Technology Task Force Report
To the
2001 Annual Meeting
of the
ARRL Board of Directors
The Amateur Radio Technology Task Force was created at Minute 54 of the 1999 Annual Meeting. Since that time, a considerable amount of input was received suggesting a variety of new technologies for the amateur radio service. The Technology Working Group, created at Minute 55 of the 1999 Annual Meeting, was assigned the responsibility of evaluating each proposal. Approximately 150 proposals were received ranging from broad philosophical suggestions to detailed technical proposals. These fell into nearly 30 topic areas, but the majority were concentrated into three general categories, with the vast majority (nearly 85%) suggesting advancement in digital technology.
Each recommendation is worthy of some level of developmental consideration, however, as the TWG noted, a "buckshot" approach (i.e., do everything) is unlikely to succeed. It is therefore our recommendation, based on the TWG's evaluation and report contained in Attachment A, that we proceed with development of three specific technologies:
High Speed Digital Networks
The TTF recommends that the Board approve the creation of the High Speed Digital Network Working Group. This group will be responsible for development work in this area with a goal to achieve speeds of 384 kbit/s in the amateur allocated spectrum to permit multimedia communications.
Digital Voice
The ARRL Board unanimously approved a recommendation from the TTF at Minute 42 of the 2000 Second Meeting to proceed with the development of digital audio. President Haynie subsequently appointed a working group consisting of Doug Smith, KF6DX, Jesse Morris, KC5GTK, & John Gibbs, KC7YXD. The groups initial report is included at Attachment B. The TTF recommends that Charles Brain, G4GUO and George Bednekoff, AC5WO, be invited to join the working group.
Software Defined Radio (SDR)
The TTF recommends that the Board approve the creation of the SDR Working Group. The group will be responsible for developing the objectives identified in the TWG report of Attachment A. In addition, we recommend that Mr. Rinaldo investigate the benefit of becoming a member of the SDR Forum. The SDR Forum is an industry association comprised of representatives from Land Mobile and Cellular type groups.
The TTF also believes it is beneficial that we foster our relationship with the FCC's Office of Engineering and Technology. A white paper should be prepared and presented to responsible people at OET during the next trip to Washington by President Haynie outlining these areas identified above. It is imperative that we keep OET abreast of our progress in these areas and offer the League's assistance to them in anyway possible.
The ARRL seriously needs a technical strategic plan. The Stratcom committee has submitted a detailed report identifying areas of importance to the amateur service. The TTF should work closely with Stratcom during the next year to better identify and further develop our microwave frequency utilization database. Since First Vice President Harrison routinely attends and presents papers at all of the VHF/UHF/Microwave conferences, he will prepare a paper for presentation at these conferences enlisting their assistance in this area.
There has been interest in the TTF's work from the general membership. Now that the TWG has completed its evaluation and submitted its final report, we must publicize the results in a QST article and update the ARRL Web Page. The TTF will work with staff in the coming weeks to address this area.
Finally, the TTF wishes to thank the TWG for completing its work in achieving our results to date.
Respectfully Submitted,
Amateur Radio Technology Task Force:
Joel Harrison, W5ZN, Chairman
Tom Frenaye, K1KI
Jim Maxwell, W6CF
Paul Rinaldo, W4RI
Ed Hare, W1RFI
Attachment A
Technology Working Group Report
to the
Technology Task Force
Radio Amateurs have responded to the Technology Working Group (TWG) survey with a strong message that two technologies are needed in the Amateur Radio Service: higher digital speeds and digital voice. This is also supported by the following:
CANDIDATE TECHNOLOGIES
Respondents to the TWG survey identified numerous technologies having merit for development. A "buckshot" approach (i.e., do everything) is unlikely to succeed. Thus, it is desirable to select specific technologies where (a) a need has been identified, (b) amateurs can make a difference, and (c) where the League's leadership is required.
1.0High-Speed Packet Radio
High-speed digital transmission rates were identified most by the respondents to the TWG survey. Amateur packet radio has been topped off at 9.6 and 56 kbit/s for the moment. A speed of 384 kbit/s is desirable to permit full frame, full motion, digitally compressed video, thus would be a good target for multimedia communications in the Amateur Radio Service. Lesser speeds also could be used at some sacrifice of motion or resolution in television but capable of handling data and voice traffic.
The following actions should advance this project:
2.0Digital Voice
Digital voice was the technology second most identified by the TTF survey respondents. Preliminary research indicates that of three digital voice technologies (LPC-10, MELP and AMBE), APCO Project 25 and other systems have selected Digital Voice Systems, Inc AMBE (see http://www.dvsinc.com). Ordinarily, it is desirable to select a vocoder that is available from multiple sources, but that does not appear to be a viable option.
