A conflict-based safety assessment technique for rear-end crash risk at signalized intersections in a lower-middle-income country: A comparison between homogeneous and heterogeneous traffic conditions

Abstract

Rear-end crashes at signalized intersections are attributed to different traffic behavior and operational characteristics. However, the influences of homogenous traffic conditions in developed countries and heterogeneous traffic conditions in low and middle-income developing countries like India are rarely explored. This study aims to examine the spatial–temporal characteristics of rear-end crash risk at signalized intersections characterized by lane-disciplined homogeneous and non-lane-based heterogeneous traffic conditions. In particular, this study has applied a conflict-based safety assessment framework to compare the rear-end crash risks across these traffic conditions. Vehicular trajectory data of vehicles at four signalized intersections, one in the homogeneous and three in heterogeneous traffic conditions, is analyzed. Rear-end conflicts position and time of occurrence in the upstream and downstream section of the stop-line are identified and modeled by extreme value models. Results show that the vehicles in heterogeneous traffic conditions are more aggressive, resulting in a higher number of conflicts and probable crashes. The presence of smaller-sized vehicles in dominating proportions and disregard of lane discipline with a higher degree of lateral movement are responsible for more traffic conflicts. The estimated extreme value models suggest that the minimum rear-end crash risk probability at signalized intersections with heterogeneous traffic conditions is about three times higher than that at signalized intersections with homogeneous traffic conditions. The observation is similar for crash risks estimated from both surrogate safety measures used in this study. The weak lane discipline and irregular lateral movement of vehicles are responsible for higher crash risks in heterogeneous traffic conditions.