Intersection Traffic State Estimation Based on Self-Organizing Maps and Connected Vehicles

Accurate traffic state estimation at intersections is essential for intelligent traffic signal control. Existing methods struggle to handle the high-volume, real-time data streams enabled by Connected Vehicles (CVs). This paper proposes a traffic state estimation method based on the Speed Transition Matrix combined with Self-Organizing Maps, designed to scale to high-frequency, large volumes of CV data streams in real time. The method is evaluated on a simulation model of an isolated intersection. Obtained results indicate that the method is well-suited for adaptive traffic signal control in the CV environment because it enables real-time CV data utilization for traffic state estimation.