A seismic shift in radio design could break the spectrum-availability bottleneck, proponents say, and open up a new frontier of opportunities for radio designers and wireless application developers. Cognitive radio one that can sense its environment and location and then alter its power, frequency, modulation and other parameters so as to dynamically reuse available spectrum is emerging, but some are already calling it the holy grail of wireless design.
Derived from a doctoral thesis submitted in May 2000 by Joseph Mitola, who is now at work on the military's cognitive-radio initiative, CR could in theory allow multidimensional reuse of spectrum in space, frequency and time, obliterating the spectrum and bandwidth limitations that have slowed broadband wireless development in the United States and elsewhere.
This new technology is a kissing cousin of software-defined radio. With SDR, the software embedded in a radio cell phone, for example, would define the parameters under which the phone should operate in real-time as its user moves from place to place. Today's cell phone parameters, by contrast, are relatively fixed in terms of frequency band and protocol.
But cognitive radio is even smarter than SDR. CR "is a radio that's aware of and can sense its environment; learn from its environment; and perform functions for its user that best serve its user," said Bruce Fette, chief scientist at General Dynamics Decision Systems and technical committee chairman of the SDR Forum.
Although the communications and wireless industries are just beginning to hear about cognitive radio, the U.S. government is already showing a keen interest in its possibilities. The Federal Communications Commission is weighing the commercial applications and the Defense Advanced Research Projects Agency is proposing a military version.
The FCC in December officially began the effort to extricate the concept from academia by issuing a notice of proposed rule making (NPRM) calling for input on how CR could be realized. Now in the response stage, that NPRM underscores the unprecedented willingness of the commission in recent years to explore innovative ways to open new spectrum to commercial unlicensed use. Examples include the release of new spectrum in the 5-GHz U-NII band last year, as well as the opening up of 7.5 GHz of bandwidth for ultrawideband (UWB) signaling in the region between 3.1 and 10.6 GHz. Though the power levels allowed for UWB were extremely low a roof of - 41 dBm the move marked the first time the FCC had allowed unlicensed use across otherwise licensed bands.
"UWB was a first shot at this [reuse of spectrum]," said Bob Brodersen, professor at the University of California, Berkeley's Wireless Research Center. "But CR is a more complete solution as it [actively] looks for unused spectrum and begins to transmit inside those bands." Cognitive radios coexist with primary users, he said, but get out of those bands if need be when the primary users show up. "UWB just hid below the noise floor," Brodersen said. "This one does it by jumping out really fast. Yet it can transmit at reasonable levels."
In his opening keynote at the CTIA Wireless show in Atlanta last month, FCC chairman Michael Powell hammered home the need to open up new spectrum and to enable spectrum reuse. The commission, he said, formed the Spectrum
Policy Task Force in 2000 "as we saw the potential for wireless broadband."
But in the current system under which the FCC operates, formulated in the 1920s, different bands are assigned to different services and licenses are then required to operate inside those bands. As a result, Powell said, the FCC has not "been able to faster optimize spectrum use. We're moving cautiously but purposefully and will look at policies that will bring it all together for consumers."
UC Berkeley's Brodersen said that President Bush "also came out recently with an edict on the same topic of better spectrum use. So politically, there's great support for this."
The FCC's Spectrum Policy Task Force (SPTF) came out with a report in 2002 that recommended the agency migrate toward a "policy-based" solution where spectrum could be used on an opportunistic basis. "This is a paradigm shift," said Jeffrey Schiffer, director of Intel Corp.'s Wireless Research Radio Communications Lab (Santa Clara, Calif.). "You can characterize it as the most exciting thing in radio that's come along in a long time, because of the opportunities that exist now and the creativity you can have around the ability to agilely move across the band and operate legally in multiple different places, using things like policy engines as a means to check whether you're legal."
In particular, the FCC is looking at a means by which the 6-MHz-wide licensed spectrum in the UHF band, currently assigned to TV broadcasters, can be reused in secondary markets as a path to last-mile data access (see story, opposite page). This proposal would set power levels up to 8 dB higher than those currently allowed in those bands, thereby greatly increasing range and coverage area. In addition, said Schiffer, the propagation characteristics at the lower UHF frequencies are particularly attractive, offering the possibility of longer distance and lower power.
