Army works with industry, academia to study micro-robotics
February 27, 2012
- Army researchers envision a scenario where an aerial vehicle might drop several ground platforms that distribute themselves, find places to hide and create a communication network
ADELPHI, Md. (Feb. 27, 2012) -- Imagine a future when American Soldiers will see threats lurking just beyond their vision with help from autonomous robots the size of bats and hummingbirds and even down to the size of a fruit fly.
"The idea is that last 100 meters coming up to a threat building, a cave or a tunnel system," said Army scientist Dr. Brett Piekarski, Ph.D., "I want to send something in to give us informational awareness before the Soldier goes into that environment."
Piekarski leads the U.S. Army's Micro Autonomous Systems and Technology Collaborative Technology Alliance, studying miniature robotics. He envisions tiny intelligent robots helping Soldiers accomplish complex tasks.
"It's all about that last 100 meters. The Soldier releases micro-robotic platforms. They find their own way into the building or whatever the environment," he said. "They navigate within that building on their own. They map out the building. They detect threats. They get information back out to the Soldier. And, they do all of this autonomously."
The reality is still years away.
The Army initiated a partnership with academia and industry nearly five years ago to conduct basic research.
A problem with today's robots is the ratio of user-to-robot. Today, Soldiers remotely control the movements of robots.
"When the Soldier is driving it, he's focused on the screen, and not protecting himself. If he were to drive it into the building and lose connectivity or communications, to get it back, he'd have to go in and get it."
The vision for the future is to have many robots overseen by one Soldier.
"That's the goal, he said. "Find your way in and get the information back out. While the robot is doing that, the Soldier can be doing something else."
When the program started, the Army Research Laboratory in Adelphi, Md., had several objectives. Scientists wanted to understand fundamental technologies that would to enable autonomous micro-robots to work together.
"We're talking about systems that are really small, collaborative, autonomous robotic platforms all the way down to the insect size," Piekarski said. "These things are operating from the handheld size all the way down to fruit-fly size platforms."
How big is a fruit fly? Consider making a functional robot that's approximately 1/16 of an inch long. ARL scientists developed and fabricated fruit-fly size flapping wings in coordination with researchers at the Universities of Maryland and Washington who are working on understanding bee and fly flight mechanics and controls. Harvard University researchers are working on insect flight at a slightly larger scale.
"One of the questions was whether flapping wing flight is better than rotary or fixed wing," he said. "A lot of what has been researched over the last four years is to understand the mechanics of wing motion."
Piekarski said a bee can right itself within a couple of turns when being hit by a wind gust -- much more quickly than existing robotics systems.
"How do they do that? They don't have the complex sensors we have. They're working with very simple systems and controls. We're trying to understand the control theories," he said. "
We looked at nature to see what we could learn from biological systems. How can we apply some of those lessons to robotic platforms to enable these small-scale platforms? It's very broad, complex and interdisciplinary."
The Army's research program, Micro Autonomous Systems and Technology, is known as MAST. It brings together a lead defense contractor, BAE Systems, to work on microsystems integration.
The University of Michigan works on microelectronics, while the University of Maryland focuses
on microsystem mechanics. The University of Pennsylvania works on processing for autonomous operations.
Researchers are studying more than 70 complex tasks. As an alliance, other additional working general members support research within the centers and thrust areas. Members include the University of California at Berkeley, the Cal Tech and the Jet Propulsion Laboratory, the Georgia Institute of Technology, the University of New Mexico, North Carolina Agricultural and Technical State University, Massachusetts Institute of Technology and Harvard University.
The researchers are developing both ground and air collaborative platforms, as well as a hybrid of the two.
The project started in 2008 as a five-year journey, with an option for another five years. Looking back, many auspicious events were happening in the robotics field, Piekarski said.
The Defense Advanced Research Projects Agency, a research organization of the U.S. Department of Defense known as DARPA, wanted to see who had the technological know-how to create the first fully autonomous ground vehicles capable of completing a substantial off-road course within a limited time. A team of Stanford University researchers won the $2 million DARPA Grand Challenge on Oct. 8, 2005. The competition was for driverless vehicles using ladar, a laser range finder. It paints a 3-D image of its environment.
"That's a pretty substantial piece of equipment, and it's not something you can put on a 20-gram robot," Piekarski said. "The question is, if you go down to a 20-gram flier that has three to five grams of payload, what kind of sensors can you put on it? The algorithms you have for the DARPA Grand Challenge vehicles have multiple duel-core processors in the back; they have big sensors, radar, ladar and more. They've even got sensors on the wheels. You're not going to put that on a 20-gram platform."
Piekarski said the challenge of autonomous microrobots is more challenging.
"Not only are we trying to navigate, we're trying to collaborate with other platforms," he said.
"We're trying to do communications and network links. We're trying to control threat detection. It's a more complex problem on a much more constrained platform."
Piekarski is optimistic because of MAST program accomplishments so far. Now, Army officials are evaluating the next step. The Army funds a large portion of the research, but industry and academia are invested.
"We're not there yet, but we've done a lot of good basic research," Piekarski said.
Energy requirements are another challenge.
"Power is a big problem for some of these smaller platforms," he said. "The lifetimes some of the emerging technologies are about five or 10 minutes. Some platforms are at 20 minutes. But, through a combination of approaches we think we'll be able to address the solution as we progress in the technology."
Piekarski hopes to leverage the huge investment in small batteries being made by industry for small electronics.
Someday the MAST vision may offer revolutionary ways for Soldiers to gather vital information on the battlefield. This research also has several applications outside the military, Piekarski said.
"We're trying to create a capability for the Soldier so we don't have to send him into a threat situation. You could apply those same things in the civilian world where there may be threats, like police, fire, rescue and disaster relief, for example. A lot of the individual technology can and is being spun out to transition to larger commercial platforms over the course of the program," he said.
"We want to increase the operational tempo to match that of the Soldier so the robotics can move and detect threats at the same tempo as the Soldier as they move through complex environments," Piekarski said.
Today's autonomous systems operate at a much slower pace.
"We're working with the consortium to identify the real focus for the next five years," Piekarski said. "As we go into the next five years, how are we going to change that focus? The base goals are still the same. It's a force multiplier. I want to be able to add to their capability where they can go into a building and distribute sensors. Robots will go in to see if there are any threats. If not, they distribute and leave behind sensors."
Piekarski envisions a scenario where an aerial vehicle might drop several ground platforms that distribute themselves, find places to hide and create a communication network.
"You can think of it as a mobile sensor network that gets left behind," he said. "In the future, a Soldier would have two or three micro-robots in his pocket. He would let them go and they would operate autonomously. He wouldn't have to drive; the platforms might have more sensors, more payload. It might detect people or chemicals, maybe even nuclear residue. In the end, there is an enhanced capability while freeing up the time of the Soldier."
The project won't develop a single product, and that's not its purpose, Piekarski said. "It's a fundamental research program. We're looking at how far we can push the technology. We want to understand the fundamentals of winged flight -- not just to see if we can make something fly, but understand why it flies. We want to develop control algorithms and develop a fundamental understanding of these technologies."
Some may see parallels with science fiction. Piekarski discussed a scene from the movie "Minority Report" where police released spider-like micro-robots. They quickly detected their target and relayed that information.
"Life has a way of imitating art," Piekarski said. "To get to where we want to go we have to have real autonomous systems navigating through these buildings and working collaboratively. "We're going to have to have integrated solutions to make those things a reality, and that's where our program is going."