The timing of traffic signals can have a big impact on congestion and travel times for drivers. Currently, traffic signal control uses settings that change based on the time of day and typical traffic patterns, along with adaptive control, which adjusts signals using real-time data to handle unexpected changes in traffic from special events, crashes, or weather.
However, these methods work less well when traffic volumes are high. To remedy this, U of M researchers have continued their study of an improved approach to traffic signal timing known as “max-pressure” signal control. This new algorithm detects the number of cars in real time at one or more intersections to determine signal timing that will maximize throughput for an entire system.
“Unlike many adaptive traffic signal timing systems, max-pressure control uses one timing algorithm that responds to all traffic demand levels instead of separate algorithms for different time periods,” says CTS scholar Raphael Stern, assistant professor with the Department of Civil, Environmental, and Geo- Engineering (CEGE) and principal investigator for the project. “Rather than maximizing traffic volume one cycle at a time, this new algorithm results in maximum traffic flows through a corridor without compromising travel time for [vehicles on] the crossing streets.”
In the second phase of this project, CTS researchers brought max-pressure control one step closer to real-world deployment. In the project’s first phase, led by CEGE Associate Professor and CTS scholar Michael Levin, the research team ran microsimulations to investigate the use of max-pressure control on seven Hennepin County intersections. Results showed that it increased vehicle throughput and substantially decreased driver delay. These microsimulations, however, didn’t test the max-pressure algorithm on actual roads.
“Before we modified real-world traffic signal hardware for field tests, it was important to ensure the algorithm could operate safely in conjunction with Hennepin County’s traffic signal controllers and traffic detector inputs,” Stern says.
To safely test the max-pressure timing method, researchers designed a platform that allowed them to test the physical signal-control hardware used by Hennepin County connected to a simulated virtual environment. This simulated environment was based on an actual four-way Hennepin County intersection using the intersection’s geometry and traffic data.
During the tests, the max-pressure-control software determined the best signal timing based on current traffic and shared it with the physical traffic signal controller, which used a virtual display to show the current phasing. Then, the physical controller and corresponding display were looped back into the microsimulation environment, allowing vehicles in the simulation to “drive” according to the signal timing displayed by the physical controller.
The tests were successful, showing that max-pressure control can be safely and cost-effectively implemented on Hennepin County’s standard traffic signal controllers.
“This long-term research highlights the potential benefits of a max-pressure traffic signal control system,” says Ben Hao, Hennepin County traffic operations engineer. Moving forward, the team aims to further validate the system’s safety, functionality, and operation by deploying it in the real world, which will “gather valuable insights and lessons learned that will benefit Hennepin County and other public and private agencies,” Hao says.
The project’s results provide the confidence needed to plan a field test of the max-pressure signal-control system. Future research would include additional testing on the traffic signal test cabinet at the Hennepin County Traffic Management Center to ensure the system can accommodate operating conditions such as flashing yellow arrows, pedestrian signal calls, emergency signal preemption, and all-red flashing. If that were successful, researchers would implement max-pressure signal control on a three- to five-signal corridor in Hennepin County.
Both phases of the research were funded by the Minnesota Local Road Research Board (LRRB).
—Megan Tsai, contributing writer
