Enhancing Vulnerable Road User Safety: A Community-Based Approach Using Data and AI-Driven Solutions

Abstract: Vulnerable Road Users (VRUs), including pedestrians, cyclists, and motorcyclists, account for over half of global road traffic fatalities, with the World Health Organization (WHO) reporting a concerning rise in these deaths. In the U.S., this trend underscores the need for innovative, community-based safety measures. This study leverages machine learning and AI to explore how socioeconomic disparities, environmental justice, and underserved communities influence VRU safety. By integrating datasets such as census tract data, land use, environmental justice, and geometric design, we aim to uncover key factors contributing to VRU crash frequency and severity. Additionally, mobility and trip data for both vehicular and VRU traffic will be analyzed to identify conflict-prone intersections, further refining safety interventions. By leveraging sources such as the NYS GIS Clearinghouse and open-source data, the study will pinpoint intersections with high conflict potential and recommend targeted solutions. The study also embraces the Safe System approach, which emphasizes collaboration across sectors to create safer environments that accommodate human limitations and shared responsibility for road safety. Key to this effort are three primary objectives: data integration and availability, predictive analytics, and the development of community-specific interventions. A case study in New York State will serve as a detailed risk analysis site, facilitating the identification of tailored, data-driven solutions to improve VRU safety outcomes in underserved communities. Through AI-driven insights, this study will pave the way for equitable, data-centric advancements in transportation safety.