UWB - second link...
____________________________________________________________
Messing with Spectrum
Is ultrawideband a brilliant revolutionary technology ? or a potential network disaster?
By Dan Sweeney
Larry Fullerton "is perhaps the keenest scientific mind I have ever encountered."
So proclaims Ralph Petroff, CEO of Time Domain Inc. (Huntsville, Ala.). Since Petroff?s father was a colleague of Werner Von Braun and he personally knew the principal architects of the U.S. space program, that?s "saying something," he says. And that?s why Petroff is using his personal fortune to back Fullerton?s business case.
Fullerton is a radio engineer with about 25 years of experience in an arcane area known as ultrawideband. His innovations are expressed in a number of patents specific to ultrawideband (UWB) radio ? a technology which until recently was scarcely known, even to most radio frequency (RF) designers.
Ultrawideband?s proponents claim existing systems meet FCC standards for spurious emissions.
Ultrawideband, according to its champions, may revolutionize broadband wireless by putting unlicensed broadband wireless operators on a par with licensed millimeter microwave carriers in terms of sheer throughput, and by removing the principal limitations of unlicensed networks: interference from other operators, high susceptibility to fades, limited throughput rates and high equipment costs. Furthermore, because the technology permits millions of non-interfering orthogonal codes, UWB promises almost unlimited network capacity without the cost premiums associated with cell division, adaptive array antennas and other capacity-enhancing expedients used in conventional wireless networks. If the more extravagant claims are to be believed, UWB provides a nearly perfect communications infrastructure that should obsolesce all other options.
Can It Be True?
"The hype is entirely justified," says DeWayne Hendricks, partner in the Dandin Group, an ultrawideband systems integrator out of Monterey, Calif. "Ultrawideband radio transmission combined with software-defined cognitive radio terminals offers significant advantages over the traditional model which treats spectrum as property."
sidebar
______________________________________________________
Ultrawideband
transmission techniques reportedly afford:
Very high throughput rates ? perhaps up to 2 Gbps
High immunity to multipath distortion
Excellent in-building penetration
Low radiated power
Inherently ultra-precise radiolocation capabilities
Co-existence with narrowband transmissions without interfering with them
Extremely simple transmitting apparatus
Low cost, single-chip receivers
___________________________________________
end sidebar
Is Hendricks speculating? Not really; his company is building a nationwide, ubiquitous network in, of all places, the South Pacific. "We have a contract to provide high-speed Internet access, telephone service and video on demand over 30 Mbps connections to every household in the Kingdom of Tonga," he states. "The system will be largely complete next year.
"The Crown Prince [of Tonga] reviewed all of the other options, wireline and wireless, and concluded that this was the best available solution," adds Hendricks. "What we do there will showcase the technology to the rest of the world."
If all of this seems incredible, Craig Matthias, analyst with the FarPoint Group (Boston) who has authored a report on UWB, also sees its potential. "We think it?s well worth watching, though it?s by no means a slam-dunk at this point," he says. "There are real issues of interference, especially with GPS [global positioning system] signals, and if the ultrawideband transmissions have to be filtered to mitigate interference, then the performance of the systems will be compromised. But I think it has promise."
UWB has its detractors, among them Jeanne Fischer, spokesperson for SBC Communications in St. Louis, Mo., and Gary Klein, vice president of government and legal affairs for the Consumer Electronics Manufacturers Association (Arlington, Va.). Both filed objections with the FCC fearing that ultrawideband transmissions will gravely interfere with licensed television broadcasts and mobile telephone transmissions ? even public safety and military communications.
"Whether it works or not, there is no way the FCC is ever going to approve it," says Bill Frezza, analyst at Wireless Computing Associates (Yardley, Pa.). "These guys are taking the same tack as the spread-spectrum guys took 10 years ago: ?We can coexist with licensed services over the same bandwidth.? Well, the spread-spectrum guys got dumped into the unlicensed bands. That?s not going to work with ultrawideband, because they have to have the bandwidth for their systems to function. Which means it isn?t going to happen ? ever."
