Listen 4 min
Share
Comment on this story Comment
Add to your saved stories
Save
At noon on a recent day, Virginia Tech researchers drove a 2022 Ford hybrid F-150 to a northern section of the 395 Express Lanes and put the truck’s automation mode to the test.
Equipped with sensors, cameras and a touch screen display on the passenger window, the truck traveled at highway speeds through a series of driving scenarios: evading debris, changing lanes out of the path of emergency vehicles and pulling over during a simulated stop by a state trooper.
Fast, informative and written just for locals. Get The 7 DMV newsletter in your inbox every weekday morning. ArrowRight
The researchers and an entourage of public safety, industry and highway officials have been using the stretch of toll road outside the nation’s capital this fall for self-driving demonstrations, hoping to develop solutions to common problems arising in the technology’s deployment. The simulations come three years into extensive research that has resulted in promising findings but has also come with moments of disappointment.
Advertisement
“These systems are coming relatively quickly, and we are trying to get ahead of them, showing how automation can overcome challenging problems,” said Zac Doerzaph, executive director of the Virginia Tech Transportation Institute, which received a $7.5 million federal grant in 2019 to conduct the research.
New habits are making more commutes miserable
Researchers are testing more than a dozen real-life scenarios focusing on how the automated driving systems (ADS) respond when confronted with public safety interactions, such as traffic stops, checkpoints, emergency vehicles and road hazards. The reversible toll lanes are hosting the simulations during the midday hour, when the road closes to shift lanes in the southbound direction.
During a recent round of demonstrations, the truck in automation mode responded smoothly to road hazards, moving to the left lane when vehicles were on the shoulder or debris was spotted. But it also encountered technical glitches that led to the computer system freezing up at a mock work zone, unable to drive between the orange cones.
Advertisement
Virginia Tech’s Mike Mollenhauer, who is leading the program, said it is a work in progress but essential to the future of the evolving technology. The research, which is expected to conclude next year, is looking at how technological and policy changes can overcome problems.
While automakers have made progress on self-driving technology, the innovation has not come without hurdles. In some cases outside the Washington region, vehicles have broken down in the middle of the roadway, blocking traffic, while in more extreme cases, they have blown past stop and traffic signs and caused crashes.
Mollenhauer said such incidents have set back progress, noting recent studies by AAA and Pew Research indicating about 25 percent of people say they would ride in an automated vehicle, down from about 45 percent a year ago.
Advertisement
“We’re kind of headed in the wrong direction,” he said, adding that he expects research can help find solutions.
Share this article Share
Data from Virginia Tech’s experiment will be available to the industry while also serving as a training ground for law enforcement and first responders on how to interact with vehicles when there is no driver.
Japjeev Kohli, vice president of technology for Transurban North America — which operates the 395 Express Lanes — said the collaboration among Virginia Tech, industry and local agencies has the potential to develop technology that could save lives. The key to the experiment, he said, is Transurban’s infrastructure, which was built with sophisticated detection systems to communicate with vehicles on the road and help them navigate the complexities they encounter, such as debris or vehicles on the shoulder.
Advertisement
As part of the program, Transurban developed software that digitally connects the vehicle and the infrastructure. The tool was designed to detect bottlenecks and to communicate with the vehicles through messages about speed and the time and distance between vehicles.
“We’re basically able to provide geolocated contextualizing insights in real time to these vehicles to allow them to make better decisions as they encounter these unexpected obstacles,” Kohli said.
The demonstrations, which are taking place again this week, involve about 13 vehicles to re-create scenarios orchestrated like a movie production on the highway.
Doerzaph said about 50 people have taken part in the experiment, including those from the research, government, industry and transportation sectors.
The goal, officials say, is to improve road safety.
Advertisement
Every year in the United States, more than 40,000 people are killed on roadways. Some transportation experts and researchers say that number could be reduced with better integration of automated driving technology.
“One of the things we know is that human errors are the largest contributor to crashes,” Doerzaph said. “And so if we can remove or reduce the human’s ability to make those errors now, we can really start to realize this worldwide effort to bring traffic fatalities down to zero.”
More coverage: Air travel, transit, railroads Air travel: Antitrust trial to determine fate of JetBlue-Spirit merger begins
Transit: Metro board says service cuts are real if funding doesn’t increase
Commuting: VRE plans new weekend service, fare increases next year
Safety: Two jets collide at Houston airport after one took off without clearance
FAA: Senate confirms new administrator, filling a role vacant for 18 months
Share
Comments
Loading...