Real-Time Soil Boundary Refinement in Excavation: A GeoBIM Framework Integrating Uncertainty Modeling Tools

This study investigates the potential of adaptive uncertainty modeling to enhance soil boundary estimation during excavation. A GeoBIM framework integrating Monte Carlo simulation and Kriging interpolation was implemented, enabling real-time refinement of boundary predictions. The results demonstrate significantly improved accuracy and reliability compared to traditional static methods, such as triangulated irregular networks (TINs) and manual excavation, especially in complex environments with limited data. The adaptive model’s ability to dynamically learn and improve as excavation data accumulate offers a key advantage for applications requiring high precision and responsiveness. This study highlights the importance of continuous data integration for subsurface modeling. Enhanced soil boundary estimations, when combined with advanced trajectory planning, can lead to more efficient, cost-effective, and environmentally sustainable earthwork operations. This research suggests that adaptive uncertainty modeling can serve as a core technology in automated and intelligent excavation and construction workflows, facilitating smarter and more sustainable earthworks. Keywords: Soil boundary prediction, Uncertainties, Intelligent excavation, Geological mapping, Building information modeling, Advanced trajectory planning, Triangulated irregular networks (TINs), Mean Absolute Error (MAE).