Research on right-turn warning and control technology for urban intersection bus

Abstract

This paper proposes a right-turn warning and control technology for urban intersection buses to address the safety issues associated with right turns. Firstly, the safety of bus-vehicle, non-motorized vehicle, and pedestrian conflicts is studied, and conflict levels are defined using the K-means clustering algorithm, trajectory smoothing, and the 85th percentile cumulative frequency curve method. Secondly, a measurement system based on the OpenCV laser triangulation method and velocity sensors, combined with YOLO v7 for target classification and recognition, is used to calculate the Time to Collision (TTC) value. A microcontroller receives the TTC signal to generate signals for vehicle deceleration or turning, thereby avoiding collisions. SUMO scene simulation is conducted for validation. The results show that the proposed system enables buses to predict traffic conflicts in advance, issue early warnings, and take evasive actions, thus reducing accident rates. This research provides inspiration for related studies on the safety of right turns for urban intersection buses both domestically and internationally.