Estimating the impacts of transit signal priority on intersection operations: A moving bottleneck approach

Abstract Transit Signal Priority (TSP) is a commonly used strategy to improve bus operations at signalized intersections. However, the impacts of TSP in mixed traffic environments has not been analytically studied. This is a challenging problem since car queues can slow down buses, while slow-moving buses can create bottlenecks for cars in mixed traffic. Furthermore, it is typically assumed that TSP is activated using bus information obtained from a fixed detector. The benefits of TSP to buses could be improved by using information from connected buses or even connected cars. To tackle these challenges, this paper models buses as moving bottlenecks, incorporating it into a kinematic wave theory (KWT) model. A dynamic programming (DP) algorithm is developed to evaluate the changes in delays to buses and cars caused by TSP using KWT and queuing theories considering the bus as a moving bottleneck. The proposed algorithm is utilized for sensitivity tests to determine the changes to car and bus delays as a result of TSP implementation as a function of the bus detector location, bus stop location and dwell duration, the existence of a downstream bottleneck, and bus detection technology. The detector location sensitivity analysis reveals that there exists an optimal (and different) bus detector location associated with each demand. The bus stop location and bus dwell duration sensitivity tests show that TSP implementation can reduce system-wide (i.e., total car and bus) delays. However, in general, it is found that the presence of a downstream bottleneck can diminish the benefits of providing TSP. Finally, different bus detection technologies are tested to quantify the benefits of using connected buses or connected cars for TSP provision in terms of car and bus delay savings. As a result of this test, it is found that connected cars can significantly improve both car and bus delays if used for TSP provision.