Bachelor-/Masterarbeit:
Industry 4.0 aims at the end-to-end digital networking of machines, plants, and IT systems. Cyber-physical systems (CPS) play a central role in this context, as they couple physical systems, also called assets, with digital models and services. The Asset Administration Shell (AAS) provides a standardized concept for this digital representation of assets and thus forms a key foundation for interoperability and scalability in Industry 4.0 applications [1].
For the interaction of CPS, appropriate modeling of the involved assets is required. In addition to the AAS itself, extensive standardizations exist, such as standardized submodel templates from the IDTA [2], submodels based on the Semantic Aspect Meta Model (SAMM) , and semantic classification systems such as ECLASS [4]. A major challenge arises from the large scope and level of detail of these standardizations, which often makes the modeling space difficult to navigate in practice. In many cases, it is unclear which concepts, submodels, or standards are best suited for a specific application scenario.
This lack of clarity not only leads to uncertainty during modeling but also increases the barrier to adopting standard-compliant solutions. As a result, real-world projects frequently rely on partially or fully proprietary modeling approaches, which limits the intended interoperability.
Recent advances in the field of large language models (LLMs) open up new possibilities for supporting the modeling process. AI-assisted approaches can help identify and apply relevant standards, templates, and modeling options in a context-aware manner, for example by integrating external knowledge sources using Retrieval-Augmented Generation (RAG). At the same time, aspects such as traceability, result quality, and determinism introduce new requirements for such approaches.
The objective of this work is to investigate existing options for modeling AAS-based cyber-physical systems using AI-assisted tools, to investigate them in practice, to evaluate them, and, where appropriate, to combine them or implement extensions.
Requirements
- Programming experience, e.g. in Python, for development of toolchains
- Familiarity with code generation tools and APIs (optional but beneficial)
- Experience with or strong interest in Large Language Models and AI-based code generation techniques
- Motivation to work with existing software components, AAS specifications, and integration infrastructures
Tasks
- Literature review on standardizations as well as modeling approaches and tools for the AAS in the Industry 4.0 context
- Setup, configuration, and practical deployment
- Creation of different modeling approaches as combination and extension of selected tools as well as implementation of additional components
- Definition of evaluation criteria focusing on result quality, determinism, usability, and extensibility
- Evaluation of the approaches using application examples and a small user study
