Smarter ground robots partnering with Soldiers
August 14, 2014
By David Vergun
HYATTSVILLE, Md. (Army News Service, Aug, 14, 2014) -- "In the Army, we always say, 'never send our Soldier into a fair fight.' Each of you here," from the robotics community, are "helping to make that happen," said Heidi Shyu, assistant secretary of the Army for Acquisition, Logistics & Technology.
Shyu, who provided the keynote address at National Defense Industrial Association's Ground Robotics Capabilities Conference & Exhibition here, Aug. 13, emphasized common architecture, open-source software and open standards for robotics development to further competition that will benefit the Army, taxpayers and industry.
The Army is working with industry partners to develop a standard architecture which will enable us to incorporate future (robotics) capabilities rapidly, keeping pace with dramatic commercial improvements, she said.
Getting the development of ground robotics right is important because the systems have become such an essential partner to warfighters, Shyu explained.
In 2004, 162 robotic systems were deployed to Iraq and Afghanistan, with a primary focus on explosive ordnance disposal, known as EOD, removal.
The use of ground robotics in combat since then has grown exponentially, with more than 7,000 systems currently deployed overseas, she pointed out. Besides helping EOD, ground robots now carry weapons, cameras and sensors for such things as detecting chemical, nuclear and biological material.
'PROPRIETARY' A DIRTY WORD
"Propriety is the worst word out there today," said Rich Ernst, interoperability lead, Office of the Secretary of Defense, referring to the opposite of open architecture, bolstering what Shyu had said earlier.
Ernst was part of an Open Architecture panel that followed Shyu's remarks.
While everyone knows the wisdom of having an open system, habits are hard to break, especially in the Defense Department, he said.
"Primes love open systems," Ernst said, "but then they'll tell you: 'just don't mess with my existing system.'"
That existing system, he said, is "a legacy environment. They want to go back to that for the next 30 years.
However, primes know they have to change because there are less programs going forward due to fiscal constraints, he added.
Besides an open architecture, Ernst said each system needs to be broken apart, made transparent and competed to the most innovative vendor, which in many cases might likely be small businesses or start-ups.
A typical system might be broken apart into 50 sub-components, he continued. The only problem is the government now has trouble managing "just one chunk." It will take a while for government to embrace this concept.
Once open standards are implemented and components are competed in the marketplace, the ground robotics systems that emerge will provide the warfighters and the taxpayers their biggest return on investments, he predicted.
Ernst also had a few choice words about "lawyers in the Pentagon who lock things down in contracts" so changes to the platforms that make sense become hard to initiate.
"I found out quickly that no matter how well we come up with the standard or specification, the lawyers undo whatever the engineers do," he said.
Ernst said he now works with the lawyers and the primes as hard as he works with the software folks to ensure things get done.
Brian Gerkey, CEO of Open Source Robotics Foundation and another panel member, agreed with Ernst's assessment. He said Robot Operating System, or ROS, builds on open architecture.
ROS is an open-source set of powerful software libraries and tools that helps anyone -- from businesses to school kids -- build robot applications and share solutions and algorithms "so you're not constantly reinventing the wheel."
ROS has about a million users worldwide, he added, including NASA, which is about to install a ROS-developed robotics application on the International Space Station.
IOP vs. ROS
Mark Mazzara, Robotics Interoperability lead for Department of the Army Systems Coordinator for Robotics, was the third panel member. He said the Army's Unmanned Ground Vehicle Interoperability Profile, or IOP, is setting the architecture standard and he hopes to see it accepted DOD-wide because "it's shown to reduce lifecycle costs."
Addressing Gerkey's earlier remarks, Mazzara said "ROS is a great thing. The difference between ROS and IOP is IOP is more focused on interoperability between subsystems -- which messages flow between them -- not the components in the black box," which can be created using ROS tools and libraries.
Studies have been conducted showing that both ROS and IOP can coexist, and both can be used to ensure the architecture stays open, he added.
A caveat to that, he said, is that IOP is being developed within the U.S. government and is being shared with allies, including NATO. Industries that want to build components for Army robots need them to be IOP-certified.
