Army: Individual Soldiers Will One Day Control Swarms of Robots

FacebookXPinterestEmailEmailEmailShare
Thanks to advances in AI, human-agent teaming and machine learning, Soldiers will provide commanders with real-time information about the enemy gathered from a variety of different sources. (Image courtesy of ARL)
Thanks to advances in AI, human-agent teaming and machine learning, Soldiers will provide commanders with real-time information about the enemy gathered from a variety of different sources. (Image courtesy of ARL)

Army robotics officials at Fort Benning, Georgia are trying to give individual soldiers the capability to control swarms of air and ground robotic systems for missions that often require large numbers of troops to accomplish.

U.S. ground forces have used small ground robots and unmanned aerial systems for years, but only on a small scale, said Don Sando, director Capabilities Development and Integration Directorate at Benning.

"To really get a large benefit from robotic systems, we have to break the one-soldier, one-robot link, because right now, you generally need one operator for one robotic system and that is effective and interesting, but when I can have dozens of robotic systems controlled by one soldier, now I have a significant advantage," Sando told a group of defense reporters today on a conference call.

A single soldier could conduct reconnaissance over "large areas with fewer soldiers and many dozens of robotic systems," Sando said.

"That starts to matter especially in conditions such as dense urban environment," Sando said. "The problem with urban environments is they consume soldiers ... limited lines of sight, tunnels, buildings -- all the things that just take manpower to overcome and control.

"If we can expand that with robotic systems, both air and ground, then that has significant impact."

The concept could be developed to enhance communications battlefields when networks are hampered by enemy activity as well as natural obstacles.

"If our communications infrastructure is going to be contested, as we know it will, then how can I regenerate quickly and effectively in a given area with robotic systems, both air and ground, to create that network?" Sando said.

CDID officials are developing a common controller that can control air and ground robots regardless of the model.

"We are very close on that; we did some assessment last year. We proved the feasibility of about three different versions of controllers that can effectively control air and ground robotic systems," Sando said. "The advantage to that is a soldier only has to learn one system as opposed to every robot has its own unique controller."

The goal is to make a decision on a common controller by late fiscal 2019, Sando said.

But the problem is more than just choosing the right controller.

"How do you train a soldier, and how do you train leaders to do that? Sando said. "It's one thing to have two hands on your rifle -- one soldier, one system. It's one thing to be a small unit leader, to have a few subordinate leaders under your control -- it's something else to have dozens of under your control."

Organizations continue to come to Benning to "practice and develop algorithms to employ swarming unmanned aerial systems," Sando said.

"The next thing beyond that is OK, how do I swarm ground robotic systems? How can I do that?" he said. "That is the thing we are least developed on and that's the thing we want to start trying to emphasize.

"We are going to continue to develop that and test that and I think that poses the next really large return on investment as we expand robotic systems."

-- Matthew Cox can be reached at matthew.cox@military.com.

Story Continues