About the Event
Intersection safety is a national, state, and local priority. In 2008, approximately 40 percent of crashes that occurred in the United States were intersection-related. However, current methods for assessing intersection safety require either a long history of crash data or subjective observation of traffic conflicts—both of which have significant drawbacks. Given the infrequent and random nature of crashes, crash-data-based methods are slow to reveal the need for improvement of either roadway design or signal control. Traditional conflict studies utilize trained personnel to identify and record conflicts observed at an intersection, so this method is time-consuming and expensive.
This presentation will highlight a new technique for evaluating intersection safety that requires neither crash history nor manual conflict survey—and overcomes the drawbacks of traditional approaches. The proposed approach utilizes the massive amount of high-resolution traffic signal event records (including both vehicle-detector actuations and signal phase changes) that has recently become available on a number of MnDOT intersections. The presentation will discuss how high-resolution traffic signal data can be used to estimate the number of rear-end and right-angle conflicts and to predict crash rate at an intersection. The proposed approach could be used to rank the safety performance for intersections within a jurisdiction in a proactive manner, thus helping agencies prioritize the intersections that need improvement.
Henry Liu is an associate professor in the Department of Civil Engineering. His primary research areas are traffic network monitoring, modeling, and control. His recent work has focused on traffic flow modeling and simulation, traffic signal control and optimization, traffic management under network disruptions, and equilibrium traffic assignment. In addition, his work on traffic signal data collection and performance measurement has been patented and licensed to a private firm for commercialization.