In contrast to conventional radios, software radios process signals mostly in the digital domain, making it possible to put radio under software control. A number of companies in Japan and Europe described their rationale for software radios and outlined areas of research that were viewed as critical for the development of software radios. Alcatel, NTT, Daimler-Chrysler, NEC, Toshiba, Matsushita, Philips, Nokia, IMST all reported ongoing research and development activities on software radios. U.S. companies that described efforts in this regard included Hughes and Rockwell Collins. At the time of the WTEC study, most companies viewed software radios as reconfigurable, multimodal, multichannel, or multiband devices. This view does not call for radios that dynamically alter error control algorithms or modulation formats as a function of the channel and traffic characteristics.
The development of a reconfigurable, multimodal, multichannel or multiband radio depends critically on high resolution, high bandwidth, high dynamic range, A/D converters. Marcell Pelgrom at Philips has forecast that the trend in higher resolution/bandwidth A/D converters will continue for another six to ten years, with three orders of magnitude improvement still left. Besides A/D converters, Alcatel drew attention to the need for frequency synthesizers and good integrable low noise amplifiers (LNAs) as additional enablers for software radios.
Dr. Hori at NEC described an interesting application of the software radio concept by calling for studies of systems in which reconfigurable radios are deployed only at base stations and not in the handsets. Base station radios would then be expected to recognize the handset type and adjust themselves to the protocol and system for which the terminal was designed. This would obviate the need for reconfigurable radios in handsets and allow for smaller and cheaper handsets.
Bo Hedberg at Ericsson, speaking about the power efficiency of multicarrier power amplifiers, reinforced Dr. Hori's view of the asymmetric deployment of software radios. Hedberg also concentrated on the base station, because he thought that the software radio solution is very costly for terminals. The technology needs to mature before it can be used at the terminals. Hedberg pointed out that power efficiency is important, not just at the mobiles but also at base stations. Apparently, the air conditioning required to keep the multi-carrier power amplifiers (MCPAs) cool at the base stations would increase the price significantly. Hedberg estimates that half the cost of future base stations will be in the MCPA. Hedberg concluded that a lot of research will be needed during the next several years to improve the MCPAs.
At the time of the WTEC study, the companies visited viewed software radios as enabling the production of multimodal, multichannel or multiband terminals. The focus is on using a software reconfigurable terminal to harmonize dissimilar standards in one lightweight, low cost package that can be mass produced and reconfigured in software. The realization of this relatively modest objective requires additional progress on (1) high resolution, high bandwidth, high dynamic range A/D converters, (2) frequency synthesizers, and (3) better low noise amplifiers.
For a time, software radios are likely to be deployed only at base stations. In this scenario, the subscriber would be shielded from the high cost, complexity, and power requirements of software radios. A primary advantage of such an asymmetric deployment is the ability to support older generation handsets even as service providers roll out newer generation services.
The base station's ability to sense and adjust to the terminal's capabilities (protocol, frequency band etc.) could be used to create new competitive models for cellular services in which users, regardless of the type of terminal they carry, can roam across competing systems. This would be in contrast to approaches in which subscribers have to make a hard decision between competing providers that use incompatible air interfaces and offer very good coverage limited to certain geographical regions. Such a model may help the cellular industry as a whole to grow by offering more assured service to subscribers.
Once hardware requirements for the current generation of software radios have been met, the rationale for software radios will evolve from lowering costs to enhancing performance at lower cost as the demand for services starts to expand and saturates the available capacity. This objective will require more adaptive software radios that will be capable of switching quite rapidly between modulation formats, error control algorithms and other such features as a function of the channel, traffic and possibly battery characteristics. The additional technical breakthroughs required will be in the area of software control of the underlying hardware.
The task of assessing the relative status of U.S. companies in the global arena is fraught with limitations. The need to provide global communications services provides a strong incentive for telecommunications companies to operate as global enterprises. Therefore at a certain level, the distinction between U.S., European, and Japanese companies is quite artificial. Yet, in the context of communications, it is quite instructive to examine the relative impact of regulatory developments, which do differ from nation to nation, on technological needs.
In an article titled "Software radio technology: a European perspective," Tuttlebee (1999) argues that the need for software radios is most acute in the United States where multiple second generation digital systems compete in a market that has yet to be saturated. Such a market could be more efficiently served through the use of software reconfigurable handsets. On the other hand, European standardization of GSM obviates the need to develop software reconfigurable for multimodal, multi-standard handsets since the European multiband GSM allocations can be accommodated with modest enhancements at the RF end without incorporating the full range of reconfigurations that software radios may be called upon to effect in the U.S. market. Therefore, Tuttlebee claims that software radios are likely to emerge first in the United States and then in Europe when there is a demand to support user roaming between GSM and third-generation UMTS systems.
To get a sense of the current state of the research activity, the WTEC panel analyzed publications on software radios that have appeared in the recent literature, using the online INSPEC service, which tracks "articles in over 4000 scholarly, journals, conference proceedings, books, reports, and dissertations in physics, electrical engineering and electronics, computers and control, and information technology." Over 1400 publications on "software radio" appeared during the period 1995 to 19993/4almost one new publication per day.
Of these 1400 publications,
With regard to author affiliations,
All but 90 of the 1400 publications were in English. Of the 90 non-English publications, 27 were in Japanese.
These numbers suggest that current publications activity is largely in line with the software radio needs communicated to the WTEC panelists: RF components, smart antennas, protocols, DSP algorithms and base station deployment. U.S. leadership in the field appears to be consistent with Tuttlebee's assessment of the critical need for early deployment of software radios in the United States. The extensive Japanese development activities discovered by this panel during the site visits appear not to be reflected in publications.
Factoring in all the information available at this time, the WTEC panel concluded that U.S., European, and Japanese companies were on par with regard to software radio technology.