners, is developing certifiable artificial intelligence (AI) solutions for aircraft production.

Using a validation platform for the structural assembly of aircraft fuselages, Fraunhofer IFAM is testing AI models for

• component detection,
• automated assembly planning and
• enhanced industrial robot precision.

Semantic data models (SDM) and agent-based assistance systems (Agentic AI) ensure that AI decisions are transparent, verifiable and certifiable for the aviation industry.

Project Objectives

Through the TrustME project, the experts for Automation and Production Technology are laying the foundation for the safe and compliant use of artificial intelligence in aircraft manufacturing. The goal is to develop AI systems in such a way that they are not only high-performing but also traceable, robust and prospectively certifiable – in accordance with EASA guidelines (EASA AI Roadmap 2.0) and the EU AI Act.

The focus is on applications that directly add value to the manufacturing of aircraft fuselages: from automated component detection and intelligent assembly planning up to increased robot precision.
 

AI-Assisted Aircraft Fuselage Assembly and Aircraft Component Detection

To address the shortage of skilled workers and the growing number of aircraft variants within individual aircraft programs, Fraunhofer IFAM is developing AI solutions for automated fuselage assembly:

• AI-powered detection of aviation-specific components based on synthetic image data from CAD models
• Determination of the assembly status – e.g., during fuselage assembly – from camera data
• Integration of the AI pipeline into a web application for easy use in manufacturing
• Validation of the AI models on fuselage demonstrators under realistic conditions

The training data used, typically images, is predominantly generated synthetically. This ensures that the approach remains scalable and transferable to other assemblies and aircraft programs.

Ontologies as Knowledge Base for Trustworthy AI

To ensure that AI decisions are transparent and verifiable, researchers at Fraunhofer IFAM are developing domain-specific ontologies for aircraft production:

• Structured representation of product, process and resource knowledge (PPR model)
• Storage of equipment capabilities and interfaces in a graph database

Ontology-based Retrieval-Augmented Generation (RAG) makes the domain knowledge of large language models (LLMs) accessible in a structured way. This allows domain expertise to be integrated in a targeted and semantically precise manner, resulting in responses that are more technically sound, explainable and trustworthy.
 

Agent-Based AI Systems for Planning and Execution

Based on these ontologies, Fraunhofer IFAM is developing an LLM-based multi-agent system for production planning and execution:

• Agents use external software functions (Function Calling), e.g., for resource planning
• Integration of real-time production data such as machine status or assembly progress
• Support for workers through natural language interaction, e.g., via a tablet

The aim is to create a scalable system that not only enables more efficient and flexible planning of complex assembly processes in aircraft manufacturing, but also makes autonomous decisions that are transparent and traceable.
 

Improving Precision: AI Enhances Robot Accuracy

For high-precision joining and machining processes, scientists are investigating how AI can increase the absolute accuracy of industrial robots beyond what is achievable with conventional calibration methods:

• Combining conventional model-based calibration with a neural correction model
   

Project Context and Funding

TrustME – “TRUSTworthy AI for Manufacturing and Engineering Processes”– is a joint project involving Fraunhofer IFAM (Stade) and Fraunhofer IGCV (Augsburg).

Consortium partner:

• Airbus Operations GmbH
• Airbus Aerostructures GmbH
• Fraunhofer IFAM
• Fraunhofer IGCV
• Helmut Schmidt University
• German Aerospace Center
• University Augsburg
• Hamburg University of Technology

The project is funded by the German Federal Ministry for Economic Affairs and Energy (BMWE) as part of the LuFo Climate VII-1 Aviation Research Program.

Funding code: 20D2401C
Duration: December 1, 2025 – May 31, 2029