Army Still Determining Best Use for Driverless Vehicles
A fully autonomous convoy demonstration
The Army’s effort to introduce unmanned vehicles into its inventory has slowed as the service’s budget tightens and questions remain on how the technology will be used.
During the wars in Iraq and Afghanistan, when insurgents were attacking supply convoys, the Army made a push to develop driverless trucks. However, the technology never reached the battlefield.
Bryan McVeigh, force projection project manager, said it is vital that the benefits are fully understood before the Army commits its shrinking resources to autonomous technology.
“We haven’t proved that the juice is worth the squeeze yet,” he said in an interview with National Defense. “We have to demonstrate that the technology is mature enough before the leadership will allow us to commit … resources.”
The current budget environment has put a strain on the development of cutting-edge technologies, said Heidi Shyu, assistant secretary of the Army for acquisition, logistics, and technology during an acquisition forum at the Center for Strategic and International Studies.
Funding for the Army’s research, development and acquisition (RDA) account has been hardest hit by budget cuts, she said. “Since 2011, [the] Army RDA account has decreased twice as fast as the Army’s top line.”
That account finances experimentation and procurement of new technologies such as unmanned systems.
The role for unmanned ground vehicles needs to be better defined so that the technology can more effectively fit within the operating concepts of its units, she said.
The Army needs to look at more specific uses for unmanned systems, said Shyu. “What we need to do is scope to specific mission space [where] we can literally use an unmanned vehicle.”
She pointed to recent testing of autonomous shuttles on Army bases as one example.
Applied Robotics for Installations and Base Operations, a SmartAmerica Challenge team sponsored by the Army’s Tank Automotive Research, Development and Engineering Center (TARDEC), has completed several successful pilot tests for autonomous vehicles on military bases. They presented their findings at the June 2014 SmartAmerica Challenge Summit in Washington, D.C., a demonstration showcasing technologies of 24 teams formed by over 100 companies, government agencies and academic institutions. The challenge commenced in 2013 as part of a White House presidential innovation fellow project.
One of the main reasons the uptake of fully autonomous ground vehicles has not progressed at the same pace as unmanned aerial systems is that the applications and requirements have not been fully spelled out, said service officials.
“It’s … a leading-edge technology, so how it fits inside the operating concepts of units hasn’t been well defined,” said Scott Davis, program executive officer for combat support and combat service support. “If you don’t understand how that’s defined, it’s very difficult to come up with individual system requirements. I think we’re starting to turn that corner now that we have less attention focused on other engagements.”
The complex terrain that ground vehicles operate in is another barrier to full autonomy, Shyu said.
Trying to design a fully autonomous vehicle that can accommodate all of the potential environments that the military fights in is extremely difficult and costly, she said.
“Unmanned vehicles that can navigate [mountainous] terrain and not fall off the cliff, are not that easy,” she noted. Muddy and sandy environments that soldiers frequently fight in also have to be considered.
It’s very difficult for autonomous vehicles to navigate that type of environment without human intervention at some point, Davis said. Even unmanned aerial systems, though advanced, require a certain level of human involvement when initiating a flight, he said at the National Defense Industrial Association’s Ground Robotics Conference in Arlington, Virginia.
“I think we’ll see [full autonomy] probably in the next 10 years if you think about highways and things that have pretty clearly defined behavioral patterns and markings,” Davis said. But when considering open terrain and combat situations, requirement definitions will continue to evolve.
“The best thing we can do today is settle on the standards and taxonomy and really understand the architectures we’ll need to build in so we’ll enable growth as best as we can predict going forward,” he added.
The wars in Iraq and Afghanistan played a role in impeding long-term adoption of unmanned ground vehicles, according to officials.
“There was a war that happened and that probably affected program of record development,” said Maj. Gen. Robert Dyess, director of the force development directorate for G8 force development.
It allowed the Army to field large numbers of robotic capabilities, but most of them were bought with overseas contingency operations funding, he said. This meant that the majority of the programs didn’t have a long-term lifespan and lacked a program structure developed in Training and Doctrine Command courses.
Davis said the biggest push during the war was for technology that improved survivability. This meant that a lot of time, effort and money went into the development of more heavily armored manned vehicles. However, there were strides made in unmanned applications regarding counter-improvised explosive device measures as well as route clearance, Davis said.
Another hurdle that needs to be overcome in order to achieve autonomous vehicles in battle is getting key decision makers onboard with unmanned systems even in the tight budget environment, said retired Lt. Gen. Rick Lynch, former executive director of the University of Texas at Arlington’s Automation and Robotics Research Institute in Fort Worth.
“When you’ve got people making decisions and they have constrained resources … by design they’re going to fund those things that they’re most familiar with and comfortable with, and not fund those things that they’re uncertain about,” said Lynch, who has a master’s degree in mechanical engineering from the Massachusetts Institute of Technology with a concentration in robotics.
