Most drivers have encountered this situation: you are waiting to turn left at a traffic light for what seems like minutes before the arrow finally turns green and allows you to make a protected left turn. A few times while waiting for the green arrow, a gap opens in oncoming traffic, and you wish that you could have gone, but alas, you still have a red arrow. At another intersection, rather than a red arrow, you encounter a flashing yellow arrow. Stopped at one of these arrows, you watch the gaps in the oncoming traffic and try to decide, “Is it safe for me to make this gap?”
How do drivers decide whether to turn or to wait? Many would say they use their intuition and experience. Traffic engineers, however, need to be able to quantify the conditions under which drivers can safely execute a permissive left turn – those turns that must yield to oncoming traffic – and when a protected left turn would be the safest option.
This challenge is something University of Central Florida (UCF) researcher Dr. Hatem Abou-Senna pondered. Eight years ago, Dr. Abou-Senna had been perusing the Manual of Uniform Traffic Control Devices (MUTCD), a resource well-known to transportation professionals, when he came across the concept of flashing yellow arrows (FYA).
History in a Flash
Flashing yellow arrows were first introduced in the United States over a decade before Dr. Abou-Senna began to pursue his research. Flashing yellow arrows were initially developed for safety purposes in the 1990s. While observing crash history, engineers noticed how some drivers assumed that a circular green signal meant they had the right-of-way to turn left across opposing lanes of traffic when this signal actually meant they needed to yield to oncoming traffic. The engineers needed a better way to communicate to drivers the difference between having the right-of-way and needing to yield. Thus, the flashing yellow arrow was developed to create a four-section signal head, complementing the steady green, yellow, and red arrows.
As flashing yellow arrows have increasingly been adopted for signaling permissive left turns, drivers have found them useful when navigating intersections. The concept is generally intuitive: the arrow points in the direction of travel, and the flashing yellow light cautions drivers to yield when making that movement. In explaining how this helps drivers, Dr. Abou-Senna describes the concept of fail-safe, saying “when people fail to understand what something means, their response is safe. The fail-safe percentage was very high for the flashing yellow arrow. Drivers take caution when they see something yellow and flashing.”
However, the problem of identifying which type of signal to use for left turns remained: when and where can traffic engineers use a flashing yellow arrow, letting drivers make decisions about taking a gap, and when should they use the green arrow to give drivers a protected movement with less potential conflict? Historically, engineers look at traffic volumes to understand at what intersections and during what times of day it was safe to allow a flashing yellow arrow – say, during the overnight hours. Dr. Abou-Senna wondered if traffic engineers could use real-time data to make the decision, thus developing the idea of a dynamic flashing yellow arrow that responds to real-time traffic conditions.
Dr. Abou-Senna worked with FDOT to understand how dynamic flashing yellow arrows may not only improve efficiency but also safety of intersections. Jim Stroz, FDOT District 5 Traffic Operations Engineer, says, “In transportation, there is often a scale of safety versus efficiency. Increased safety can result in a decrease in efficiency and traffic flow. On the other hand, an increase in efficiency could result in a decrease in safety. There really is a balance that we were able to understand with flashing yellow arrows through this research.”
Supporting Decisions: Real-Time Simulation and Field Testing
To begin their research, the research team identified factors for evaluating gaps. These factors – for example, the number of lanes crossing, how many oncoming vehicles are present, and how fast those vehicles are traveling – are what drivers intuitively consider when deciding whether to turn during a gap. Then they developed a computerized decision support system (DSS) to interactively evaluate when flashing yellow arrows could be activated. Based on the number of left turns requested and other parameters in the virtual system, the DSS provided a decision on whether a permissive left turn would be both effective and safe under the simulated conditions. The researchers then used the DSS to determine appropriate conditions for a flashing yellow arrow left turn using real-time traffic volumes.
The next phase of the research moved beyond simulation into field application of the DSS. The team created hardware to talk to the controllers that run traffic signals. But there was a challenge: in Florida, several types of traffic signal controllers made by various companies are used, so the hardware needed to be applicable across different controller types.
In order to check the permissive flashing yellow modes in the field, the team created a universal tool that could talk to various types of controllers and tested it at the UCF Intelligent Transportation System (ITS) lab. In total, six locations throughout Seminole, Orange, and Volusia counties were chosen for field application. Then, this tool was set up in the UCF ITS lab to mimic conditions as if it were installed and responding at these six intersections. The DSS used real-time data from in-field traffic signal detectors to adjust the signal phasing for the lab testing.
Following the tests, minimum gaps in traffic were calculated from the field data and used to analyze how often traffic conditions would warrant a flashing yellow left-turn arrow.
Expanding and Deploying: The Future of Dynamic Flashing Yellow Arrows
The in-field testing showed that dynamic flashing yellow arrows reduce both queues and delay for left turns. A dynamic flashing yellow arrow was found to be more efficient with longer cycle lengths – the time it takes for a full intersection to complete its signal phasing before repeating – than with shorter ones. Overall, the testing confirmed that the algorithm and logic developed throughout these projects actually works in the field and can optimize safety and efficiency at signals with four-section heads.
Even though the initial project has successfully progressed from research to field application, there is still more work to be done. “I never thought this research would evolve like it has, but FDOT became very interested in the dynamic flashing yellow arrow concept and how it can be applied to their systems,” says Dr. Abou-Senna.
Jim Stroz feels that future pilot testing will be even more beneficial to FDOT and the state: “Now we have the hardware, so the next step is to manufacture more DSS devices and deploy them to various locations.” Plans for this effort are to expand field testing and perform thorough before-and-after evaluation of intersections deployed with the dynamic flashing yellow arrow. While the research to date has focused on FDOT District 5, drivers across the state may soon be finding themselves moving more freely with flashing yellow arrows.
BDV24-977-21 Dynamic Flashing Yellow Arrow (FYA) – A Study on Variable Left Turn Mode Operational and Safety Impacts – Phase III – Developing a Decision Support System Hardware Platform
Final Report | Summary
BDV24-977-10 Dynamic Flashing Yellow Arrow (FYA) – A Study on Variable Left Turn Mode Operational and Safety Impacts – Phase II – Model Expansion and Testing
Final Report | Summary
BDK78-977-15 Dynamic Flashing Yellow Arrow (FYA) – A Study on Variable Left Turn Mode Operational and Safety Impacts
Final Report | Summary