Estimating Traffic Conflict by Incorporating the Heterogeneity in Indian Traffic Condition Using Real-World Trajectory Data

Signalized intersections have the second-highest number of accidents after un-signalized intersections according to the Road Accidents in India (2022) report by Ministry of Road Transport and Highways of India. Safety analysis using accident data is a reactive approach. Traffic conflict techniques on the other hand are proactive approaches that identify observable critical vehicle interactions (conflict) that could have led to a crash. In this study, the traffic conflicts are identified by using a methodology that incorporates the heterogeneity and the disordered nature of the Indian traffic condition. The conflicts are estimated by modifying conventional Modified Time to Collision (MTTC) and Deceleration to Avoid Crash (DRAC) by incorporating vehicle heading direction, relative position, speed, and acceleration between vehicles to estimate rear-end and side-swipe conflicts. A threshold value suitable for disordered traffic conditions is estimated to segregate the conflicts and non-conflicting interactions. The variation in MTTC value with respect to the DRAC of vehicles showed that the lower MTTC values are obtained for motorized two-wheelers and motorized three-wheelers indicating that smaller vehicle types contribute to more critical vehicle interactions. A conflict Severity Index (SI) was developed with values closer to zero indicating a non-risk event and closer to 1 indicating a high-risk event. The temporal variation in SI was above 0.9 during the first one-third of red time for rear-end conflict and the first half of green time for side-swipe conflict.