The Rising Spectrum Crunch—Cognitive Radio to the Rescue

E-mail Sinisa Peric

The radio frequency (RF) spectrum is in high demand; the skyrocketing rise in mobile broadband traffic and the recent drive to "have all things wireless" is taxing the availability of the current spectrum more than ever before. Can the latest push for cognitive radio regulation and standardization help by enabling dynamic spectrum allocation for unused portions of the spectrum?

One hundred years ago, the RF spectrum was a free resource available to all. However, since radio devices were not then widely available, relatively few people took advantage of it. As the U.S. government, private companies, and amateur radio operators began to use the RF spectrum, they experienced interference between transmissions in the same spectrum in the same geographical location. People realized that if everyone could use any part of spectrum, anywhere, at any time, and for any purpose, interference would become both inevitable and unpredictable, as would the operation of any devices using the RF spectrum.

The U.S. Government Steps In
The U.S. government was given authority via the Radio Act of 1912 to regulate how the RF spectrum is used. The government licenses parts of the spectrum in a certain area, for a certain amount of time, to an exclusive user, and usually for a specific purpose. The government also allows license-free use of certain parts of the spectrum if the use is within certain transmission parameters and serves a specific purpose (e.g., the industrial, scientific and medical (ISM) bands). Because of the limited amount of the RF spectrum that can be effectively used by existing technology, and the significant demand for it, the RF spectrum has become a precious resource.

Recent examples of this thirst for the RF spectrum are the FCC's repurposing of the SMR 800 band from iDEN to CDMA, which uses the spectrum more efficiently, and an addition of 5 MHz of Block G spectrum to the PCS band. Sprint is now deploying CDMA in the SMR 800 band and LTE in PCS Block G nationwide, allowing Sprint to serve more customers and enable higher data rates for those customers. By 2020, up to 500 MHz of the spectrum suitable for mobile and fixed wireless use should be freed up. It is a painful, costly, and slow process to move a spectrum allocation from one use(r) to another.

Alternative Solutions
So could there be a better way? Cognitive radio (CR), conceptualized back in 1998, is based on the principle of dynamic spectrum allocation (DSA). Cognitive radios first perform spectrum sensing, a process of measuring the amount of RF power available in particular parts of spectrum. If a part of the spectrum exhibits very low or nonexistent RF power, cognitive ratio can use it for its own communications. This could apply to either licensed or unlicensed portions of the RF spectrum. For cognitive radios to communicate among themselves via the spectrum they have sensed to be unoccupied, the radios must first inform each other of the particular open frequency and other parameters. Cognitive radios also need to continue assessing the spectrum to recognize whether the seemingly-unused part of the spectrum remains unused while the radios are operating. This is typically done by scheduling some quiet time for all the CR nodes, during which CR radios check to see whether the portion of the spectrum they are using has started experiencing any non-CR activity. If this occurs, the radios would have to vacate it. For this reason, CR nodes must also sense whether another part of the spectrum might be better suited for their communication. Transition to a new part of the spectrum requires coordination and punctuality to prevent the interruption of ongoing communications.

If the CRs used are independent and decentralized, a mechanism must exist to discourage CRs from interfering in a portion of the spectrum that is sensed to be unused. Usage of this portion of the spectrum should be efficient for all CR nodes wanting to communicate. That is why standardizing CR spectrum-access policies and protocols is necessary. The first rules regulating CR devices appeared in a FCC Second Memorandum Opinion and Order on September 23, 2010. The corresponding IEEE standard 801.22 was completed in July of 2011.

Based on these standards and regulations, CR technology operates in the vacated analog TV channels or the unused space between TV channels from 54–698 MHz. These CR devices are called TV Band Devices (TVBD) and are allowed to transmit up to 50 mW of power. If, in addition to sensing an unused channel, a TVBD also knows its location and confirms that this channel is allowed at this location by querying a database of channels in use, the TVBD can transmit up to 100 mW. Fixed TVBDs (Base Stations) transmissions are allowed up to 4 W, but cannot be placed at locations more than 76 m above the average terrain and cannot have antennas higher than 30 m above ground. TVBDs require no license and their data rate is scalable with the number of free TV channels (with data rates up to 19 Mbps per TV channel) that can be used at a given location. The communication distance is up to 30 km from the Base Station.

So, in light of these CR developments, maybe we should not ask ourselves, "Can we free up some more spectrum?" but rather, "Can spectrum be free…again?"

Sinisa Peric is the Chief R&D Engineer at QRC Technologies, Fredericksburg, VA, where he is responsible for test and measurement algorithms for various cellular systems and measurement devices. He is a Senior IEEE member, was awarded a U.S. patent and has multiple patent applications pending. He can be reached by email at [email protected].