Thanks to the blogosphere hype machine, most people associate automated, driverless vehicles with the cute, self-driving Google car. Google’s technology is charming, and suggests an idyllic morning commute in which we’re all chauffeured to work by robots. But the future of driverless vehicles is much more mundane.
Trucks. The future of driverless driving is all about trucks. So forget about that sensor-equipped Volkswagen Passat, and get ready for a 40-ton Peterbuilt 18-wheeler.
In the wake of new U.S. Department of Transportation guidelines for automated-vehicle testing, experts such as Ümit Özguner, a professor with Ohio State University’s Center for Automotive Research, are predicting that the first wave of operational autonomous vehicles will be devoted to long-haul deliveries. This is about moving cargo, not people.
“The trucking industry is very interested in going from single trucks to convoys of trucks. One human driver with perhaps three other trucks behind it,” Özguner told TechHive. “Those three wouldn’t necessarily have a driver in them. Eventually you could imagine removing the first driver too.”
The technology, which is being developed (and in some instances, deployed) for use right now, consists of a lead truck operated by a human driver followed in close formation by a small fleet of driverless vehicles “tethered” by a series of sensors.
Big in Europe
In late May, the U.S. Department of Transportation opened the door for road-testing self-driving trucks by publishing policies for the three states that currently allow driverless vehicles: California, Nevada, and Florida. These guidelines open the door for driverless tests of all kinds. Overseas, meanwhile, experiments with driverless truck convoys—also known as “platoons”—have been under way for years.
For example, the E.U.-sponsored Safe Road Trains for the Environment (or “SARTRE”) program ran from 2009 to 2012, and employed a mixture of radar, fixed lasers, and cameras to create a platoon of highway vehicles featuring one lead vehicle—typically a truck—with a series of cars driving behind in close formation. The follow vehicles operated completely under the control of the lead truck, allowing drivers in the rear cars to sit back and enjoy the ride, completely hands-free.
SARTRE was created specifically with both trucks and cars in mind, but one of its principal architects sees greater feasibility in automated trucking.
“[Long-haul trucking] is the most realistic starting point for the commercial adoption of the technology. The long-haul vehicles have the most to gain, both in terms of safety and economic benefits,” says Mike Baker, the chief engineer at Ricardo UK Ltd, the lead firm of SARTRE. “The fuel savings witnessed by trucks in a platoon has a significant impact on the operating profits of the operator, not to mention the environmental impact of reduced CO2 and emissions.”
Big in Japan
A similar Japanese caravan program is researching methods to use computers in choreographing a fleet of driverless trucks in close proximity. Similar to the way competitive bicyclists drive in close formation to create a slipstream of lower air resistance, the tailgating trucks would work together to reduce fuel consumption.
Although the Japanese study is still in progress and has yet to demonstrate significant reductions in fuel consumption, Baker says that SARTRE has already produced reliable fuel savings.
As gas prices continue to fluctuate on a grand scale, keeping fuel costs down is a major concern of any business that relies on over-the-road vehicles to function. However, the more immediate economic opportunity of connected caravans will be found in supplementing (or completely replacing) human workers—along with all those associated annoyances such as offering decent pay, providing healthcare benefits, and making sure drivers get enough sleep.
Highway to automation
Driving in urban or suburban areas is riddled with variables. While self-driving cars have demonstrated the ability to successfully operate among all the stoplights and changing traffic patterns our roadways can throw at them, they haven’t yet proven their ability to operate among erratic, emotional, and occasionally undisciplined human drivers. The same holds true for driverless convoys, which would likely need to pick up a human operator before exiting the relatively simple interstate highway system and entering congested urban centers.
In the future, cars will be driving themselves through all sorts of environments, “but that’s quite a ways off,” says Dan Flores, General Motors’ advanced-technology spokesperson. “We’re [developing semiautonomous technologies] only with highways in mind because you have the ability to stay in one lane for a long period of time. There’s a lot more predictability there.”
In the coming years, car manufacturers will continue to introduce incremental, semiautomated technologies designed for situations that don’t require interacting with too many other drivers (examples already include automated parallel-parking systems and cruise control). Flores told TechHive that before the end of the decade, GM hopes to release a technology called Super Cruise that will marry adaptive cruise and lane controls, and will allow any vehicle to safely navigate itself over long stretches of highway.
Highways are the ideal setting for all types of automated vehicles, while driving among humans in stop-and-go urban traffic is a different story altogether.
“Humans willfully disobey the rules at times. The mixture of automated vehicles and humans brings in all sorts of problems in regards to predicting what humans will do—going through red lights, not stopping at the stop sign, and so on,” says Özguner. “[Connected convoys on a highway] are [one] of the safest scenarios you can think of with autonomous cars driving along with lots of human drivers.”
Aside from all the safety issues relating to automated cars in congested urban centers, the military-grade technology that Google is using remains prohibitively expensive. And even if it were to cost less over time, a number of legal issues have yet to be sorted out. For example, who pays if a self-automated car causes an accident—the hardware manufacturer, the driver’s insurance company, or the driver himself?
Trucking is a dirty job
On the surface, driverless trucks might appear to be just another way for technology to rob human beings of jobs. But it turns out that trucking might be one of those dirty jobs that are best handled by a machine.
There’s a curious and severe shortage of truckers in the United States. According to the American Trucking Association, 2012 saw a 98 percent turnover for long-haul truckers, and despite the nation’s stubbornly high unemployment rate, the industry remains short of qualified drivers by more than 25,000.
“It’s not seen as a desirable lifestyle for a lot of obvious reasons. Leave home on Monday, don’t come home until Saturday—that’s not fun,” says Richard Wallace, the director of transportation systems analysis for the Center for Automotive Research (unrelated to OSU’s C.A.R. program). “If we could use advanced technologies so one driver could handle a platoon of trucks with a little automation, there’s a lot we could do to increase the quality of life in that industry, as well as increase safety,” Wallace says.
While the solitude of the highway may be a decreasingly desirable career choice for many Americans, far more precarious working situations have catapulted driverless trucks from the theoretical realm into the practical.
Trucks at war
Driverless convoys are already being used in war zones, where supply trucks are subject to attack. Robotics firm ASI has devised a driverless convoy system that uses sensors in combination with a physical Kevlar leash. A single truck leads a small battalion of unmanned vehicles behind it, eliminating the need for human drivers in a dangerous environment.
Similarly, mining operations in Chile and western Australia are using driverless dump trucks to transport ore and other materials, allowing machines to take over in remote locations where few human workers are willing to go.
Media coverage of driverless vehicles does a great job of explaining current technology, but hasn’t done enough to reality-check what we’ll actually be using in the next two, five, or even ten years. So, left to our own imaginations, we envision a future where we simply utter, “Take us to the movies, Mr. Car,” and off we vroom, with drinks in hand. But the unromantic truth is that such technology won’t hit the roads for at least the next two or three presidential administrations—if by then.
And until then? Keep your eyes on the road, and mind the automated convoys.
This story, "The first driverless cars will actually be a bunch of trucks" was originally published by TechHive.