Proactive Freeway Crash Prevention Using Real-Time Traffic Control

This paper makes use of a probabilistic model, that predicts the likelihood of crashes (crash potential) on freeways on the basis of traffic flow conditions, in real-time crash prevention. The model was developed using incident logs and loop detector data collected over a 13-month period on the Gardiner Expressway in Toronto. The previous work suggested that an increase in levels of traffic turbulence generally yields high crash potential. Traffic turbulence was defined in terms of a series of crash precursors which represent traffic conditions that were present prior to crash occurrence. To apply the model in crash prevention, the link needs to be established between crash potential and real-time safety intervention. The objective of this paper is to explore this link for different thresholds of crash potential. The paper discusses the guidelines for evaluating the safety benefit of one crash prevention strategy (variable speed limits) and suggests the risk-based evaluation framework for real-time traffic control.