Arterial Queue Spillback Detection and Signal Control Based on Connected Vehicle Technology

Abstract

Queue spillbacks are a major problem in urban signalized arterials because such spillbacks can lead to gridlock and excessive delays. Several methods based on fixed-location detector data have been proposed to identify the occurrence of queue spillbacks and implement signal control strategies to mitigate their impacts. This paper presents two queue spillback detection methods based on connected vehicle (CV) or probe data. The first method requires only the use of CV data and is based on the notion that nonequipped vehicles in queue that arrive after the last CV-equipped vehicle can be modeled by using a geometric distribution. The second spillback detection method combines CV data with information about the signal settings at the upstream intersection and is based on a kinematic wave theory of traffic. The authors also developed a signal control strategy to mitigate queue spillbacks once they were detected. The proposed queue spillback detection methods and alternative signal control strategy were tested through simulation on a four-signal segment of San Pablo Avenue in Berkeley, California. The results show the penetration rate thresholds of CV-equipped vehicles required for accurate queue detection. The proposed signal control strategy improved traffic operations for the upstream cross streets without compromising traffic operations on either direction of the arterial traffic and substantially reduced the variation of the queue length on the critical arterial link.