.Published: Saturday, May 11, 2020
The humble traffic lights, which have been controlling the flow of traffic in cars and trucks for over a century, could be about to undergo a major change as vehicles become smarter and connected.
Researchers are looking at ways to improve traffic safety and efficiency by using features of modern cars such as GPS. The upgrades may eventually eliminate the current red, yellow, and green traffic lights, giving control over to driverless vehicles.
Henry Liu is a professor of civil engineering at the University of Michigan who is leading the study. He said that the implementation of a new system of traffic signals could be closer than most people think.
He said, “I think artificial intelligence is coming.”
The U.S. hasn’t seen many changes in traffic lights over the years. Megan Kate Nelson, a historian for Smithsonian Magazine, wrote that Cleveland introduced what was considered the “first municipal traffic control system” back in 1914. Engineer James Hodge’s invention was powered by electricity from the city’s trolley line. It featured two lights, red and green – colors that railroads have long used. The police officer in the booth had to flip the switch from his position on the sidewalk to change the light.
William Potts, a Detroit policeman, is credited for adding the yellow traffic light a few years later. However, as a city worker, he was not able to patent it. Nelson wrote that by 1930, most major American cities, as well as many smaller ones, had at least one electric traffic signal.
Traffic signals have been given a whole new dimension with the advent of automated and connected vehicles.
Ali Hajbabaie is an associate professor of engineering at North Carolina State University. His team has been rethinking traffic flow. Hajbabaie proposes adding a fourth traffic light, perhaps white, to signal when there are enough autonomous vehicles on the road.
Hajbabaie’s team, which used small model cars, explained that they would stop at an intersection if the light was red, and proceed if green. If the white light is on, then you simply follow the vehicle ahead of you.
He said that while Hajbabaie refers to “white phase”, and perhaps even a light of white, it is not important what color the light is. Modifying the current lights to flash both red and green at once could suffice to indicate that driverless cars have taken control. It’s important to make sure it is universally accepted, just like current signals.
Hajbabaie admitted that such an approach is years away. It would take 40 to 50 percent of the vehicles on the roads to be autonomous in order for it to work.
Sandy Karp, a spokesperson for Waymo, pointed out that Google’s self-driving subsidiary launched a ride-sharing service with full autonomy in Los Angeles and Austin (Texas) even without adding a fourth traffic signal.
Karp, in an email sent to The Associated Press, said that while it’s good that people are thinking about ways to make AVs safer and more reliable at this stage of development, they should not rush into AV-specific investments which may be premature or unnecessary.
Researchers at the University of Michigan have adopted a new approach. The researchers conducted a pilot in Birmingham, a Detroit suburb. They used the data from General Motors’ vehicles on speed and location to change the timing of the traffic lights in that city. Researchers recently received a grant from the U.S. Department of Transportation under the bipartisan Infrastructure Law to test the real-time implementation of the changes.
The Michigan research is more likely to be implemented in a wider range of vehicles than Hajbabaie’s goal, because it deals with vehicles which have drivers and not those that are fully autonomous.
Liu, the Michigan researcher who is leading the Michigan project, says that even if only 6% of vehicles in Birmingham are connected to the GM System, the data provided will be enough to change the timing of traffic lights and smooth out the flow.
Birmingham’s 34 traffic lights were selected because, as with more than half the signals across the country, they are set on a fixed schedule and lack any cameras or sensors that monitor congestion. Liu stated that although there are more sophisticated solutions for monitoring traffic, these require expensive and complex upgrades.
Liu stated, “The beauty is that you don’t need to change the infrastructure.” “The data does not come from the infrastructure.” The data is coming from car manufacturers.”
Danielle Deneau is the director of traffic safety for the Road Commission of Oakland County in Michigan. She said that the initial data from Birmingham adjusted the timing of the green lights only by a few second, but this was enough to reduce the congestion. The new grant-funded study could bring even bigger changes, automating traffic lights at a location to be announced in the county.
Source: ABC News