Leveraging Information Technology to Understand and Assist Vulnerable Road Users

Information and communication technologies can greatly improve traffic capacity, efficiency, and stability, but most connected-vehicle and signal management advances focus on motor vehicles, leaving a major gap for vulnerable road users. We propose the Vision Enhanced Non-motorized Users Services (VENUS) smart node, a cooperative SPaT infrastructure that uses edge-based computer vision and AI to detect each pedestrian and cyclist’s location, class, pose, and mobility status, and to generate real-time, directional crossing requests. VENUS also serves as a communication hub, sharing SPaT data and bidirectional messages with signal controllers, connected vehicles, and personal devices. In tests with 1,076 users at six intersections, VENUS achieved 90.24% accuracy for directional-aware crossing triggers and 89.87% for mobility status estimation across normal users and four disability groups, while remaining low-cost and compatible with existing connected-vehicle infrastructure.