Researchers Push the Radio Rainbow’s Limits

Guevara Noubir

Back when walkie-​​talkies and car radios rep­re­sented the height of wire­less tech­nology, there were plenty of fre­quen­cies to go around. The spec­trum of radio waves was easily parsed into dis­crete packets: one for the oldies sta­tion, one for playing cops-​​and-​​robbers in the woods behind your house, sev­eral for real police offi­cers and for mil­i­tary communications.

But with the increasing use of wire­less devices, that spec­trum is get­ting clogged. “You cannot create more fre­quency,” explained Gue­vara Noubir, a pro­fessor in the Col­lege of Com­puter and Infor­ma­tion Sci­ence. “The spec­trum is like a nat­ural resource—it has to be shared by everybody.”

The problem of wire­less inter­fer­ence isn’t lim­ited to things like dropped calls, which we might be more familiar with. Noubir said it also presents a sig­nif­i­cant secu­rity chal­lenge as adver­saries may inten­tion­ally jam communications.

The Fed­eral Com­mu­ni­ca­tions Com­mis­sion, or FCC, announced in June that it would expand a pres­i­den­tial direc­tive to iden­tify more fre­quency band­width through policy and best prac­tices. But as quickly as band­width is freed up, new wire­less tech­nolo­gies enter the market demanding even more.

Since cre­ating new spec­trum would be as easy as gen­er­ating petro­leum fuel out of thin air, researchers like Noubir are set­ting their sights on better wire­less sys­tems that more effec­tively use the fre­quen­cies we’ve got. Toward that end, the Defense Advanced Research Projects Agency, also known as DARPA, has rolled out a first-​​of-​​its-​​kind Spec­trum Chal­lenge to design and build a soft­ware defined radio from the ground up that address today’s challenges.

In a pre­lim­i­nary round ear­lier this year, 90 research teams from around the country watched their software-​​defined radios attempt to dom­i­nate their com­peti­tors in one sec­tion and then to coop­er­a­tively share the spec­trum in a second. In that first event, the pool was reduced to just 18 teams, which faced off again ear­lier this month.

Com­paring wire­less com­mu­ni­ca­tion sys­tems and algo­rithms is can be dif­fi­cult, given the chal­lenge of repro­ducing the envi­ron­ment for radio wave prop­a­ga­tion. In order to create a level playing field for wire­less sys­tems com­par­ison, the tour­na­ments were phys­i­cally run on a syn­thetic testbed at Rut­gers University.

Noubir’s team, which also includes former grad­uate stu­dent Bishal Thapa and cur­rent doc­toral can­di­date Triet Vo-​​Huu, has con­sis­tently dom­i­nated the pack. Since the 1990s, when Noubir helped develop the world’s first 3G net­work, he and his col­leagues have devel­oped var­ious tech­niques for effi­ciently blocking com­mu­ni­ca­tions as well as over­coming smart attacks trans­mitted by others on the same fre­quency wave. “Our tech­niques draw from com­mu­ni­ca­tions and coding theory, algo­rithmic game theory, and soft­ware defined radio,” he explained.

Their soft­ware earned Noubir’s team first place in this month’s coop­er­a­tive sec­tion and third in the com­pet­i­tive sec­tion. Theirs was the only team to place in the top three in both sec­tions, as well as the only team to retain its high-​​level standing in the pre­lim­i­nary round. Their achieve­ments were met with a $25,000 prize from DARPA.

The next event in the chal­lenge will be held in March 2014, when the same teams will com­pete again with sim­ilar rules and toward a sim­ilar goal. “We’ll be spending the next six months making smarter tech­niques on top of the foun­da­tion we’ve already cre­ated,” Noubir said.