
As societies become increasingly mobile and more dependent on information technologies, their radio spectrum needs change. The basic spectrum management structure in all countries was developed over a long period, and much of the basic framework has remained stable for decades. Much of this framework was also developed when the available frequency range for commercial systems was much lower than it is now, and the bands used had different propagation characteristics than most bands used today. (After all, 330 MHz is called high frequency because it was considered "high" in the early days of radio.)
The Federal Communications Commission (FCC) has been reviewing basic assumptions of spectrum policy in the past few years, and its landmark Spectrum Policy Task Force (SPTF) report in 2003 outlined a variety of possible new initiatives [1]. In May 2004 the FCC followed up on this report and released its proposals in Docket 04-186 [2] dealing with the possible use of cognitive radio technology for low-power unlicensed devices to share spectrum in the VHF and UHF television bands.
Cognitive radio technology refers to a new type of radio that uses real-time interaction with its environment to determine transmitter parameters such as frequency, power, and modulation. Simple versions of this technology have been in use for several years. For example, many cordless telephones select the frequency to use at a given time based on sensing other users. However, in the past, use of such technology has been limited to bands occupied only by unlicensed devices. If the FCC allows this technology to be used where television broadcasting is unambiguously "primary" with respect to other users, this would be a significant change.
The FCC was interested in making more spectrum available to unlicensed devices because of the explosive growth of IEEE 802.11/Wi-Fi unlicensed wireless LAN technology, and because the FCC recognizes that there is a great deal of "white space" [2] that could be exploited by unlicensed devices. That is, at a given time and place, monitoring usually shows much apparently unused spectrum that could, in theory, be used by new devices without adversely impacting existing users even though frequency assignment data shows there is little or no unassigned spectrum in most bands of interest.
There are many reasons for this white space. Some is due to the large peak-to-average use ratio of many systems that have dedicated spectrum, such as those used for public safety mobile users that have dedicated spectrum. Some white space is due to spectrum assignments designed to accommodate practical receiver limitations, such as limited adjacent channel and image frequency rejection. Finally, some white space is simply due to the fact that population and hence demand for spectrum is not spatially uniform.
A measure of this white space can be seen in a survey funded by the National Science Foundation in New York City during the 2004 Republic National Convention. Even though the convention led to higher than average spectrum use, there was still significant white space.1
The FCC proposal selected the TV bands as an initial home for cognitive radio based on several factors. First, the frequencies used for TV would give greater range in rural areas than the microwave frequencies used for Wi-Fi, and would also give better building penetration than Wi-Fi in all areas. Second, TV broadcast systems usually use high antennas, and the intended receivers need greater than 10 dB signal-to-noise ratio (SNR) to function (higher for analog NTSC). These large SNR ratios simplify the technology needed to detect whether a channel is currently in use. Finally, TV transmitters are left on more or less continuously, and infrequently change location or frequency. Thus, it would appear that it would be simpler to use cognitive radio in TV bands than in any other band. While these factors would also apply to AM and FM broadcasting, the 6 MHz bandwidth of TV channels makes them particularly attractive for this use. Since only a minority of households depend on over-the-air TV reception,2 the impact of any errors in selecting unused frequencies would be very small.
The FCC's recent proposal discussed three possible techniques unlicensed devices might use to determine whether white space spectrum is available for use at a given location:
References
[1] http://www.fcc.gov/sptf/reports.html
[2] http://hraunfoss.fcc.gov/edocs_public/attachmatch/DOC-247169A1.pdf
[3] http://gullfoss2.fcc.gov/cgi-bin/websql/prod/ecfs/comsrch_v2.hts?ws_mode=retrieve_list&id_proceeding=04-186
[4] http://gullfoss2.fcc.gov/prod/ecfs/upload_v2.cgi Enter "04-186" in entry box 1, "Proceeding".
[5] http://www.fcc.gov/realaudio/presentations/2003/021203/featuredetection.pdf
[6] http://gullfoss2.fcc.gov/prod/ecfs/retrieve.cgi?native_or_pdf=pdf&id_document=6516883656 and http://gullfoss2.fcc.gov/prod/ecfs/retrieve.cgi?native_or_pdf=pdf&id_document=6516883657
[7] http://gullfoss2.fcc.gov/prod/ecfs/retrieve.cgi?native_or_pdf=pdf&id_document=6516983275
Biography
MICHAEL J. MARCUS [F] received S.B. and Sc.D. degrees from MIT in electrical engineering. He retired from the FCC in 2004 where, among other things, he helped create the unlicensed bands that are home to Wi-Fi and Bluetooth. He now resides in Paris, France, where he does research and consults.