By

To meet the growing demand for commercial wireless applications, the Federal Communications Commission is gradually opening new spectrum for unlicensed commercial use. One such example is allowing ultrawideband (UWB) devices to operate in the 3.1 GHz to 10.6 GHz band, but at very low transmission power. In addition, the FCC has recently proposed allowing unlicensed use of vacant TV bands. This willingness of the FCC in recent years to open up spectrum for unlicensed commercial use will likely continue and eventually break the spectrum scarcity bottleneck.

Harnessing the newly allowed spectrum in an efficient and cost-effective way requires a radical shift in radio and protocol design. Features in such designs include the ability to adapt to evolving FCC regulations, reconfigurability and the ability to allow sharing among non-primary users. All must be met with high regard to linearity, emission suppression and mixed signal requirements.

One approach is a spectrum-agile wireless communications system. Specifically, this system focuses on broadband wireless Internet access as the primary consumer application for both in-home and last-mile distribution. This thereby enables, for example, remote patient monitoring via the TV bands, pending an FCC ruling.

Design considerations
In current commercial wireless networks (operating both in licensed and unlicensed bands), devices increasingly are built using programmable platforms. Programmable platforms provide flexibility of design during manufacturing and upgradeability after deployment. Which functionality of spectrum-agile radios is upgradeable after deployment depends upon the design and architecture of such programmable platforms — for example, hardware-software partitioning. Key metrics being considered in architecting programmable platforms are cost, power and size to meet given application performance requirements such as throughput, delay, range and robustness. For example, a low-throughput application can be enabled by a complete software-defined-radio (SDR) in some frequency bands, but high-throughput applications (such as several Mbits/s to Gbit/s radio) need more hardware components to meet commercial requirements.

Flexibility to adapt a radio system's transmission characteristics (spectrum usage) based on external factors will be key to future commercial wireless networking devices. In fact this is one of the key functions needed in spectrum-agile radio that may not be needed in an SDR. In particular there are two external factors: real-time spectrum usage by other radio systems in its vicinity and also evolving FCC policies. Our proposed system retains this flexibility, while addressing cost, power and size requirements for a commercial broadband wireless application discussed above.

Proposed agile system
Spectrum-agile radios operate in radio spectrum originally licensed to other radio systems but that is currently vacant. Licensed radio systems are also referred to as incumbent or primary radio systems, with TV broadcast and radars being examples. To harness such vacant bands using spectrum-agile radios we propose the following solution:

• Radio design: A radio based on a multicarrier modulation scheme such as OFDM (orthogonal frequency division multiplexing). Unlike today's physical layers that are somewhat inflexible (for example, Wi-Fi can operate in predefined 20-MHz bands), our proposed physical layer will provide a broad range of flexible parameters to adapt transmission characteristics on the fly. The proposed scheme allows us to take advantage of the independence of subcarriers by assigning higher energy to subcarriers where, for instance, no primary is detected. In addition, frequency bands with higher SNR can be modulated to carry more bits per second. Parameters that can be set as needed include: number of subcarriers, intercarrier spacing, subcarrier power, subcarrier modulation order and signal bandwidth. Such a flexible architecture allows users to adapt its transmission characteristics to the required operating condition. Nevertheless this flexibility is achieved while meeting commercialization requirements (e.g. size, cost) for the broadband wireless application.

Linearity, filtering, out-of-band emission suppression and mixed-signal requirements may not be the same across the whole spectrum of interest. Therefore, front-end component design (radio frequency circuits and antenna design) issues will need to be considered, depending upon the width of the band in which spectrum-agile radio intends to operate. For example, operating in the 300-MHz to the 10-GHz band poses a challenge in both antenna and RF circuits design to meet a certain performance and commercialization metric. Nevertheless, a number of architectural trade-offs exist. For example, more than one front end can be considered, with each one tuned to a certain portion of the spectrum while maintaining the performance and commercialization requirements. On the other hand, a single unified architecture can also be considered that can take advantage of future semiconductor and antenna technologies.

• Protocol design: A key function of wireless protocols is providing access to the shared wireless medium in a fair and efficient manner. Beyond sharing between similar devices (existing protocols such as Wi-Fi), we propose protocols and algorithms for sharing with licensed radio systems and among dissimilar non-primary radio systems.
• Spectrum sharing with licensed radio systems: The opportunistic usage of spectrum by spectrum-agile radios should not cause harmful interference to the operation of licensed radio systems, thus requiring their detection even at very low signal powers. To improve detection reliability beyond what is possible today (for example, with carrier sensing in Wi-Fi), we propose the use of the incumbent's transmission signal characteristics to identify its presence. The spectrum-agile radio can thus adapt its spectrum usage to external factors, in this case presence of an incumbent.
• The need for reconfiguration of spectrum-agile radios after deployment may arise if new primary radio systems are allowed to operate in an unlicensed frequency band. Examples of such new primary radio systems include public safety systems. While the design of such reconfigurable systems remains a challenge, preliminary solutions do exist.
• Spectrum sharing among non-primaries: Another key opportunity for spectrum-agile radios is to design-in coexistence among yet-to-be-defined dissimilar non-primary radio systems sharing the same spectrum (similar to Wi-Fi and Bluetooth with equal rights to spectrum access in the ISM band). Dissimilar spectrum-agile radio devices need to coordinate spectrum opportunities, which may require communicating the existence of opportunities, or may be based on voluntary sharing rules such as spectrum etiquette.
• Adaptation to evolving FCC policies: Our vision is based on the DARPA XG program, a framework for flexible radio regulation. FCC policies are described in a machine-understandable form with the help of XML-files. Regulatory rules that are described in such files are interpreted by spectrum-agile radios, and, upon changes, the radio devices reconfigure their way of operation to adapt to the new regulatory environment.

FCC policies may evolve either due to the introduction of new primaries in unlicensed bands as discussed above, or due to the availability of new technologies that allow better use of spectrum, for instance, to more reliably detect existing primaries.

Kiran Challapali (kiran.challapali@philips.com) is principal member of research staff/project leader; Dagnachew (Dan) Birru (Dagnachew.Birru@philips.com) is senior member research staff/project leader; and Stefan Mangold (stefan.mangold@philips.com) is senior member of research staff in the Wireless Communications & Networking (WiCAN) group of Philips Research-USA (Briarcliff Manor, N.Y.).

See related chart