The acquisition process itself is too slow and cumbersome, he added. He said the robotics enhancement program — an Army effort rolled out in April that allows inventors and manufacturers to submit robotic technologies and solutions online — is a great idea, in theory, to overcome some of the problems with current acquisition procedures, but does not go far enough.
Getting the technology into the hands of soldiers from the very beginning so they can provide feedback to improve the product and accelerate fielding is a great idea, Lynch said. However, “As soon as you start going through the bureaucratic process of acquisition, you’re talking decades, not days,” he said, pointing to the program’s process of relying on a council of colonels for submission review and approval.
Another flaw is not having the testing community involved from the beginning. The testers and soldiers should receive the equipment simultaneously, not sequentially in order to streamline the process, Lynch said.
McVeigh said completely autonomous vehicles will not be achieved in the near term, but partially autonomous capabilities can be added to the current fleet with a digital architecture that can be upgraded with sensors in the years to come.
He pointed to recent experiments executed by TARDEC as proof that unmanned vehicles are making progress.
In 2014 TARDEC, in conjunction with Lockheed Martin, demonstrated fully autonomous convoys operating in urban environments. These experiments were performed with multiple vehicles of different models as part of the Army and Marine Corps’ autonomous mobility appliqué system (AMAS) program.
Vehicles on which the appliqué kit had been installed, exhibited the ability to stop when there was oncoming traffic, follow the rules of the road and avoid pedestrians and obstacles at speeds up to 25 mph during the first capabilities advancement demonstration in January 2014. The second demonstration in June 2014 focused on establishing the convoy-like scenario through increased speed and additional vehicles.
The technology in the AMAS kit comprises GPS, light detecting radar and automotive radio detection and ranging. It also includes commercial automotive sensors to make it more affordable.
These experiments illustrate ways in which the Army can start to utilize autonomous capabilities — drive by wire, lane departure warnings and collision avoidance — in the near term with a system that can be upgraded in the future, Davis said. “It’s really trying to design an architecture that fits the system today and provides some capability in terms of driverless safety and then provides the framework or underpinning for future additions that would allow us to ever increase the robotic capability of our manned platforms.”
He said the Army would continue to move forward with the experiments to further flesh out the technology, but more importantly to get additional vehicles in the hands of soldiers so they can begin employing them in more tactically realistic scenarios. That will help them understand exactly how they would use the technology, determine what shortfalls still exist and improve soldier confidence in the equipment, Davis said.
“This is one of those things where I don’t think we have to wait for the perfect technological breakthrough. I think all of the elements of robotics, from providing some assistance all the way up to full autonomy, has value. So it’s really trying to figure out where those two things meet — where we can get a substantial increment of technology that we employ on the vehicles, and it fits with what the soldiers believe that they can deal with,” he said.
Questions that will still need to be answered going forward are how removing drivers from the equation will affect the security of a convoy and how an autonomous vehicle will navigate complex urban scenarios where it will come in contact with other cars and pedestrians, he said. Some of these elements haven’t been fully explored and the experiments’ results will help shape the requirements.
When considering the shift from manned to fully autonomous systems, the immediate effects of such a transition also have to be considered, said Lt. Col. Larry Dring, product manager for unmanned ground vehicles.
When there is no longer a need for human drivers, the trucks can run 24 hours a day, but that can wear down the vehicles and create the need for additional mechanics because they are coming in for repairs three times as often, he said. That begs that question, “What are the real savings?”
The immediate future for unmanned capabilities in the Army is in large semi-autonomous systems used for bomb detection and disposal, McVeigh said.
Two large vehicles that are currently in development are the route clearance and interrogation system (RCIS) and the husky mounted detection system (HMDS), he said.
When the RCIS system identifies an improvised explosive device or some other abnormality in the road, it will allow the soldier to excavate the device remotely, removing troops from the “boom,” he said. The request for proposals will be released in the October to November timeframe and the program is about two years away from awarding a contract, McVeigh said.
The HMDS system is moving on a similar trajectory. “Right now there’s a soldier in there and [the vehicle] goes through and identifies where there’s a disturbance in the ground and uses a radar to see if it’s something we should be looking at,” said McVeigh. If the soldier can be taken out of the vehicle, then the suspicious object can be identified without putting the fighter’s life at risk, he added. The autonomous version of that system is still a few years away, he said.
RCIS will establish the building block for leader-follower and autonomous convoy operations, which provide clear benefits, McVeigh said. There is less chance of a soldier being harmed by an explosive device, and as the Army downsizes the number of active duty soldiers, automating convoys can free up more of them for war fighting.
“We see the requirements [for leader-follower] emerging through the process … I believe we will have a program of record for leader-follower by the turn of the decade,” he said.
Photo Credit: Army