Dilemma Zone Identification and Mitigation Approach at Signalized Intersections under Mixed Traffic Conditions Using UAV Data

Abstract

Traffic safety at signalized intersections is significantly affected by the driver’s behavior, psychology, and the design of traffic signals, especially in mixed traffic conditions. At signalized intersections, the traffic signal design influences the traffic behavior during the yellow change interval. The indecisiveness or the dilemma in decision-making during the yellow change interval often results in hard deceleration or red-light running, which concerns the safety of the road users in the form of rear-end and side-angled conflicts. Thus, it is essential to understand how traffic behavior and signal design changes affect safety. The present work investigates traffic behavior at signalized intersections operating under mixed traffic conditions using high-quality vehicular trajectory data obtained using unmanned aerial vehicles (UAVs). The analyzed traffic behavior is related to Type I and Type II dilemma zones at intersections with two types of traffic flow conditions: oversaturated and undersaturated. However, it is found that the position of these dilemma zones is near the stop line for an oversaturated intersection and away from the stop line for an undersaturated intersection, as it depends on the approach speeds of vehicles. The driver’s decision-making behavior, which depends on the driver’s psychology, is also analyzed, and a binary logistic regression model is developed for determining the Type II dilemma zone by considering various parameters at the onset of the yellow change interval. Further, the variation in the length of the Type I dilemma zone is studied by performing a sensitivity analysis and proposing measures for mitigating its negative effects. The findings underscore the critical importance of considering both traffic behavior and signal design in the safety analysis of signalized intersections, particularly in mixed traffic environments where driver behavior, psychology, and vehicle interactions are more unpredictable. The findings from this study have significant practical applications for enhancing traffic safety at signalized intersections, especially in mixed traffic conditions like those in India. Utilizing high-quality trajectory data collected via UAVs, this research precisely identifies dilemma zones and comprehensively analyzes driver behavior during the yellow change interval. One key application is in optimizing traffic signal design and timing. By accurately locating Type I and Type II dilemma zones, traffic engineers can adjust signal timings and phase durations to minimize hard decelerations and red-light running, thereby reducing rear-end and side-angle collisions. The development of a binary logistic regression model for predicting Type II dilemma zones can be integrated into intelligent traffic management systems. This model allows for dynamic signal timing adjustments based on real-time traffic conditions, improving the adaptability and efficiency of traffic signals. Additionally, the proposed mitigation measures, such as adjusting the yellow change interval length and implementing speed control measures, can be directly applied by traffic authorities to enhance intersection safety, particularly in mixed traffic conditions.