Detecting Signals of Bottleneck Activation for Freeway Operations and Control

Abstract

Any application of intelligent transportation systems (ITS) to freeway operations and control relies on understanding the behavior of freeway bottlenecks. By improving our understanding of the formation and dissipation of queues at active freeway bottlenecks (points characterized by the presence of queued traffic immediately upstream and unqueued traffic immediately downstream) we can improve the techniques used to manage freeway traffic—one of the hallmarks of ITS. It is shown that freeway bottlenecks' activations can be diagnosed definitively using archived inductive loop detector data (measured at 20- or 30-sec intervals) from two sites in Toronto, Canada. Once the bottlenecks' locations and times of activation were determined, potential signals of their activations were explored. It is shown that certain potential signals were evident immediately before upstream queue formation, and that these were reproducible from day to day (over three or four days examined) at the two sites studied. As a result, it is suggested that these signals be used for systematic investigation of bottleneck behavior in either real time or in retrospect, on a more widespread basis as part of a real-time ITS control system.