Case study Knowledge4Retail – Digital twin for retail
Retailers increasingly require integrated, intelligent systems to adapt to dynamic consumer behavior and operational demands. In the Knowledge4Retail (K4R) project, fortiss designed semantic digital twin models that bridge physical retail spaces with AI-powered services across partners.
The result is a scalable, open-source infrastructure enabling automation, planning, and personalized services in future retail environments.
Challenge
Retailers face fragmented data silos and incompatible systems, making it difficult to deploy AI solutions that require semantic, structured, and real-time data. There was no unified way to model retail environments that could support AI-based planning and service orchestration. fortiss was brought on to design a modular, semantic modeling framework that supports integration and intelligent reasoning across diverse use cases.
Solution
fortiss developed a domain-specific semantic modeling architecture based on the Industry 4.0 Asset Administration Shell (AAS), extended to support real-time 3D retail environments and AI service integration. They created modular ontologies and interfaces to describe physical objects, services, and spatial data, enabling interoperability across K4R partners. The models supported use cases ranging from in-store robot navigation to AI-driven planogram optimization. fortiss contributed to the K4R open-source stack and coordinated knowledge modeling across distributed teams.
Result
- Created a semantic digital twin infrastructure aligned with AAS and industry standards.
- Enabled AI-based services like robot navigation, dynamic shelf management, and route optimization.
- Designed reusable ontologies and interfaces for modular modeling of retail objects and services.
- Supported integration of AI tools across multiple partners with a unified data representation layer.
- Contributed to the K4R open-source ecosystem to ensure long-term accessibility and collaboration.
- Demonstrated real-world use cases, including semantic robot planning and personalized customer services.
Outcome
Semantic modelling lays the foundation for the scalable integration of AI by providing machine-readable and context-aware data. Modularity and standardisation promote collaborative development and ensure long-term sustainability. It is also becoming clear that digital twins go far beyond mere 3D models and must also comprehensively map behaviour, semantics and service interfaces. Furthermore, open-source infrastructures play a key role in enabling innovation across industry and partner boundaries.
Project partner
