Multimodal Machine Learning Technologies in Tunnel Geotechnical Applications - NAT2024

Applications of machine learning and robotics in the tunneling industry are still nascent. Research and development are ongoing in TBM guidance; prediction of TBM cutter wear and tunneling induced ground deformations; and robotic arms for segment installation, inspection, cleaning, and maintenance of TBM cutting tools. This paper presents a case study of geological profile reconstruction using multimodal machine learning. The results from this study can aid the development and application of multimodal AI technologies to generate data reports and geological profiles in real time, taking voice recordings, photos, or videos obtained directly from the field as inputs.