Mazzara said he can't predict the future of IOP, and whether or not the government will turn it over to industry, or to a non-profit robotics association that implements standards down the road.
"We'll just have to wait and see how it plays out," he said.
Mazzara added more to his thoughts on what the future holds for ground robots.
He thinks that an industry like agriculture could benefit from using some of the same or similar platforms the Army uses. Although the payloads would obviously be very different, a common mobility platform would make a better business case for internal investments, meaning quantity would drive down the cost of production.
The Army is now focused on modularity, ensuring components can be installed and removed in the "plug-and-play" mode that Shyu mentioned earlier, he said. The next phase, which will happen very soon, will focus on interoperability protocols between robots and manned ground vehicles, ground robots to ground robots and ground robots to unmanned aerial systems.
Besides those interoperability requirements, the Army will soon turn its attention to interfacing geospatial data, databases and even cloud computing with the ground robots so they can become smarter and more autonomous.
A key to all this, he said, is to surf the wave, keeping abreast of developments or emerging technologies in the automotive, mobile phone, software, Unmanned Aircraft Systems, and robotics industries. These are overlapping technologies that have applicability.
COMBAT PATCHES EARNED
While the panel sees a bright future ahead once a few clouds move away, Shyu pointed to two examples where robots are being used successfully today on the battlefield in Afghanistan.
The Mini-EOD, referred to as "Devil Pup," can locate, identify and disarm explosives, she said. It's so small and light that a Soldier can carry it in his or her rucksack on a long foot patrol.
Some 300 of them have been in theater over the last few years, at a cost of $35 million.
"It's truly saving Soldiers' lives," she said. "That's the power of robotics."
The other is the six-ton, M160 Anti-personnel Mine Clearance system, which can clear minefields in urban areas and practically any field condition. The M160 has "rendered previously unusable roads functional again," she said.
Near-term Army plans for robots include replacing the Talon Family of Robots with the Man Transportable Robotics System, or MTRS, a process that will take at least seven years, she said, noting that more than 2, 200 Talons have seen combat service over the past decade, and they're now past their service life.
Both the Talon and MTRS are tracked vehicles, with the Talon weighing 115 to 140 pounds and the MTRS 164. They can carry a number of payloads used for missions ranging from EOD to surveillance, with MTRS having planned chemical detection capability as well.
Between now and 2021, the existing Talons will get upgraded sensors and payload capacity, as a "bridging strategy" until MTRS can come online, Shyu explained.
Returning to her theme of common architecture, Shyu said MTRS will definitely have a capability so that if a camera, sensor, arm or other component becomes obsolete, a new device can be fitted to its common chassis in a "plug-and-play" fashion.
As it stands now, the MTRS Increment II program will soon conduct an analysis of alternatives, "which will determine the best acquisition strategy to gain cost and performance efficiencies across multiple Army formations," according to the Program Executive Office for Ground Combat Systems.
ROBOTS ON THE PROWL
"The future of ground robots depends on their ability to operate in a very diverse and constrained environment," Shyu said. "Commercial autonomous vehicles today maneuver very well on well-defined roadways, where GPS maps are available."
However, formations have to navigate through challenging terrain like deserts, unpaved roads, rocky hillsides, jungles, and urban areas, often in adverse weather like snow, ice and sandy deserts with temperatures in the triple digits.
Add to that contested environments where jamming and possible capture are possible.
"Efforts to overcoming these challenges are essential," Shyu said.
Despite tough fiscal environments, "our robotics industry continues to innovate," she concluded. "The future for ground robots has absolutely unlimited potential. Opportunities for invention and innovation are limited only by our own creativity and our willingness to take risks and take on new challenges."
The Army recognizes the value of science and technology efforts going into robotics, she added.
Despite fiscal challenges, the service is "working very hard to protect its S&T portfolio," she explained. "It used to be the Army's fourth biggest portfolio behind aviation, mission command and ground combat systems. It's now the Army's third biggest portfolio.
"I'm excited to see what academia and industry can bring in terms of innovative solutions to solve some of our most difficult challenges," Shyu said.
(For more ARNEWS stories, visit http://www.army.mil/ARNEWS, or Facebook at www.facebook.com/ArmyNewsService)