The upcoming project SAGE: “LLM-based tools for the automated implementation and energy optimization of building automation systems,” funded by the Federal Ministry for Economic Affairs and Energy, is being conducted in collaboration with RWTH Aachen and Robert Bosch GmbH.
Increasing automation in buildings is necessary for energy-optimal control. The current building stock in Germany does not possess the prerequisites for such automation due to heterogeneous planning data. LLMs can extract this data and help construct a knowledge graph, which can then be used to increase automation.
The Chair of Automatic Control is responsible for designing a pipeline for automatic controller synthesis. This includes identifying the model structure, system identification and controller design using the constructed knowledge graph and LLMs. Key challenges are the robustness of such a pipeline and the interaction with system designers and end users.