"Still, you have to operate in 6-MHz channels, so you won't get the 54 Mbits/ second you get with .11g or a [Wi-Fi]," he said. Nonetheless, Schiffer says, data rates "will be respectable," especially in rural areas where channels can be concatenated.
But the FCC's interest also extends to higher frequencies. "If you look at the entire RF frequency up to 100 GHz, and take a snapshot at any given time, you'll see that only 5 to 10 percent of it is being used," said Ed Thomas, chief engineer at the FCC. "So, there's 90 GHz of available bandwidth."
The military, too, has latched on to the concept of cognitive radio under the umbrella of the Darpa-funded XG or Next Generation Communications program. Its aim is to develop technology that allows multiple users to share spectrum in a way that coexists with, and complements, sharing protocols included in today's Wi-Fi technologies. "They're exploring the idea in the most extreme way," said John Notor, a system architect at Cadence Design Systems Inc. (San Jose, Calif.) and a leading researcher in the area of CR. "They're working on a 'dc-to-daylight' box covering a broad range of frequency bands up to the microwave area and which can use any spectrum at any time and adapt accordingly."
The government's interest in and definition of CR goes way beyond spectrum reuse, said Fette of General Dynamics. Indeed, CR and Darpa's XG technology are not identical, as many in the industry believe, he said. "The spectral-efficiency technologies as exemplified by XG [and the FCC's efforts] are one important facet of CR, but not the only facet."
From a military perspective, CR's ability to handle functions that best serve its user translates to sufficient situational and mission awareness to help the soldier reach an objective. "The kind of data required for this is still in the labs," said Fette. "Academicians are studying this and we want to work with them." Examples of data for a soldier, he said, could include local topography, mission objectives and time scales for those objectives, as well as knowledge of and access to the radio networks in the area as well as "the location of friendly and enemy positions and artillery."
WLAN launch pad
Researchers believe the technical foundations established by wireless LANs provide a launching pad for CR. WLANs already incorporate essential CR features such as dynamic frequency selection and transmit power control. Also, while the RF front ends may require wideband receivers and transmitters, Fette said, the hardware exists now, "and the software is a matter of sitting down and doing the software engineering" to make functions like filtering, band selection and interference mitigation "available as plug-in software modules for the radios. That's stuff we know how to do."
Fette sees the technology being implemented in a different style for the commercial market, "but the FCC's interest in spectrum efficiency will make that happen fairly quickly in five years. And that's not so bad," he said.
That's not to say there are no technical hurdles to surmount. "Some fundamental ideas need to be ironed out, such as interference temperature," said Brodersen of UC Berkeley. Interference temperature, a metric the FCC included in its December rule-making notice, involves helping a radio that's intending to transmit determine how much interference it will cause on other radios in the locale to ensure the interference doesn't cross a certain yet-to-be-determined threshold, said Brodersen. "But the problem is how you, as a transmitter, can detect how much interference you're causing this other receiver? That's a tough problem," he acknowledged.
Researchers are also grappling with how to identify channels and so-called "white space" within the TV spectrum, as well as the uncertainty surrounding FCC certification rules for radios with no assigned frequency bands.
Protocols and communications channels are also a thorny issue, since they're required to help disparate radios negotiate optimum communication parameters to minimize their respective interference levels.
"We're working on a prototype CR right now, but the problem is the protocol and how to figure out how to set up these calls," said Brodersen, referring to the kind of data that should go back and forth between the physical and media-access control layers. "So you have to design a protocol stack that's very different from what we've had before as you have this sensing thing built into it now."
"There's some research to be done, especially in the incumbent profile detection space," said Cadence's Notor. "Radios don't exist in isolation. Turn on a bunch of radios and how do they detect each other? How do they get out of the way if a licensed user comes up? None of this has been addressed in any standards I'm aware of."