Comparative Evaluation of Benefits from Traffic Signal Retiming and Green Light Optimized Speed Advisory Systems

摘要

One of the ways to reduce excessive stop-and-go driving on urban streets is to optimize signal timings. More recently new methods in traffic signal optimization have incorporated changes in drivers’ behavior to achieve optimum performance at signalized intersections. One such application is called GLOSA - Green Light Optimized Speed Advisory. GLOSA is a method that uses traffic signal information to provide drivers (through infrastructure-to-vehicle communication) with speed advice for a more uniform commute with less stopping time through traffic signals. Recent results showed that a GLOSA system works efficiently only if exact durations of signal phases are known. Objective of this paper is to further evaluate performance of a GLOSA system in real-world-like conditions. Could GLOSA and similar speed-advisory methods replace a need for retiming traffic signals? This paper attempts to answer this question by applying a GLOSA approach on a set of (optimal and suboptimal) signal timings from an urban corridor in Salt Lake City, UT. A VISSIM model of a 5-intersection corridor, calibrated and validated with field data, is used to test two types of signal timings: fixed and actuated. The field signal timings from the case-study corridor are optimized by VISGAOST, a Genetic Algorithm Stochastic Optimization tool based on VISSIM evaluations. The results suggest that the GLOSA does not have an equal effect on traffic with fixed-time and actuated-coordinated signal timings, where the latter ones are collected as averages from historic records and embedded into the GLOSA algorithm. If the phase durations are predictable, as with fixed-time signal timings, then GLOSA has a significantly positive effect on number of stops and fuel consumption. However, if accurate signal timings are not known then it is likely that GLOSA will not bring a positive impact on traffic performance.