Here are some steps that should expedite digital voice activities:
3.0Software Defined Radio (SDR)
Most software receivers have an analog front end consisting of band-pass filtering, a low-noise RF amplifier to set a low system noise level, a local oscillator and mixer to heterodyne the signal to an intermediate frequency (IF) where analog-to-digital (A/D) conversion, digital filtering and demodulation takes place. Recently, however, there are some software receivers that perform A/D conversion immediately after the antenna.
Most efforts by individual amateurs have focused on use of digital signal processing (DSP) boards (evaluation boards and PC soundboards) to perform digital filtering and demodulation, such as for RTTY and data modes, at baseband (not even at IF). Amateur Radio equipment manufacturers have included DSP features in transceivers.
FCC OET Chief Hatfield has also identified software-defined radio (SDR or simply software radio) as a technology to which amateurs can contribute. SDRs, cognitive radios and similar devices are also the subject under consideration by the FCC TAC Spectrum Working Group. (See http://www.jacksons.net.tac.) (Also see Hatfield's keynote address at the 2000 SDR Forum meeting, http://www.fcc.gov/oet/speeches/sdrforumsph.htm.)
The FCC has initiated a Notice of Inquiry regarding software-defined radios (ET Docket No. 00-47) and the League has timely filed comments. The NOI asks a number of detailed questions about software radio technology capabilities, commercial deployment, international developments, facilitation of interoperability, movement toward standards, improvement of spectrum efficiency and impact on FCC rules. The Commission is understandably concerned about privacy of communication, prevention of using unauthorized frequencies and equipment certification issues.
SDR Forum is a nonprofit group dedicated to the development of advanced wireless systems. The chairman is Stephen Blust, WA4IPI. A membership fee of $2000 (for nonprofit organizations) is required to participate or obtain documents.
Here are some possible actions:
CONCLUSIONS
a. ARRL needs to develop a technical strategic plan to identify key technologies where amateurs can make a difference.
b. The League should work with the FCC Office of Engineering and Technology.
c. QST and QEX should be used to communicate the need for amateur participation in new technologies.
d. As technical projects invariably need a "champion" to succeed, it is important to find and designate such individuals and appoint responsible groups.
e. Recognize that some of these technologies, particularly digital voice, have international implications and we should work with interested sister societies.
It has been our pleasure to serve the League in this capacity.
Respectfully Submitted,
Technology Working Group:
Rich Moseson, W2VU
Chairman
Attachment B
Digital Voice Working Group Report
to the
ARRL Technology Task Force
1-15-2001
Overview Certain technological goals for digital voice in Amateur Radio are fairly obvious. Among those are the performance advantages that we assign to digital modes, such as high recovered SNR during otherwise noisy conditions, error detection and correction, interference reduction, and so forth. Signal propagation varies so much over the amateur spectrum, though, that modulation formats and data rates appropriate for one band are inadequate for another. Further, conditions on a particular frequency may change rapidly with time.
Logistical goals involve another set of variables, such as money, time to implementation and recruiting experimenters. Getting digital voice into the hands of amateurs quickly is an attractive prospect; but it may conflict with our wish to find the very best solutions.
General Technological Goals
What's out there now? Does it do what we want? Well, what DO we want? We predict that digital voice will not be popular unless it: 1) Sounds natural and telling that the voice was coded is difficult; 2) Operates under reasonably low SNRs, 3) Is compatible with existing ham transceivers in the form of an affordable external box or kit; 4) Is something that programmers can put into their future designs in embedded form at low power consumption, and 5) Employs an open architecture so that future improvements are always possible via software or firmware. We have seen the success of Linux and what is possible with open systems.
Designers have been pretty hard-pressed to achieve the data rates required by current voice codecs on the HF bands. To get a digital voice system meeting the above to operate under busy HF conditions is a tall order, to be sure. Charles Brain's system requires a fairly high SNR on HF, but the modems he used are not terribly sophisticated by today's standards. Frederick Electronics (now NSG Datacom?) made reliable HF modems at 2400 bps and above, but they were not cheap. What is needed on HF is a very robust waveform to do it in 3 kHz of BW. We conclude that HF would be a tough introductory medium for Amateur Radio digital voice. At the other end of the spectrum, 35-mW microwave radios are very cheap these days: A Compaq or Samsung PCI card complying with 802.11b at 11 Mbps costs about $200. It plugs and plays and you are on 13 cm in nothing flat. That is enough BW for very nice DTV, but the spread-spectrum technology is unusable at lower frequencies. Possibilities for the future are limited by unavailability of development platforms. ITU-standard protocols, such as G.723.1, are out there. Internet protocol already supports digital voice modes, including H.323 and H.324, that may meet requirements; but IP may not be the most efficient system for the bands where most hams spend their time. Before discarding it, though, the TTF should think about how it would add network connectivity for UHF bands where data and voice emissions may coexist. Command, control, repeating and trunking capability come along for the ride with some of those systems.
