Remember when our autonomous vehicles needed to stop for red lights? What a waste of time.
University researchers are creating a way for self-driving vehicles, now under heavy development by several companies, to zip effortlessly through intersections with nary a brake applied. Their idea is to build computerized traffic managers that would sit at crossroads like electronic air traffic controllers and direct approaching vehicles.
“If we put a communications device in these vehicles that can talk to our system, then we don’t have to use traffic lights anymore,” Hesham Rakha, an engineering professor and the director of the Center for Sustainable Mobility at Virginia Tech’s Transportation Institute, tells Txchnologist. “Just imagine—no more need to wait for yellow or red lights.”
One system devised at Virginia Tech is called iCACC, short for Intersection management using Cooperative Adaptive Cruise Control. Another version, created by University of Texas at Austin computer science researchers, is called Autonomous Intersection Management (AIM).
Simulations run by Rakha’s research team have found that employing iCACC to direct traffic can reduce total delays at intersections by up to 90 percent and diminish fuel consumption by as much as 45 percent.
The Virginia Tech team’s latest findings on iCACC are being presented at the 92nd Transportation Research Board Annual Meeting, which is wrapping up today.
Logic at the intersection
In concept, iCACC would be a unit installed at intersections and roundabouts that receives location, speed and acceleration data from autonomous vehicles when they are within 600 feet of it. Analyzing the same data for other vehicles within range along with current road and weather conditions, the iCACC controller would determine the optimum strategy for getting all the vehicles through the intersection as efficiently as possible.
(A screen shot from Virginia Tech’s iCACC used for simulating autonomous vehicles traversing an intersection. Courtesy Hesham Rakha.)
Assuming that self-driving vehicles will at least initially operate alongside those piloted by people, Rakha says that cameras installed at the intersection could monitor unexpected changes in speed and direction by human drivers.
“All of the instructions from iCACC to the vehicles are happening at the speed of computer processing, so if a driver takes over and doesn’t follow instructions, the system can handle that,” he says. “A person looks only at themselves and their own actions, but our controller constantly looks at the whole situation of the intersection zone.”
A major safety improvement, one day
Like autonomous vehicles, which offer an alternative to accidents caused by distraction, intoxication and other human errors, safe control of vehicles barreling into intersections could mean a major step forward for public health.
According to a National Traffic Highway Safety Administration report, about 40 percent of the 5.8 million crashes across the U.S. in 2008 occurred at intersections. That’s a huge number considering the small proportion of the entire roadway system that intersections represent.
The system could also tightly regulate speeds to avoid conflicts and allow vehicles to zip through at nearly full speed. It could prioritize movement through intersections for certain classes of vehicle. First responders? Full speed from start to destination when they’re answering a call. Slower moving public busses? Empty HOV lane and all others stop when they’re coming through.
“We’re looking at optimizing all movements for all vehicles coming into the intersection—preventing collisions while at the same time minimizing the total amount of slowdowns,” Rakha says.
So far, the Virginia Tech team’s work has been limited to advanced simulations, but they are hoping to run field tests in the next several months to get real-world data. Unfortunately, they have no autonomous test cars, so human assistants behind the wheel will execute instructions in real time from the iCACC program. “It won’t be the same as an automated vehicle that would perform the instructions by the letter because with people doing it, they might introduce errors,” he says.
The Texas group does have their own robotic car and they’ve been running AIM through trials. The video below shows their work.
Robotic cars becoming reality
The two teams’ efforts are part of a quickly expanding field of autonomous driving research and development being done by companies and academics. Car manufacturers like Mercedes, Audi, Cadillac and others are testing robotic technologies. The golden date seems to be around the middle of this decade when consumers will start to see autonomous driving options for certain situations, like moving through slow-moving traffic and parking.
During this year’s CES, Audi showed off a robotic function that will park a vehicle for an owner and a traffic jam assistant that will take over driving in congestion where vehicles are moving at speeds less than 37 mph. At the same show, Toyota showed off a semi-autonomous Lexus prototype it has been working on. These efforts and others join Google’s advanced project to bring driverless cars onto the market.
So far, legislation has been passed in Nevada and California that allow robotic cars on their roads and other states are considering similar moves.
“No one is going to be throwing away their steering wheel anytime soon,” says Rakha, “but it seems we’ll see autonomous vehicles on the road in the future. It’s a reality whether we like it or not.”
Top Image: Marvin, the University of Texas at Austin’s autonomous car, is putting the Autonomous Intersection Management system through its paces. Courtesy Tsz-Chiu Au/University of Texas.