Sunday, April 28, 2013

Media Overload - Air Force Drone

Welcome to the Age of Big Drone Data

BY SPENCER ACKERMAN
04.25.13
6:30 AM



An MQ-9 Reaper at Fort Drum, New York. Photo: U.S. Air Force

James is the Air Force’s deputy chief for intelligence, surveillance and reconnaissance, giving him the flying service’s drone portfolio. During a rare public talk yesterday in Washington, James let on that “sustainment” of the drone fleet is his next big task. That means focusing less on designing new robots, as the Air Force’s new budget indicates, and more on the human problem of managing the absolutely enormous amount of data that its Predators, Reapers, Global Hawks and Sentinels generate.

“The future is going to be taking all sources of information and developing knowledge and intelligence from that,” James said. He’s working on some software fixes for that, as well as some data-storage farms. Welcome to the age of Big Drone Data.

The Air Force has actually lived in it for a long time. Last year, Secretary Michael Donley lamented that it will take “years” for Air Force analysts to swim through the oceans of imagery that the drones yield. One of the major purposes of the drone fleet is to hover over an area longer than a plane with a pilot in a cockpit can, snapping photos and streaming video down to the ground. And when you’ve got a robot doing that for 16 to 22 hours at a stretch, the length of a typical drone combat-air patrol, all that data piles up.

James doesn’t have ready-made solutions, but he said the Air Force is starting to look long and hard at its big-data challenges. First comes upgrading its network infrastructure “to move the data around, store it as you need to and to do that securely.” (Indeed.)

Next comes an improved suite of software tools to integrate the video feeds with other forms of imagery, harvested from drones, satellites, piloted spy planes and other sources. It’s got to work so that “I’m not relying on the human eyeball to look at FMV, full-motion video, all the time,” he said, “the tools are doing that for me.” Forthcoming algorithms will find something from a database of electro-optical information, connect it with something from the signals database “and bring it together in a fused fashion,” James said. No timetable on when that’ll come online.

In the meantime, Air Force isn’t totally shying away from developing new kinds of drones. It’s got a “micro-aviary” of tiny, insect- and bird-like ones, currently in the research phase. And the Air Force has a long history of developing planes in secret. As aviation journalist extraordinaire Bill Sweetman wrote on Tuesday, there’s likely a new secret drone design in the works right now; and in any event, the Air Force wants its next long-range bomber to be pilot-optional.

But other challenges that James wants his drone fleet of the future — even if it looks mostly like the one he’s got today — to meet are going to compound the Big Drone Data problem.

The first is increasing the time the Predators, Reapers, Global Hawks and Sentinels can stay aloft, an engineering challenge. The second is getting ever-powerful sensors in the bellies of the drones, so they can loiter further away from the targets they spy on. That’s a big issue, since drones are really easy to shoot down — they fly slow and aren’t built to maneuver — and not every place the military wants to send them lacks sophisticated air defenses a la Afghanistan and Iraq.

Yet the longer the drones are in the skies and the better their sensor packages are, the more data they’ll produce — bringing the Air Force back into its data-management problem. James levels: they’re working on it.

“The software tools will lead the way,” he predicted. “And it’s not just the military that’s worried about how you handle this big data. There’s lots of corporate and commercial interests out there in terms of video and imagery and what do I do with it and how can I track things and see them.” Until then, the Air Force has a video glut on its hands.


No comments:

Post a Comment