TALENTA succeeds: A Milestone for Intelligent Infrastructure Management
The TALENTA research project, launched in September 2023, provided a significant impetus for intelligent, interoperable, and practical infrastructure management. Together with project partners concedra and Constructor University Bremen (formerly Jacobs University), the vision emerged to develop a high-performance prototype that reduces technological complexity while setting new standards for collaboration, transparency, and efficiency.
The project focused on developing an innovative asset management platform for digital twins in the transportation sector. Particular attention was paid to the needs of small and medium-sized enterprises (SMEs) in order to facilitate their access to modern technologies and promote their active participation in the digital transformation.
Read about the results and challenges of this project in this blog post.
The Challenge: Fragmented data, lack of standards
Digital twins are considered a key technology for the future of transportation infrastructure. Nevertheless, their widespread adoption is often hindered by fundamental obstacles:
- lack of a shared understanding of data models
- challenges in integrating heterogeneous systems
- limited access to relevant data
- lack of transparency in AI-driven decisions
- high costs and a lack of viable business models
This is exactly where TALENTA comes into play—with the goal of systematically overcoming these barriers.
The Approach: Semantics, AI, and Interoperability
TALENTA’s approach is based on the targeted integration of semantic technologies, artificial intelligence, and interoperability. At its core is an ontology-based information model that enables a unified semantic understanding across organizational boundaries. Building on this foundation, a platform was developed that not only integrates data but also transforms it into a context that can be intelligently utilized.
A key component is the ability to semantically integrate heterogeneous data. TALENTA automatically brings together a wide variety of data sources—from traditional relational databases to unstructured text documents, complex 3D models, and image data. This information is structured within a knowledge graph, linked together, and thus placed in a holistic context.
The custom-developed TALENTA ontology serves as a common language for all stakeholders. It provides a standardized representation of assets, their properties, and their relationships, thereby laying the foundation for a consistent, cross-organizational understanding of data.
Another key component is the integrated explainable AI toolset. It ensures that AI-driven decisions are not perceived as a “black box,” but remain traceable and transparent. By combining established methods such as SHAP, LIME, partial dependence plots, and Grad-CAM with a complementary evaluation metric, the explainability of models becomes measurable and comparable.
This approach is complemented by a scenario-based search mechanism that automatically translates natural-language queries into technical queries. This allows relevant assets to be efficiently identified and retrieved from the knowledge graph in a context-aware manner.
The platform is rounded out by an intuitive asset management portal that offers a user-friendly interface. This portal enables digital assets to be made available, used, and reused across organizations—supported by secure authentication and authorization mechanisms.
A Practical Approach: Focus on SMEs
One of TALENTA’s key objectives was to make it easier for small and medium-sized enterprises and local governments, in particular, to access modern technologies. With this in mind, new business models were specifically developed to enable flexible and economically attractive use of these technologies. Through approaches such as “Digital Twin as a Service” and “Machine Learning as a Service,” companies can leverage innovative technologies without having to make large upfront investments, while at the same time tapping into new economic potential.
Research meets Application
Over a period of 30 months, the project was implemented in six sequential work packages. The scope ranged from the development of the ontology to the integration and processing of data, and finally to the full integration of all components into a prototype platform. Throughout the project, a consistent focus was placed on international standards, open data sources, and reproducible methods in order to create a sustainable, interoperable, and future-proof solution.
Results and Added Value
As part of the TALENTA project, key technological building blocks for intelligent digital twin asset management were developed and successfully integrated into a platform prototype. These include an ontology-based asset model, methods for the semantic integration of heterogeneous data, a knowledge graph, and an XAI-supported analysis and search concept.
The results clearly demonstrate the technical feasibility of the secure and interoperable use of digital twins across organizational boundaries. At the same time, TALENTA should be understood as a proof of concept (PoC) and not as a finished product. Building on these results, a detailed implementation plan was developed that outlines the path toward a market-ready platform. The focus is on service-based approaches such as “Digital Twin as a Service” and “Machine Learning as a Service.” In this way, TALENTA lays the foundation for future applications in infrastructure management and specifically paves the way for the next steps toward product development and market launch.
Looking ahead
Even though individual components—particularly in the areas of system integration and optimization—can still be further developed, TALENTA provides a robust foundation for future applications and product developments. The commercialization strategy developed during the project clearly demonstrates that the technologies developed have the potential to be transferred to productive use and successfully established on the market in the long term.
Conclusion
TALENTA exemplifies a new generation of digital platforms that are intelligent, interoperable, and consistently designed with practical applications in mind. The project demonstrates how the combination of semantic technologies, knowledge graphs, and explainable artificial intelligence can create genuine added value for infrastructure management.
Thus, the successful completion of this project marks not only the end of a research project, but also the beginning of new opportunities for digital transformation in the transportation sector.
We would like to thank our project partners concedra and Constructor University Bremen for their excellent collaboration, as well as the research program MFund for making this project possible.
About the BMDV’s mFUND Funding Program
As part of the mFUND funding program, the BMDV has been supporting research and development projects focused on data-driven digital innovations for Mobility 4.0 since 2016. Project funding is complemented by active professional networking among stakeholders from politics, business, government, and research, as well as the provision of open data on the Mobilithek portal. For more information, visit www.mfund.de.
