By Linzy Novotny
To the average driver timing green lights, a surefire way to improve fuel efficiency and reduce travel time, is an educated guessing game. It’s easy to learn the lights along your most frequented routes, and slow and accelerate accordingly, but once off the beaten path, you’re at the mercy of timers and traffic. Until now. Yes, there is an app for that (shudder).
Doctoral candidate Emmanouil Koukoumidis of Princeton and the Massachusetts Institute of Technology, along with the co-advisement of professors Li-Shiuan Peh of MIT and Margaret Martonosi of Princeton, have developed an app that helps a driver hit green lights instead of red, according to research Koukoumidis conducted.
“I’ve tested SignalGuru extensively in Singapore and Boston,” Koukoumidis said. “Some other researchers tested it as well and were very excited about it.”
How Does Signal Guru Predict Red Traffic Lights?
The SignalGuru utilizes green light optimal speed advisory (GLOSA) technology to detect the speed a vehicle approaching a red traffic signal should travel to avoid coming to a stop at the signal, but rather, roll up to the signal as it turns green. The program displays the amount of time until the immediate traffic signal will turn green, the amount of time until the signal will turn red again and the speed the vehicle should travel according to GLOSA. The smartphone needs to be mounted onto the vehicle’s windshield so that photos can be taken by the phone’s camera in order for a series of steps to be taken by the program to determine GLOSA.
“It is a collaborative system,” Koukoumidis said. “If some other smartphone has done all that then it can provide your smartphone with the necessary information, [which is the] predicted time that the specific light will turn green.” This is especially useful when determining the signal schedule for traffic-adaptive signals.
One obstacle Koukoumidis had to overcome was determining when pre-timed and traffic-adaptive signals would change from green to red and vice versa. Pre-timed signal intervals can be looked up in a database, while intervals of traffic-adaptive signals are predicted based on the history of settings using machine learning. This means that over time the photos taken by the phone can determine the schedule of traffic-adaptive signals and the SignalGuru can gather data about the traffic light from other smartphones that have already collected the information.
According to Koukoumidis’ research, the SignalGuru comes within 0.66 seconds of the time lights will turn green for pre-timed signals and within 2.45 seconds for traffic-adaptive signals.
Koukoumidis compared the costs of the SignalGuru to alternative ways of determining traffic scheduling, but only one other device is able to determine the predictability, offer continuous advisory and advance advisory, just as the SignalGuru does. The Audi Travolution meets these needs, but comes at the cost of owning an Audi.
When Will Signal Guru’s Green Light Giving Traffic Software Be Available?
Although the SignalGuru shows a lot of promise, “SignalGuru is still in the prototype stage,” Koukoumidis cautions. “We are looking into making it available commercially as we have received very good traction so far. Many users ranging from regular U.S. drivers to truck drivers from Australia have already contacted us and requested information about downloading the SignalGuru application.”
“We are definitely interested in making it available and we will try to do that as soon as possible,” Koukoumidis said.