Investigation on impact of vehicle types on right-turn crossing conflicts at unsignalized T-intersections using Generalized Poisson Regression Model

In India, crossing conflicts between right-turn vehicles from minor roads and conflicting vehicles on major roads have intensified at unsignalized T-intersections recently. The rapid increase in vehicular traffic, including motorized two-wheelers, auto-rickshaws, and others, has led to a rise in right-turn crossing conflicts (RTCC) at these intersections. The severity of RTCC is influenced by the characteristics of both the right-turning vehicles and the conflicting through vehicles. This study explores the impact of vehicle types on RTCC. Observations of RTCC were made using time-to-collision (TTC) at four un-signalized T-intersections across Patna city in India. The RTCC were categorized into critical and non-critical based on TTC threshold values determined by the k-means clustering algorithm. The Generalized Poisson regression model is an alternative approach to modeling equidispersion, over dispersion, or under dispersion in count data. The study revealed that a higher composition of two-wheelers, auto-rickshaws, and cars among both right-turning and conflicting through vehicles significantly influences the severity of RTCC. Additionally, the model results indicated that the speeds of right-turn and conflicting through vehicles, the volume of conflicting through traffic, right-turn traffic, vehicle gaps, waiting times, and abnormal driving paths significantly affect RTCC at un-signalized T-intersections. The findings of this study assist traffic engineers and safety experts in identifying critical unsignalized T-intersections by analyzing the number of right-turn crossing conflicts.