Improved Time-to-Collision Definition for Simulating Traffic Conflicts on Truck-Only Infrastructure

Abstract

Transportation agencies have traditionally relied on historical crash records as the primary measure to evaluate the safety of roadways. The infrequent and sporadic occurrence of accidents and the long period required to collect accident data have led to the use of surrogate safety measures. The use of microsimulation modeling for conflict analysis has been popularized for evaluating experimental changes to existing road networks. Previous freeway studies have used a simplified time-to-collision definition, which produces unrealistic conflict situations. The definition included situations with no collision path, such as when two vehicles were traveling at the same speed or when the leading vehicle was speeding away from the following vehicle. A revised conflict definition is developed to address these issues and is then contrasted with the simplified definition used in earlier studies. An investigation of acceleration rates demonstrates that the revised approach retains the meaningful conflicts produced by the previous definition but eliminates the situations that are unlikely to be conflicts. This revised conflict definition is used to investigate the evaluation of a truck-only highway in the greater Toronto, Ontario, Canada, area to observe the effects on traffic conflicts. In general, it was found that although providing a separate highway for trucks did reduce truck-related conflicts, car lane-change conflicts increased because of the cars' increased maneuverability and presence on the truck-free highway.