Evaluating Left-Turn Conflicts under Permissive Phases Using Trajectory Data: A Comparison of Circular Green and Flashing Yellow Arrow

Abstract Left-turn maneuvers at signalized intersections pose significant safety challenges, particularly under permissive phasing, where drivers are required to judge gaps in opposing traffic. To assess the safety implications of novel signal technologies such as the flashing yellow arrow (FYA), this study examines left-turn conflicts using high-resolution vehicle trajectory data collected at a representative signalized intersection in Tucson, AZ. Specifically, we compare the traditional circular green (CG) indications and FYA within the permissive phase of protected-permissive left-turn signal operations. Safety performance functions are developed with postencroachment time (PET) serving as a surrogate safety measure to quantify traffic conflicts. Conflict frequency is modeled using zero-inflated negative binomial regression to address overdispersion and excess zeros in the data. The results indicate that FYA is associated with a statistically significant reduction in PET-based conflicts compared with CG. While lower left-turn speeds and higher conflicting volumes are consistently correlated with increased conflict frequency under both indications, the effect of opposing-vehicle speed differs between them. Specifically, higher opposing speeds are associated with more frequent conflicts under CG but fewer conflicts under FYA, reflecting distinct driver adaptations to the two signal indications. Behavioral analysis reveals that FYA is associated with milder acceleration and more moderate deceleration during left-turn maneuvers, suggesting more cautious driver behavior. These findings highlight the importance of incorporating signal indication type, geometric design, and traffic dynamics into left-turn safety evaluations. From a practical standpoint, the findings suggest potential safety benefits of FYA and offer actionable insights for traffic engineers aiming to improve intersection safety through signal design and operational refinements.