Google is convinced that traffic lights are malfunctioning. And he has a new system to prove it

Why does it take so long for that traffic light to turn green? Traffic in urban centers can exhaust anyone’s patience, and although traffic lights are there to avoid chaos, it is inevitable to think that they could work better. That’s just what they are trying to prove at Google.

Project Green Light. The problem of traffic light synchronization has been debated for some time, and Google set out to try to improve the situation. A year ago they launched the so-called Project Green Light, an initiative that uses data from drivers and cars to optimize the operation of traffic lights.

We spend too much time at the red light. A 2021 study by the company Inrix cited in WSJ reveals that in the United States, drivers spend 10% of their short trips stopped at traffic lights. This causes not only possible anger, but also an increase in emissions: at intersections, pollution is up to 29% higher than on interurban roads.

Google Maps as a data source. But Google believes they can improve this situation thanks to the data of users who use applications like Google Maps. Researchers at the University of Michigan did something similar with data from General Motors cars to do a small experiment in Detroit. They evaluated the synchronization of 34 traffic lights in 2022, and although the changes suggested by the data were not huge, they had a notable impact on traffic flow, especially at peak times.

A cheap way to improve things. Making modifications to traffic light networks can be very expensive. Traditionally, vehicle count studies were commissioned at a given intersection, and each such study costs, according to WSJ, about $5,000. Modernizing an intersection can cost around $250,000, and maintenance adds even more costs. Google’s proposal can offer a much more affordable alternative.

Dynamic patterns. Most traffic lights work according to a fixed pattern that is difficult to customize. At most they have a configuration for peak hours and normal hours, and the ideal would be to have completely dynamic systems that adjust to the traffic at all times. That challenge is too complex, but Project Greenlight can at least help optimize the functioning of these patterns.

How Project Greenlight works. This system makes an information panel available to traffic agencies where they can see recommendations to improve the configuration of traffic lights at various intersections. In addition, Google’s system makes use of AI mechanisms to evaluate how traffic patterns and their flow change after the changes, which allows us to analyze what impact these changes have had.

Promising results. The project has already been launched in the form of small pilots in cities such as Abu Dhabi, Hamburg, Seattle and Kolkata. The data reveal that thanks to it, stops at traffic lights are reduced by 30%, a certainly notable figure that helps improve traffic flow, reduce emissions and of course, also avoid driver anxiety.

An expanding project. The official Project Greenlight website offers a contact form for traffic agencies. There is a waiting list, but the initiative is expanding to new cities and researchers say that any startup could do something similar. The methodology is publicly available, they say, and anyone is free to take advantage of it. As Henry Liu, head of the University of Michigan study, explained, “the mathematics was already there. The beauty of our solution is that the previous models were too complicated, and we have simplified them.”

In Xataka | Spanish traffic lights are more dangerous than the rest due to a peculiarity. And the DGT knows it

Image | Eliobed Suarez