DARPA Challenges Industry To Make Adaptive Radios With Artificial Intelligence

9 September 2016

The Pentagon’s research agency has a new challenge for scientists: make wireless radios with artificial intelligence that can figure out the most effective, efficient way to use the radio frequency spectrum, and win a pile of cash.

Winners of the Defense Advanced Research Projects Agency’s (DARPA) Spectrum Collaboration Challenge (SC2) could take home up to $3.5 million, but to do that, teams will have to demonstrate new technologies that represent a “paradigm shift” with both military and commercial applications, said Paul Tilghman, a DARPA program manager who is leading the challenge.

“The real crux of the problem is — when you look at users of the spectrum, whether they are commercial users of the spectrum, whether they're consumers or they're the military —  the thing that is ubiquitously true is we all are placing more and more and more demand on the spectrum, and all of that demand is really adding up and going to stress the way that we actually manage the spectrum,” he said. “Where do we put our communications systems? Where do we put our radars? Where do we put our [electronic warfare] systems”

The government has managed the radio frequency spectrum for pretty much the same way since the invention of the radio in the early 1900s: dividing it into pieces and distributing it to various government and commercial entities, he said. That lack of agility results in a lack of efficiency, as, at any time, portions of the spectrum could be overtaxed while others are not in use.

DARPA wants challenge participants to figure out how to make radios adapt to the current spectrum environment and keep pace with any real-time changes, Tilghman said.

“The goal of the Spectrum Collaboration Challenge is to take a bunch of radio systems that have no prior knowledge of each other and say: 'Your goal is to get the best, most optimal use of the spectrum you can possibly can.' We’re going to do this in a large, open tournament where multiple teams come in and compete to bring their best designs forward,” he said.

What sets the Spectrum Collaboration Challenge apart from other adaptive radio efforts is the focus on artificial intelligence (AI) as a key enabler.

“This idea of AI systems that can learn to collaborate with each other is really sort of a fundamentally untapped area,” he added. “And very specifically, we want to know if AI can tackle this problem."

Although the Sept. 2 deadline to submit a proposal to DARPA has already passed, teams can still vie for a chance to participate in the challenge through the “open track” meant to engage nontraditional participants without experience in bidding for government contracts. The agency will award a contract to the vendors who put forward the best proposal, but participants in the open track stand to win prize money of equivalent value if they can make it through several “hurdles” meant to weed out technologies that need more work.

BAE Systems is one of the companies that submitted a proposal in hopes of participating in the challenge, confirmed Josh Niedzwiecki, director of BAE's sensor processing and exploitation group. The company had already been developing artificial intelligence technologies, but the DARPA effort offers a new way to apply them.

"BAE Systems has been researching technology similar to the challenge to create a radio, and other sensors, that can adapt to use the spectrum more effectively using machine learning," he said. "BAE Systems' research in this area has touched upon the different pieces of the Spectrum Collaboration Challenge, but SC2 is a unique opportunity where it pulls all the different pieces together into a single, collaborative system."

Eventually, teams from each track will advance to the championship round, scheduled for December 2019, and use their radios to compete against other participants in the “coliseum” — a testbed that simulates a spectrum environment crowded with other adaptive radios, non-adaptive devices and interference. The team that best manages the spectrum wins the prize.

DARPA has not yet decided how it would move this technology forward if the challenge is successful, including whether it would pursue a follow-on program to continue development. Tilghman noted the broad commercial applications for communications systems and said the agency hopes the effort will generate excitement in commercial industry and academia, which could continue maturing such technologies.


Source : defensenews.com