Traffic Conflict Prediction for Mixed CAV Traffic Flow on Highways

Aiming at addressing the safety evaluation challenges in mixed traffic flow with connected and autonomous vehicles (CAVs), this paper proposes a conflict identification model that integrates Time-to-Collision (TTC) and Safe Stopping Distance (SSD). A conflict probability formula is developed using logistic regression. Simulation experiments are carried out in a highway scenario implemented in SUMO. The results demonstrate that the proposed traffic conflict prediction model achieves an R2 value of 0.9386. Furthermore, the impact of CAVs on traffic flow is analyzed, revealing that when the CAV penetration rate exceeds 0.8, the reduction in traffic conflicts becomes significant, with the number of conflicts decreasing by more than 43.8%. This study offers a quantitative tool and a strategic foundation for the management of mixed traffic flow.