Transport

Google's Green Light project retimes traffic lights for 30% fewer stops

Google's Green Light project retimes traffic lights for 30% fewer stops
Google says it's using Maps data to help cities optimize their traffic signals
Google says it's using Maps data to help cities optimize their traffic signals
View 2 Images
Google says it's using Maps data to help cities optimize their traffic signals
1/2
Google says it's using Maps data to help cities optimize their traffic signals
Traffic engineers receive simple timing change recommendations they can implement using existing systems
2/2
Traffic engineers receive simple timing change recommendations they can implement using existing systems

Through its Maps app, Google has access to a monster trove of daily traffic data – and it's using it to provide cities with free, AI-optimized traffic signal timing suggestions that could reduce stop/starts by 30% and intersection emissions by 10%.

Project Green Light has partnered thus far with 12 cities worldwide, to provide timing recommendations on 70 different intersections. It gives these recommendations free of charge, for now, via a little web interface, and the company says that rolling the changes out can be a five-minute job that uses the city's existing management systems.

Data is one key here; city traffic engineers simply don't have access to the breadth and quality of information that Google gets simply by virtue of having so many Android devices and in-car navigation systems online and constantly reporting location data.

Another is the AI tools and expertise available to the Google Research team, which has built a system that uses AI and Maps data to build models of individual intersections and their traffic signals and patterns, and then scales these models out to include other intersections in the same area, in order to best sync things up and keep cars moving.

Traffic engineers receive simple timing change recommendations they can implement using existing systems
Traffic engineers receive simple timing change recommendations they can implement using existing systems

The tool can model and analyze thousands of intersections simultaneously, to develop a city-wide picture of traffic flow that can be experimented on virtually, with a view to giving as many drivers as possible a "green wave" that'll cut down transport time, fuel burn and emissions.

"Early numbers," reads the project website, "indicate a potential for up to 30% reduction in stops and 10% reduction in greenhouse gas emissions." That's emissions at a given intersection, of course, not city-wide. But it's the kind of thing that might genuinely move the needle once rolled out at serious scale.

The first cities already experimenting with Project Green Light are Rio de Janeiro, Seattle, Hamburg, Bangalore, Haifa, Budapest, Kolkata, Abu Dhabi, Hyderabad, Manchester, Bali, and Jakarta. That's a pretty diverse range of driving conditions, and Google says the 70 intersections currently being trialed represent a chance to affect somewhere around 30 million monthly car trips.

"Green Light identified opportunities where we previously had no visibility, and directed engineers to where there were potential benefits in changing signal timings," says David Atkin, Analysis and Reporting Manager at Transport for Greater Manchester. "This provided valuable insights for our city with 2,400 traffic signals. Both the Green Light and Transport for Greater Manchester teams brought expertise and ideas to the table to improve journeys and reduce emissions."

Google is inviting other cities to sign up on a waitlist, with further cities to be added over time. Check out a video below.

How AI is fixing traffic lights | Project Green Light

Source: Google Research

5 comments
5 comments
paul314
As the world transitions to electric vehicles, the gains from this kind of thing (at least in terms of emissions) will be much less.
asdf asdf
google executives would of course have a speedy ride
asdf asdf
dissenters will be left in traffic
PAV
San Diego please sign up please sign up
BlueOak
“ Data is one key here; city traffic engineers simply don't have access to the breadth and quality of information that Google gets simply by virtue of having so many **Android devices**”

So are we to believe Apple iOS Google maps app users are spared from Google’s privacy issues? Ditto, Google’s Waze app on Apple devices?