Transit signal priority along arterials using advanced detection

This research developed and tested the concept of advanced detection and cycle length adaptation as a strategy for providing priority for transit vehicles. In a departure from control strategies that rely on detection only a few seconds in advance of the stopline, a control algorithm was developed in which transit vehicles are detected two to three cycles in advance of their arrival at an intersection stopline, and phase lengths were then constrained so that the transit-serving phase was green for a 40-s predicted arrival window. Methods were developed for selecting whether to extend or compress phase lengths to shift a green period to cover the arrival window. Adaptive control was combined with actuated control using traffic density and queue length estimation, transit stopline actuation, and peer-to-peer communication for coordination in the peak travel direction. The method was applied by simulation to Boston, Massachusetts' Huntington Avenue corridor, which is served by a light-rail line running partly in mixed traffic and partly in a median reservation. The prediction/adaptation algorithm resulted in 82% of the trains arriving during the green phase. This control strategy resulted in substantial improvements to transit travel time and regularity with negligible impacts on private traffic and pedestrians, and was found to be more effective than simple preemption.