AI4TWINNING

The project aims to automatically generate a system of interconnected digital twins of the built environment that span multiple resolution scales and provide rich semantics and coherent geometry. To this end, we will explore a multiscale, multisensor, and multimethod approach that combines terrestrial, airborne, and space-based data acquisition, various sensors (visible, thermal, LiDAR, radar), and different processing methods that integrate top-down and bottom-up AI approaches. The key concept of this proposal, which represents a groundbreaking advancement, lies in deriving building information and intelligently fusing the resulting information through AI-based methods, thereby closing information gaps and increasing the completeness, accuracy, and reliability of the resulting digital twins. To simplify the process and improve the results, we make extensive use of informed machine learning by leveraging explicit knowledge about building design and construction. The goal of the project is not to create a single monolithic digital twin, but rather a system of interconnected twins at various scales that enables the seamless integration of city, neighborhood, and building models while keeping them up-to-date and consistent. Munich’s city center—an area surrounding the central TUM campus where large datasets from various sensors are already available—serves as the test and demonstration scenario.