Estimating intersection delays to transit vehicles using archived AVL/APC data

Many transit agencies have deployed automatic vehicle location (AVL) systems and/or automatic passenger counting (APC) systems on a portion of their fleet of transit vehicles. Data from these systems are used for real-time system monitoring and control and archived data are often used for service performance reporting and for service planning. This paper proposes and demonstrates a method by which these archived data can be used to estimate the mean and variance of transit vehicle delays caused by signalized and unsignalized intersections as well as the horizontal extent of the queue. These quantitative measures are of particular value for identifying intersections at which transit priority measures would be most useful. The proposed method is suitable for application to most transit AVL/APC databases and is demonstrated using data from Grand River Transit, the public transit service provider in the Region of Waterloo, Ontario Canada. These results suggest that the proposed method has practical application for the identification and prioritization of candidate transit priority measures, including transit signal priority (green extension, early green, special transit phases) and queue jump lanes.