Improving traffic signal warrant systems by incorporating robust collision analyses
University of New Brunswick
Traffic signal warrants (TSWs) are typically the first tool used by traffic engineers to determine if a stop-controlled intersection should be signalized due to their shorthand, objective, and consistent analytical methods that estimate net operational benefits. The majority of Canadian jurisdictions rely on the TSW procedure published by the Transportation Association of Canada (TAC), and this procedure is notably different from most others due to its exclusion of any empirical safety measures. Changes in collision severity and frequency are two of the most significant impacts of signalization, so the lack of a collision history component in the TAC warrant procedure limits the robustness of any findings. This dissertation presents collision adjustment factors (CAFs) that allow crash history to be incorporated into the TAC warrant procedure. Developing the CAFs required novel statistical analyses of the severity/cost and frequency of collisions at stop-controlled and signalized intersections in North America, evaluation of the safety benefits of signalization, and conversion of those benefits into points used by the TAC warrant procedure. Each of the research phases are presented in separate chapters of this dissertation. There are several primary findings derived from this research. It was found that the intersection characteristics with the greatest influence on collision severity and cost were posted speed limit, land use (rural/urban), and the presence of divided approaches. In developing North American collision prediction models it was found that the uncalibrated models in the Highway Safety Manual were not a good representation of the average collision frequency at stop-controlled and signalized intersections. By combining the analyses of collision severity and frequency it was found that the majority of intersection configurations studied were not projected to exhibit a safety benefit from signalization. Lastly, the recommended method for converting the safety benefit/cost of signalization into TAC warrant points was through expert opinion on the valuation of collisions versus delays. Further areas for study include surveying practitioners to gain a more robust expert opinion on the relative values of collisions and delays for TSWs and replication of the statistical analyses in this dissertation using alternative data sources.