Over the past decade, USDOT and state DOTs have invested heavily in advancing roadway safety and operations through data-driven research, traditional GIS methods, and artificial intelligence-based methods. These efforts have led to meaningful insights into identifying crash hotspots zones and infrastructure vulnerabilities, improving existing data sets and capabilities, and developing policy recommendations, particularly in rural and hazard-prone areas. However, as the landscape of geospatial technology evolves, so too must our approaches. Tried and proven 3-D technologies — from UAV-based photogrammetry and LiDAR to neural rendering models like Neural Radiance Fields (NeRFs) and 3-D Gaussian splatting — now offer an unprecedented opportunity to elevate the practice and research from flat and map-based assessments to immersive and data-rich spatial intelligence systems. With the integration of these tools, DOTs can reconstruct, analyze, and simulate rural transportation networks in full three-dimensional fidelity, opening the door to proactive, predictive, and deeply contextualized safety interventions. As such, building on the existing foundation, this proposal aims to formally inaugurate 3-D geospatial technologies into the traffic safety framework — setting a new benchmark for traffic operations and safety with a specific focus on spatial analysis, disaster preparedness, and real-world impact. To operationalize the proposed vision, this research is structured around a set of interrelated objectives that collectively transform how traffic operations and safety could be measured, modelled, and managed — particularly in rural and hazard-prone areas. These objectives are:
1) perform 3-D roadway geometry analysis and visibility assessment based on the application of artificial intelligence techniques on 3-D data,
2) simulate real-world driving scenarios using 3-d rendering, and
3)evaluate the impact of flora cover on roadway safety and augment the framework with additional geospatial data sets as needed. Each objective builds upon the next, beginning with data acquisition and moving through semantic analysis, scenario simulation, and long-term infrastructure monitoring.
Exhibit D
