🧠 What Happens When AI Controls Traffic Lights?

Sign Up | Home Page | Meet The Team

In today’s sunday special:

Read time: 7 minutes

📜 THE PRELUDE

You lean into the steering wheel as the car barely inches forward, met with a river of red brake lights that stretches endlessly ahead, stealing your time. The rising heat radiates from the asphalt, shimmering across the restless vibrations of hundreds of idling engines.

Finally, the traffic light turns green, transforming the suffocating standstill into a chaotic crawl. You ease off the brake and roll forward a few feet, only to settle back into stillness. The horns of angered drivers honk in impatient bursts as a motorcycle weaves recklessly between lanes.

According to the INRIX, outdated traffic light management contributes to at least 10% of total traffic delays nationwide. In the U.S., traffic lights are supposed to be retimed every 3 to 5 years to account for current traffic flow. In reality, most major U.S. cities do it far less often because retiming requires hiring expensive transportation engineers to collect current traffic data and simulate new car count scenarios that involve pedestrians, bicyclists, and transit.

Google Research recently launched “Project Green Light,” which harnesses AI to help you save time and money on the road by reducing stop-and-go traffic. So, how exactly does it achieve this without tedious manual car counts and pricey video detection cameras?

🚦 A BRIEF HISTORY OF TRAFFIC CONTROL

In the late 1800s, traffic constables relied on hand signals and brass whistles to guide the flow of horse-drawn carriages. But as urbanization surged, intersections became progressively congested. In 1866, about 1,102 people were killed, and around 1,334 people were injured on roadways throughout London, UK.

In 1868, 40-year-old railway superintendent J. P. Knight invented the first traffic light to manage the high-fatality roadways of Victorian London near the Palace of Westminster. It was a lantern-style gas lamp mounted with three mechanical arms. During the day, a police officer relied on a pulley system, raising the mechanical arms parallel to the ground to signal “stop,” and lowering them at a 45-deg. angle to signal “caution.” At night, the gas lamp was fitted with green and red glass panels to signal “go” and “stop.”

In the early 1900s, the U.S. automobile industry exploded, with the number of motorized vehicles rising from 8,000 to more than 8,000,000. As a result, intersections become increasingly dangerous, with traffic lights rapidly switching between red and green.

In 1920, 37-year-old Detroit policeman William Potts invented the first four-way, three-colored electric traffic light to provide a safe transitionary period between stopping and going. In 1949, the Geneva Convention on Road Traffic standardized color-coded traffic lights, making the “red-yellow-green” protocol universal.

By the 1970s, major U.S. cities began implementing “green waves.” This strategic initiative involved synchronizing traffic lights across intersections through centralized traffic control systems to create a continuous flow of green lights for motorized vehicles.

This strategic initiative led to the implementation of “fixed-time control,” in which traffic lights operated on pre-programmed cycles, switching phases (i.e., “green,” “yellow,” and “red”) at predetermined intervals. For decades, this method worked well, as traffic patterns were relatively predictable.

By the 1980s, major U.S. cities began embracing adaptive traffic management systems such as SCATS, which continuously collect real-time traffic data from ILDs: in-pavement metal detectors that utilize the magnetic field to sense motorized vehicles by measuring changes in the magnetic signal. SCATS leverages this real-time traffic data to dynamically adjust three things:

🏁 HOW AI REDUCES STOP-AND-GO TRAFFIC

Google Research recently launched “Project Green Light,” an AI-assisted traffic light optimization project that taps into real-time traffic data from Google Maps to enhance existing intersection attributes like green splits: the right-of-way time and order. In other words, Google Maps essentially turns a driver’s car into a “mobile sensor” to record the current flow of traffic instead of relying on traditional adaptive traffic management systems that require expensive roadside infrastructure.

⛽️ THE REAL-WORLD IMPACT?

It’s currently live in 18 cities across 4 continents, helping over 47 million monthly drivers experience up to 30% fewer traffic stops. It’s also cut the traffic light retiming costs per intersection by up to 80%, saving major U.S. cities at least $500 million every 3 to 5 years.

Boston’s traffic engineers have implemented the actionable recommendations suggested by “Project Green Light” at 114 intersections. As a result, unnecessary stops were reduced by 20%, and total stops fell by 33%, saving nearly 4,000 gal. of fuel annually, with Bostonians experiencing over $2,000,000 in annual gas savings.

🔑 KEY TAKEAWAY

Until recently, optimizing traffic lights required expensive roadside infrastructure. Now, AI can analyze billions of real-world drivers to simulate thousands of “signal-timing strategies,” helping major U.S. cities improve traffic flow and reduce fuel waste at virtually no cost to them.

📒 FINAL NOTE

{{rp_personalized_text}}

Share your unique referral link: {{rp_refer_url}}