A Static–Dynamic Driving Risk Field Model for CAV at Signalized Intersections: Enhancing Efficiency, Safety, and Decarbonization

Intersections are critical bottlenecks in urban traffic, contributing disproportionately to crashes, delays and emissions. This study proposes a unified risk field model for connected and automated vehicles (CAVs) at signalized intersections that integrates static environmental factors (markings, signal timing and geometry) and dynamic factors (moving vehicles) within a potential energy field framework. A risk perception circle mechanism is introduced to simulate drivers’ risk perception while reducing computational complexity. A gradient-based control law converts the risk field into a desired acceleration and is written back in a VISSIM-Python co-simulation platform to evaluate its effectiveness. Results show a 6.8% reduction in average travel time, an 8.1% reduction in average delay, shorter weaving areas, and reduced acceleration fluctuation. Additionally, safety analysis reveals a 27.8% decrease in TTC variance, a 22.6% increase in mean PET, and a 21.8% decrease in PET variance. Emissions of NOx, HC, CO, and CO2 are reduced by 5.3%, 5.5%, 4.5%, and 3.3%, respectively. These findings demonstrate that the proposed model simultaneously improves travel efficiency, traffic safety and decarbonization, offering a quantitative and interpretable tool for CAV control at signalized intersections, supporting intelligent, sustainable and resilient intersection management.