Research Project "KI-DEG"
Funding: The KI-DEG project is supported as part of the R&D project and pilot line funding program through funds from the EFRE Technology Funding Program 2021–2027. This funding contributes to enable method-based innovation projects with transfer potential in the manufacturing sector and to strengthen the competitiveness, particularly of SMEs.
The KI-DEG research project is developing methodological frameworks to process heterogeneously distributed production data in a structured manner and make it usable in a context-specific way.
The goal is to make it easier for small and medium-sized enterprises (SMEs) in particular to access data- and AI-driven decision support without requiring the involvement of specialized data analysis teams.
Project partners
The project is being carried out by a consortium of industry and research partners:
GTV – Gesellschaft für Technische Visualistik mbH
Development of adaptive visualization logic and user-centered methodological approaches.
Fraunhofer-Institut für Werkzeugmaschinen und Umformtechnik (IWU)
Methodological modeling of semantic data structures and the integration of industrial production contexts.
Background
Many manufacturing companies have extensive data sets stored in various systems, formats, and levels of quality. This fragmentation makes it difficult to identify patterns, assess situations accurately, and make informed decisions – especially in dynamic manufacturing environments.
Project Objective
KI-DEG follows a methodological research approach..
The focus is on developing a so-called Visual Semantics Layers – a semantic mediation layer that structurally enriches production data and places it within a domain-specific context. Building on this, the project explores approaches to how AI-supported methods, in conjunction with semantic models, can recommend suitable visualization options to support decision-making processes. The final decision-making authority always remains with the users.
Expected results
The project does not produce market-ready software products, but rather transferable methodological results, including:
- Concepts for the semantic structuring of heterogeneous production data
- Models for role- and context-dependent visualization logics
- A research-based demonstrator for the methodological validation of the developed approaches
These results form a methodological foundation for their integration into existing software tools or further research and transfer projects after the project’s conclusion.