Traffic lights are a necessary nuisance, but they're also incredibly bad for the environment. Idling vehicles in backed up traffic soak up fuel, and are a source of needless emissions. Even if the wait at a light isn't long, the acts of stopping and revving up to speed from a standstill are the parts of the driving cycle that use up the most fuel. People also tend to drive headlong into red lights, slam the brakes, and then wait. What if, instead, they knew exactly the speed to drive that would let them see green at every intersection.
An application called "SignalGuru," designed by researchers at Princeton and MIT, uses a network of smartphones mounted on car dashboards to estimate traffic light patterns to do just that. The application snaps photographs of an intersection from several approaching cars. With that information, the program estimates the time between light changes.
As the driver approaches a traffic light, the display on the app shows a suggested driving speed and lists the time (in seconds) until the light is due to turn green. The speed is calculated so that the car reaches the traffic light just as it turns green—if the driver plays it right and slows down accordingly, the can zip right through the intersection without coming to a full stop (the app makers decided to not allow it to tell you how fast to speed up to catch a soon-to-change green light). In a single stroke the app can decrease both emissions from cars as well as the amount of fuel they consume.
The app makers used iPhones and 20 cars to test the system in Cambridge, MA, and in Singapore. In the Cambridge test, SignalGuru got drivers to cut fuel use by 20%, and were accurate within a range of 0.66 seconds in predicting when traffic lights turned.
The test in Singapore was a little trickier—the traffic system there uses a traffic-flow sensitive light system, which varies how long traffic lights stay red and green. For now, the application is best suited to working with traffic lights which are on fixed schedules, which is how most traffic lights in the U.S. work.
Emmanouil Koukoumidis, one of the app's creators, anticipates that the version of the app that people can buy and use will probably have an audio feature, which will speak out instructions so drivers don't have to look at a screen. Also, combined with existing navigation software, SignalGuru 2.0 could also suggest alternate, intersection-free routes for the car to take. Either way, prepare to get a lot less used to seeing red while driving.
[Image: Flick user Magic Mazdik]