Next time you avoid sitting at a red light in Seattle, Google’s “Project Green Light” might have something to do with it.
The tech giant is working with 13 cities around the world — with Seattle being the first in the U.S. — to help optimize traffic lights and make traffic flow more efficiently. Google’s ultimate goal is to reduce greenhouse gas emissions from vehicles stuck in stop-and-go traffic.
The Google Research initiative relies on artificial intelligence and driving trends from Google Maps to model traffic patterns and make recommendations for optimizing existing traffic light plans. Tweaks can be applied to existing infrastructure in as little as five minutes, according to a Google blog post last month.
Google began working with Seattle in late 2022 as one of its earliest pilot cities. Juliet Rothenberg, a group product manager for Climate AI at Google, said the city was selected because of a familiarity with cutting-edge technology and the sophistication of the Seattle Department of Transportation’s traffic engineering department.
Google is working with SDOT by sharing recommendations for traffic light schedules. These recommendations are based on aggregated, anonymized driving trends.
Project Green Light has currently adjusted timing by 10 seconds at two intersections in the 15th Avenue Northwest corridor in Ballard, resulting in more streamlined traffic from Northwest 53rd Street to Northwest 65th Street and reduced stop-and-go traffic across five major intersections.
Stop-and-go traffic has also been improved at an intersection in the Greenwood neighborhood.
Prior to Project Green Light, cities would try to optimize traffic lights using expensive sensors or time-consuming manual vehicle counts. SDOT said it typically receives community feedback about signal timing along intersections or corridors, and then assigns engineers to investigate. The agency then uses a combination of traffic data, field observations and traffic modeling software to determine the best timing plan to remedy the situation.
Green Light is capable of analyzing thousands of intersections simultaneously, and SDOT said the system saves the agency time and gives it the ability to look citywide for timing issues and improve efficiency.
“The system combines many manual efforts into a simple, easy to implement process,” an SDOT spokesperson told GeekWire. “It allows us to be much more proactive in the way we discover and remedy signal timing issues and lets us perform routine optimizations more quickly.”
For now, Green Light timing adjustments are being used in Seattle alongside the more traditional process and data sources, serving as a supplement rather than a replacement.
“There is still an art to timing a complex signal system to efficiently serve people in cars, on bikes, in buses and on foot — having easy access to good data makes a big difference,” the spokesperson added.
After implementation of the timing improvements, Green Light has provided quick feedback on the effectiveness of the changes, according to SDOT. At one intersection, it was quickly identified that the timing change improved congestion for one direction of travel but did not have a benefit to overall intersection/corridor operations. With the data provided, SDOT was able to quickly make the decision to reverse that timing change.
The global transportation sector is responsible for a significant amount of the world’s greenhouse gas emissions. In the U.S., the Environmental Protection Agency reported that burning fossil fuels for cars, trucks and more generated 28% of U.S. emissions in 2021.
Google pointed at research that says pollution can be 29 times higher at city intersections than on open roads.
“We’re really pleased with the results we’ve seen so far leveraging Google AI,” Rothenberg said. “Early numbers indicate a potential for up to 30% reduction in stops and an average 10% reduction in estimated CO2 and pollution. This is just one example of the crucial role that AI can and must play in facilitating the transition to a lower-carbon world.”
