Kevin Curtis has worked road construction in New Jersey for years, but he still has mixed feelings about the industry.
Though the work is seasonal, when it arrives, it’s stable. The job helps him provide for his young son, and he’s met many close friends through the business.
Still, the hours are long and the labor is hard. A day on the job could entail anything from toiling over hot tar to wrangling a jackhammer, he said.
“Every day I come home, something new hurts,” Curtis said. “I go through so much aspirin I should buy stock in it.”
But, Curtis admits, those physical pains are the least of his concerns.
“There have definitely been some close calls—a little too close sometimes,” he said. “[Drivers] don’t see the cones in time, or they’re on their phone and don’t see you at all. You’re always aware of the possibility that someone might hit you.”
Curtis has reason to worry. The construction industry was the country’s most deadly in 2015, with 937 workers dying on the job, according to the Bureau of Labor Statistics’ most recent figures.
As a place where the standard rules of the road don’t apply, construction sites can be unpredictable and dangerous for drivers and workers alike. And soon another unknown—and potentially unsafe—factor will be unleashed upon them: self-driving cars.
The Roads Have Eyes
The problem stems from the way autonomous vehicles view the world.
To “see” where they’re going, self-driving cars rely upon a number of outlets, including sensors, cameras, and a series of maps and markers determined by their engineers and refined by their GPS. The information combines to create a schematic diagram that helps the vehicle understand when to stop, turn, merge, or yield, among other actions.
A construction zone, however, turns this ordered view of the world into chaos.
On- and off-ramps appearing clear to an autonomous auto could be blocked off. Lanes could merge, move, or disappear altogether. And even the most sophisticated technology could be fooled by the sudden precedent of a construction worker’s wave over a traffic signal.
Compounding the problem is the seeming randomness with which these sites occur. No national database tracks the location of ongoing construction zones, forcing computers to rely on information posted by a wide assortment of state and local agencies—which may or may not be keeping real-time updates of their progress.
No national database tracks the location of ongoing construction zones, forcing computers to rely on information posted by a wide assortment of state and local agencies—which may or may not be keeping real-time updates of their progress.
And even if the site’s location can be pinpointed, configurations within it could change at any time, Curtis said.
“There are times when we don’t know what the site will look like until we show up in the morning,” he explained. “Or we’ll finish one section, then move things around to work on another area.”
For their part, the Federal Highway Association (FHWA) is working on developing applications that will “broadcast information about work zones, such as lane shifts,” according to agency representative Donna Hannah.
The FHWA is “working with vehicle manufacturers and road owners and operators regarding how data from the infrastructure can be provided to vehicles, such as lane closure information, presence of workers, and queue development, to fuse with on-board vehicle sensor data to provide drivers—human or automated—with information on work zone status,” Hannah said.
Still, that type of effort and coordination could take years to perfect. In the meantime, motorists can step in to commandeer self-driving models that include steering wheels and pedals, but the pod-like prototypes being pushed by Silicon Valley will have no such backup option.
Instead, they’ll have to master another human skill: communication.
Reading the Signs
The key to imbuing our vehicles with such knowledge may, somewhat ironically, lie in the heart of a construction zone.
Michigan workers are busy installing experimental infrastructure along a 17-mile stretch of Interstate 75, which will become the nation’s first “smart highway” equipped with signs and other infrastructure that can “talk” directly to the vehicles riding along it.
The project utilizes “dedicated short-range communication” (DSRC) technology, which is similar to WiFi, allowing cars and the road to talk to each other. When complete, the corridor will be open for testing by any automaker—an option that will become essential in the future, as the government recently required all vehicles to have such communication abilities by 2020.
“From a state perspective, we find it vitally, vitally important,” said Michigan Department of Transportation (MDOT) representative Rob Morosi. “As vehicles become increasingly connected, the infrastructure must also be updated, not only for safety, but for reliability with this new technology.”
Construction on the first 2.5-mile segment should finish this September. It will include 15 signs, 10 barrels, and 10 “road side units” (RSUs) capable of sending messages directly to any connected car driving past them.
While instructions on the signs will be readable by humans, about half of the equipment will also include 2D barcodes—block-like symbols that look similar to the QR codes used to produce promotional deals when scanned. But instead of transferring coupons to a cell phone, the cyphers will translate the signs for the car.
“There’s no difference to John and Jane Q. Public with a non-connected vehicle,” Morosi said.
“They can still read the signs. It’s the same type of pavement markers that you’d find anywhere. The barrels look the same. The signs look the same. But as a human will pick up the words ‘Road Work Ahead,’ the car will pick up the barcode.”
Such easily designable and portable “smart signs” could help keep connected cars abreast of the current configuration of a construction site. But eventually, the technology may not just get our cars to go the right direction, but also the right speed through work zones, Morosi said.
With the help of the RSUs, which will relay information to both each other and passing vehicles, a connected car could be privy to the fact that it’s approaching a construction zone and prepare itself accordingly.
“The future goal would be to have the car slow itself down, without human assistance, when it receives a message like that,” Morosi said.
Taking the Scenic Route
Still, many automakers have promised self-driving cars on a tight schedule, giving engineers little time to develop such technology. Without the proper programming in place yet to steer their vehicles clear of such issues, some have come up with alternative plans to deal with work zones.
Nissan recently proposed a system called “Seamless Autonomous Mobility,” which would allow connected cars to send a signal alerting a human when it approached a construction site—or any other unexpected situation.
The operator could then use “teleoperating” technology to take control of the car temporarily, giving it various commands to navigate the scene by utilizing the vehicle’s suite of cameras and sensors, among other tools.
Still others have decided the best way to approach the problem is to avoid it all together.
If technology can be developed to allow vehicles to reliably detect upcoming work zones, an autonomous auto may simply be directed to circumnavigate the spot, like choosing a reroute option on a GPS.
Such a plan would take a dangerous factor out of the mix for construction workers, likely making their places of work safer. And while it could lead to other issues, like traffic jams on roads around the work zone, it may just give passengers of such future vehicles a rare chance to see a path less traveled—and remember why humans liked driving to begin with.