As Army soldiers take up secure positions behind a wall, they deploy a small reconnaissance team - a very small one. Some hopping, some flying, the stealthy recon squad vanishes into a suspicious building for long minutes, then relays the all-clear back to its relieved partners outside.
It’s an intriguing scenario; it just hasn’t happened yet. But the chances that tiny intelligent mobile robots could someday assist U.S. armed forces and other personnel are improving, thanks to a major research program led by BAE Systems and including the Georgia Institute of Technology among 10 principal and general members.
Called the Micro Autonomous Systems and Technology (MAST) Collaborative Technology Alliance Program, the new five-year initiative is sponsored by the U.S. Army Research Laboratory (ARL).
The Georgia Tech Research Institute (GTRI), Georgia Tech’s College of Computing and the School of Aerospace Engineering and School of Physics are involved in the program, with Georgia Tech researchers contributing to three of four primary MAST research teams.
“At this point, it’s difficult to say which practical systems will come out of this, because we’re doing basic research at the moment,” said Mike Heiges, a GTRI senior research engineer. “By bringing together world-class expertise from several different fields, it’s hoped that within five years real-world applications can be developed.”
The program includes four principal research teams:
Other institutions participating include Boston University, Harvard University, the Massachusetts Institute of Technology, Vanderbilt University, the University of Milan (Italy), the University of Sydney (Australia); and two companies: Centeye and Daedalus.
Heiges, working with GTRI principal research engineers Jim McMichael and Lora Weiss, is supporting the integration effort. The GTRI integration team is collaborating with Dimitri Mavris of Georgia Tech’s Aerospace Systems Design Laboratory.
The integration team’s immediate focus involves determining the most desirable capabilities for the palm-sized robots. Currently, the Army uses unmanned systems that are sizeable and send back a raw data stream that must be constantly monitored by human operators.
“What the Army wants is very small robots smart enough to go off on their own and then alert you when they find something,” Heiges said. “These wouldn’t be dumb sensor carriers; they would interpret what their sensors are telling them.”
Also participating in the program is assistant professor Daniel Goldman of the Georgia Tech School of Physics. Collaborating with the University of Maryland-led microsystems mechanics team, Goldman is developing models of different ground surfaces to better understand locomotion on complex terrain like sand and leaf litter.
In another Georgia Tech effort, GTRI principal research engineer Tom Collins is leading a team focused on the processing for autonomous operations. Collins is collaborating with several investigators from the College of Computing, including Regents’ professor Ronald Arkin; Henrik Christensen, director of the Robotics and Intelligent Machines Center, and associate professor Frank Dellaert. Working with the University of Pennsylvania and others, the Georgia Tech researchers are contributing to the intelligence technology necessary for small-robot operation.
It’s a distinct challenge, Collins says. Potential operating environments are still unknown, and that’s problematic since a small robot can’t necessarily handle the same physical challenges as a larger robot.
“We’re looking at very complex locomotion mechanisms that we have to plan paths for - things that may jump, crawl or fly,” Collins said. “We have to look at everything in a rather abstract and general sense, while also dealing with all the problems of scale.”
Also at issue is how the robots will interact. Intelligence could be distributed to each individual machine, or the robot ensemble could send data to one highly developed processing node. Currently, research favors a decentralized approach in which individual robots would share information among their companions to form a more complete picture.
It’s likely that robot ensembles will be programmed quickly by people with limited training. Moreover, operators won’t be able to carry elaborate gear in the field.
“We’ll need to keep it simple,” Collins said. “The operator could specify the mission type, then choose a few parameters like direction and the number of robot participants. Then, relatively quickly, you’ve got a bunch of robots programmed and ready to go.”
Electronic Systems Laboratory (ELSYS) - http://www.gtri.gatech.edu/elsys
Robotics Trends would like to thank the Georgia Tech Research Institute for permission to reprint this article. The original can be found at http://www.gtri.gatech.edu/news/micro-brigade-miniature-robot-ensembles-could-aid-.