AI-Enhanced Data Collection and Integration

Transportation agencies often face persistent data challenges: information is fragmented across jurisdictions, inconsistent in quality, and incomplete across modes. Reliable, real-time data on emerging systems, such as connected and automated vehicles (CAV), micromobility, and active transportation, is especially limited. These data gaps make it difficult to anticipate system needs, identify safety risks, or design coordinated, multimodal networks.

The AI for Transportation Planning and Design (AI for TPD) initiative is exploring how artificial intelligence can transform how transportation data is gathered, cleaned, and integrated. Using computer vision, machine learning, and Internet of Things (IoT) technologies, the program supports new methods to automatically extract information from diverse sources such as satellite and street-level imagery, LiDAR, sensor feeds, dashcam video, crowdsourced inputs, and vehicle probe data.

AI algorithms can classify roadway elements, detect changes over time, and estimate travel behaviors at scales and speeds that are not feasible through manual analysis. These approaches improve data coverage, timeliness, and consistency, creating a stronger foundation for safety analysis, performance monitoring, and automated decision support. By building high-quality, continuously updated datasets, agencies can more effectively plan, evaluate, and maintain safer and more resilient transportation systems.

Building on improved data integration, the AI for TPD initiative is advancing new tools that use artificial intelligence and large language models (LLMs) to make transportation data easier to explore and apply. These tools will allow users to query complex datasets in natural language, asking questions such as "Where are the highest-risk pedestrian corridors?" or "How might proposed designs affect freight travel times?" and receive clear, visual responses.

By combining AI analytics, geospatial visualization, and dynamic simulation, agencies can test scenarios, compare design options, and monitor performance across safety, mobility, and efficiency metrics. This reduces the need for manual data processing and specialized modeling, enabling faster, more transparent, and evidence-based decision-making across transportation networks.