Real-time LiDAR applications for assessing the safety of vulnerable road users at urban intersections

In recent years, the use of real-time LiDAR (Light Detection and Ranging) technology has gained increasing relevance in the field of intelligent transportation systems, particularly for its ability to capture high-resolution spatial and temporal data of dynamic environments. This paper presents a methodology that integrates real-time LiDAR scanning to analyse the behaviour of both vehicular traffic and pedestrian movement in a complex urban intersection, with the overall goal of assessing road safety conditions and supporting the planning of targeted interventions. The study focuses on the continuous monitoring of a selected urban intersection using a multi-beam LiDAR sensor positioned to maximize the field of view of all incoming and outgoing traffic lanes as well as pedestrian crossings. The sensor identifies the class of road user—whether motorized vehicle’s driver, pedestrian, or motorcyclist – and captures real-time three-dimensional point clouds allowing the extraction of trajectories, speed profiles, and interaction patterns between different road users. Through advanced data processing and object-tracking algorithms, it becomes possible to identify potentially hazardous situations, such as near-misses, sudden braking events, and unsafe pedestrian crossings, with the aim of deepening the knowledge of road user behaviour. Preliminary results confirm the feasibility of identifying and analysing interactions between vehicles and Vulnerable Road Users (VRUs). These interactions are assessed through Surrogate Safety Measures (SSMs) able to detect potential conflict events, using thresholds established in literature, namely Time to Collision (TTC), and Gap Time (GAP). The proposed methodology is scalable and adaptable to various urban environments, making it a practical tool for the assessment of safety conditions at road intersections and the consequent design of effective safety countermeasures aimed for safer and smarter cities. In addition, this study contributes to the expanding field of digital road safety analysis by integrating real-world LiDAR-derived data into virtual simulators, such as driving, motorcycle and pedestrian simulators, improving accuracy of driver behaviour studies and urban mobility planning.