Safe Routes for All (SR4A) - Using AI-Based Image Recognition and Machine Learning Algorithms for Network-Wide Assessment and Routing of Multimodal Trips Based on Level of Traffic Stress

This proposal will leverage artificial intelligence (AI) and machine-learning (ML) to fundamentally alter the approach to data generation for Complete Streets, and dramatically improve the capabilities of agencies looking to provide network-wide assessments and prioritization of improvements to multimodal facilities. Specifically, this research will (1) use AI-based image processing to extract missing geometric attributes from high-resolution aerial imagery to generate all necessary inputs for segment and crossing level of traffic stress (LTS) methods, and (2) use ML-based routing algorithms and optimization to identify low stress routes that match user preference and help identify most critical infrastructure investments to unlock the low-stress network. The research will overcome a major obstacle in the evaluation of complete networks that are safe and comfortable for all users, which is the availability and coverage of spatial data. This research will generate all necessary input data for industry-standard LTS methods for both segments and crossings, and integrate the results in network-wide routing that considers LTS and user preferences. Using AI and ML technologies, this research will fill critical data gaps needed to empower agencies to accurately and efficiently assess their multimodal networks, and inform investment decisions for creating complete networks for walking and bicycling.