ANALYSIS OF LARGE-SCALE TRAFFIC INCIDENTS AND EN ROUTE DIVERSIONS DUE TO CONGESTION ON FREEWAYS

En route traffic diversions have been identified as one of the effective traffic operations strategies in traffic incident management. The employment of such traffic operations will help relieve the congestion, save travel time, as well as reduce energy use and tailpipe emissions. However, little attention has been paid to quantifying the benefits by deploying such traffic operations under large-scale traffic incident-induced congestion on freeways, specifically under the connected vehicle environment. New Connected and Automated Vehicle technology, known as “CAV”, has the potential to further increase the benefits by deploying en route traffic diversions. This dissertation research is intended to study the benefits of en route traffic diversion by analyzing large-scale incident-related characteristics, as well as optimizing the signal plans under the diversion framework. The dissertation contributes to the art of traffic incident management by 1) understanding the characteristics of large-scale traffic incidents, and 2) developing a framework under the CAV to study the benefits of en route diversions.Towards the end, 4 studies are linked together for the dissertation. The first study will be focusing on the analysis of the large-scale traffic incidents by using the traffic incident data collected on East Tennessee major roadways. Specifically, incident classification, incident duration prediction, as well as sequential real-time prediction are studied in detail. The second study mainly focuses on truck-involved crashes. By incorporating injury severity information into the incident duration analysis, the second study developed a bivariate analysis framework using a unique dataset created by matching an incident database and a crash database. Then, the third study estimates and evaluates the benefit of deploying the en route traffic diversion strategy under the large-scale traffic incident-induced congestion on freeways by using simulation models and incorporating the analysis outcomes from the other two studies. The last study optimizes the signal timing plans for two intersections, which generates some implications along the arterial corridor under connected vehicles environment to gain more benefits in terms of travel timing savings for the studies network in Knoxville, Tennessee. The implications of the findings (e.g. faster response of agencies to the large-scale incidents reduces the incident duration, penetration of CAVs in the traffic diversion operations further reduces traffic network system delay), as well as the potential applications, will be discussed in this dissertation study.