It seems practical to adopt a system that operates well at a data rate of 9600 bps or above for VHF and UHF, and operates reasonably on HF at 2400 bps or thereabouts. Higher data rates should be possible at microwave frequencies. Extremely low rates, along with a store-and-play-back feature, would allow voice contacts over very bad links-- ones where voice communication was not possible before. Questions: Will this affect phone contests? Will digital voice contacts be counted the same as the analog? What happens when KF6DX lights off his digital-voice rig through the local repeater? Economy dictates that the modulation standard change with the data rate. It is likely that the coding scheme would also have to change. A graduated scale of data rates-- and therefor quality levels-- ought to be possible that allows one to enter the fray at whatever level one can afford. PC-based solutions should be sought, if possible, but it may be that many PCs do not have enough processing horsepower for the job if a modem is included.
DSP evaluation kits are a great place to start. Flexibility is the key. Many kits include data acquisition hardware (digital-to-analog conversion). Such a system is so generic that one can program it to be almost anything. Look at PSK31 and its development for a glimpse of what improvements are possible in open systems.
Logistical Goals
It seems logical to divide our task into short-term and long-term actions. Short-term actions should, of course, include those that make the long-term actions possible. A recommendation to approve expenditures and commitments of time to conduct experimentation must naturally begin with finding volunteers. We have a few now who would be willing to give of their time to do testing, both on the bench and on the air. Those experiments should determine what performance levels are possible with existing equipment and the magnitude of interference caused and sustained among both analog and digital phone stations. We further recommend the continuance of efforts to solicit writers on the topic and perhaps arranging for a survey in QST or on the ARRL Web site. The toughest part will be finding people to help communicate the message and do the work.
We can get copies of the Charles Brain/TAPR codec (no modem), H.323 hardware from several vendors and G723.1 gear from Analog Devices. APCO 25 gear is also available at a dear GSA price. As some of these units incorporate proprietary coding systems, though, we have another thing to consider.
It is easy to use a digital voice codec chip such as the DVC unit. It seems to embody many well-documented techniques, but we cannot be sure, since it is proprietary. We recommend an open system that allows for future improvements through software. As we consider coding methods, we must be careful to look at intellectual-property issues. We would not like to lock ourselves into a single-source solution, nor would we like to step on anybody's patent rights. The cellular-phone market has undoubtedly spawned many such patents in going digital. We recommend research into the legal state of the art on that subject; also, regarding what coding methods we may legally transmit. That is, can we transmit digital voice that uses a proprietary (non-public) code? We recommend contact with the FCC before testing begins in any case.
Long-term planning shall emphasize developmental aspects and Amateur Radio's dedication to them. That is one reason why an open architecture is so important. Interference issues must be carefully explored. In addition, paths for future growth must be clear. Our ultimate recommendations shall stress that our system encourages experimentation and makes it reasonably practical. After reducing the results of our initial testing, we will further explore development plans for an external digital voice codec in software. We have a few general comments about that, as follows.
Speech Coding Research
A lot has been done on this in the last few years. Development is driven by the desire to pass audio and video information over the Internet at relatively low data rates, by the DTV gang, and by the recording industry. It is difficult, though, to make digital voice a BW-efficient proposition. While digital modulation has the advantages mentioned above, it may not ever be the most BW-efficient way to send the information.
We nonetheless recommend increased contact with individuals in those and other related industries. The trade magazines are full of articles on MPEG coding and similar technologies. We should try to get a few for QST and QEX.
Modulation Formats
Orthogonal frequency-division multiplexing (OFDM) is getting a lot of press lately. The next best bet these days seems to be straight M-ary PSK, much like what Brain and Talbot used. Many great minds, however, have wrestled with the linear nature of that modulation and the high peak-to-average ratio it produces on SSB. Data waveform shaping (MSK, GMSK) helps a lot, and we think that sort of thing will be hard to beat in the short haul.
Modems may be implemented in firmware, which makes it possible to build an entire system on a generic DSP evaluation platform. Economical stand-alone modems are currently available, though, at speeds sufficient for VHF and above. We propose to experiment with some of those.
Protocols
With an open architecture, a number of protocols can be supported. Market forces may ultimately decide what works the best and what costs the least. It is desirable that some kind of standard preamble or identifier be embedded in digital voice transmissions. Such identifiers allow some measure of synchronism with what follows, but digital voice modulation ought not be solely dependent on one, initial preamble. Synchronizing aids and identifiers are usually embedded continually in the data stream with little overhead.
Respectfully submitted,
Doug Smith, KF6DX
Chair, ARRL Digital Voice Committee