Army, Car Makers Push Ahead With Driverless Vehicle Research
Japanese, German and U.S. car manufacturers are investing research and development funding into the same idea, but for everyday life.
Self-driving vehicles is a technology trend that has largely been out of the general public’s eye. But there are signs that this is about to change.
Already some of these capabilities are making it into luxury cars. Vehicles that can parallel park themselves are being advertised on TV. The Defense Advanced Research Projects Agency sponsored the DARPA Grand Challenge in 2007, where driverless cars navigated a closed course on a military base in California. The X Prize Foundation — with great fanfare — offered $10 million for suborbital space travel. It is now gearing up to offer a similar cash award for a driverless vehicle.
Google is investing money in the concept. Toyota, Volkswagen, General Motors and other vehicle manufacturers are doing so as well.
Along with the military, NASA funds the development of vehicles that can autonomously explore the surfaces of planets or moons in the solar system.
Meanwhile, the National Highway Traffic Safety Administration — part of the Department of Transportation — is taking a small step toward autonomy by experimenting with cars that warn each other of potential collisions.
“I think over the last couple of years we are starting to converge,” said John Augustine, managing director of intelligent transportation systems at the Department of Transportation. He has an annual R&D budget of about $110 million to spend on such initiatives. He recently traveled from Washington, D.C. to the Army’s Tank and Automotive Research and Development Center in Michigan, to see how his program and the Defense Department can leverage each other’s initiatives.
Defense and Transportation Department personnel, along with NASA and industry representatives, met at a joint Association for Unmanned Systems International-National Defense Industrial Association conference devoted to bringing the military, federal, and commercial “intelligent transportation” worlds together.
Both the Defense and Transportation Departments have a goal of saving lives. For the military, it wants to protect those who are driving supply trucks in hostile domains. For the last four years, it has invested in convoy active safety technology, or CAST, which allows hands-free driving.
As for the Transportation Department, it wants to sharply reduce the large numbers of victims who die on U.S. roads each year. About 6 million accidents occur in the United States annually, resulting in about 33,000 fatalities. The vast majority of these accidents are caused by driver error.
“Safety is our number one mission,” Augustine said. The death toll does not include serious injuries resulting from the car crashes, he noted.
Jose Gonzalez, director of land warfare and munitions in the office of the secretary of defense, said with the grand challenges, prize competitions and research and development dollars going into autonomous vehicles, the question is now: where is all this heading?
“Cultural acceptance is now our biggest challenge in DoD,” he said. Combatant commanders have seen the value of ground robots that assist explosive ordnance disposal teams dismantle roadside bombs, but the U.S. military as a whole has not moved beyond them.
The CAST system allows trucks to follow a leader vehicle and their drivers to keep their hands off wheels and their feet off brakes. It has been under development for a number of years, but has not made it into battle zones. One issue has been the price tag. TARDEC has been trying to bring the price to under $20,000 for each vehicle.
“It’s not just technology. It’s affordable technology,” Gonzalez said. “We can’t come in with high-price solutions.”
There needs to be a firm set of requirements from the users, and they have not been forthcoming, Gonzalez added.
Jim Overholt, senior research scientist for robotics at TARDEC, said the original idea for CAST was to lessen the number of personnel traveling in the convoys, and to therefore, reduce the number of potential casualties. Combatant commanders don’t seem to need that, though. They want the same number of troops in the convoys. However, they do see the value in letting the robotic system drive the trucks so soldiers can keep their eyes peeled for ambushes and improvised explosive devices.
“If you don’t have requirements, you’re not really going to get money to transition this technology,” Overholt said. He is spending a lot of time with those who write requirements to ensure the proper documents are produced.
The mine-resistant, ambush-protected vehicles that were sped into the field to combat IEDs are notoriously difficult to drive, he noted. He asked whether robotic systems integrated into these vehicles could help take some of the burden off drivers.
“Now we are starting to see more and more of this notion of an optionally manned vehicle,” he said.
Optionally piloted aircraft, which feature cockpits that can accommodate a pilot, or can be flown remotely, is a trend in the unmanned aerial vehicles realm. Overholt sees a world where any kind of ground vehicle, tracked or wheeled, from Humvees all the way up to class-A trucks, can accommodate a kit that will allow these options.
A soldier could potentially jump out of his vehicle with a controller, or maybe use his hands to gesture what he or she wants the vehicle to do, and then the vehicle will continue driving autonomously.
At a robotics rodeo at Fort Benning, Ga., last year, engineers from the Southwest Research Institute demonstrated a system added on to a Ford SUV that had six modes: manual driving, remote control, pedestrian following, vehicle following, supervised autonomy and full autonomy. With supervised autonomy, a driver is still behind the wheel, but he or she makes small adjustments to the wheel or brake when the robotic system gets off track.
Overholt noted that B-kits added to vehicles that can provide some of these modes could help answer the congressional call to make one-third of all military vehicles autonomous by 2015.
That would still be an expensive proposition, he said. One-third comes to about 200,000 vehicles and could cost $10 billion to $25 billion.
A request for information on such technologies was planned for this summer, with an industry day forthcoming, he said.
“That is the challenge for the community. Can you develop something that can interface with the vehicle?” Overholt asked.
As for the Department of Transportation, “our number one goal is not cars that drive themselves, but cars that don’t crash,” Augustine said.
The first step is developing a low-weight, low-price system that warns drivers of an impending accident. For example, one of the most common accidents is a rear-end collision when one car abruptly stops or slows down, and the car behind fails to brake in time.
The system will rely on a 5.9-gigahertz transmitter and GPS receiver that sends out about 10 messages per second to nearby vehicles. A warning device will tell drivers when a crash is imminent and it will be up to them to take evasive actions. Nothing will take control of the car or truck, Augustine said. At a blind intersection, for example, cars will be able to “talk with each other” and warn drivers that they are approaching too fast even if they cannot see each other’s vehicles. The department believes the system has the potential to reduce crashes by 81 percent.
Of course, for it to work, it will have to be installed on every car on the road, Augustine noted.
That is where the cultural acceptance issue comes into play in the civilian world. The public will have to be assured of privacy and that the system can’t be hacked into and corrupted, he said. There is also the cost. Customers who buy luxury vehicles that are currently featuring crash avoidance technology can absorb the increased price. It will have to be more affordable for it to proliferate, he added. The system can be integrated into new cars, but it takes about 20 years for all the vehicles in the United States to be replenished. For older models, there will have to be incentives for owners to add the system, Augustine said.
The underlying technology for the program is relatively mature, Augustine said. It is the cultural acceptance, policy and “human factors” issues that must be worked out. For instance, do drivers respond better to a warning that comes as a beep, buzz or a vibration under the car seat?
A consortium of eight car manufacturers from the United States, Japan, Germany and South Korea is part of the research and development efforts.
A large-scale test organized by the National Highway Traffic Safety Administration is scheduled for early 2012 at a yet to be determined location. There will be anywhere from 2,000 to 3,000 vehicles outfitted with warning devices in a real-world scenario, Augustine said.
As for letting vehicles take over control from a driver in order to avoid a crash, the program is looking into that, he said.
“When do you let the vehicle take control? We are going to have to … prove out the research. But we are on that path,” Augustine said.