The chief stumbling block that UWB must overcome before it sees any significant commercial deployment is its relationship to spectrum allocations. Since UWB occupies huge slices of spectrum already allocated to a range of licensed services, it may require special exemptions from the FCC if it is to exist. Proponents insist that the radiated power of the signal can be kept below permissible levels of background noise and thus within Part 15 of the FCC rules governing spurious emissions from electronic devices not intended for radio service. But so far nobody?s willing to set up a network and test such assumptions in court.
sidebar
__________________________________________________
Legal Free for Alls
Sept. 1, 1998 ? The FCC releases a Notice of Inquiry (NOI) to investigate the possibility of revising Part 15 of its Rules to permit UWB radio systems to operate legally in the United States. Part 15, which forms the legal basis for unlicensed broadband wireless services and private networks, stipulates that radios must accept whatever interference is received and correct whatever interference they cause. The rule allows operators to transmit without a license and without the need for frequency coordination.
Originally envisioned for governing short range, low-power private equipment, the rule now encompasses a number of long range, high-capacity networks which can tolerate considerable interference through use of advanced modulation techniques. The FCC also has allocated certain bands for unlicensed operation which are closed to everything but data communications devices.
All unlicensed radio equipment used in public or private networks operates within well-defined bands. The difference is that in unlicensed systems, the FCC puts no restrictions on the number of users or their physical location; it?s a free for all, with power constraints supposedly preventing anyone from monopolizing the spectrum and, with advancing technology ? more specifically, spread spectrum techniques ? enabling tolerance of interference.
UWB would radically change the application of the rule in that the unlicensed transmission would no longer be confined to restricted unlicensed bands but would be permitted to extend across the usable spectrum.
When commercial CDMA became the topic of discussion in the early 1990s, advocates maintained that CDMA could coexist with simultaneous narrowband transmissions. Indeed, the FCC entertained the notion of joint occupancy of satellite telephony bands by CDMA and TDMA systems, but it never occurred.
With CDMA, restriction to licensed bands has been the norm. Within the CDMA bands, signals interfere only with each other and the issue of coexistence is moot. But in UWB transmissions, the signals interfere with everything ? even one another.
Television broadcasters view ultrawideband with alarm. So do users of GPS devices which occupy 20 MHz of protected bandwidth and which must be able to receive extremely weak signals. Additional broadband noise in the GPS spectrum, it?s claimed, will cause the devices to malfunction, resulting in potential aviation and maritime mishaps.
The FCC plans to field-test the devices this year. No one knows how the commission will rule. "They haven?t been friendly in the past," notes Jerry Ross, president of Anro Corp. (Sarasota, Fla.), and a pioneer in ultrawideband development.
Nevertheless, the FCC was clearly informed of recent developments and on the technical subtleties of ultrawideband, and the NOI?s text acknowledged some of the claimed advantages. The tenor of the document was not dismissive.
Some manufacturers expect a partial accommodation, with certain bands off-limits. This presents a problem, because the employment of conventional notch filters engenders time dispersion, smearing the pulse and garbling the code. Most believe restrictions are possible, and any company which expects to survive had better come up with a solution.
________________________________________________
end sidebar
The FCC at press time was considering modifying the rules to accommodate UWB. In the meantime, it remains the province of small entrepreneurial companies, failing to win the commitment of major wireless infrastructure providers. In fact some of the "majors" are downright hostile.
"No question there is opposition to ultrawideband ? a lot of it coming from incumbents wed to traditional technology," says Hendricks. "I?ve been to conferences on ultrawideband where speakers from major companies were suddenly yanked off the agendas. Top management does not want to be seen as endorsing ultrawideband."
Yet, Petroff says the tide is turning. "We?ve been visited by 318 organizations thus far this year," he asserts. "Those included top telecom and consumer electronics manufacturers. Since our own intention is to provide a chipset, not to compete in legacy markets, we see no reason that they should oppose us."
Making Market
Ultrawideband was not developed for broadband data. Most applications, real or proposed, are subsets of radar and radiolocation. UWB?s unique ability to penetrate physical obstructions, including walls and earth embankments, coupled with its high spatial resolution and precise distance-gauging capabilities, make it ideal for finding and tracking objects. Work has been done to develop radar rescue devices, in-building surveillance, collision-avoidance radar, devices for measuring liquid levels in tanks and even radio-positioning tags.
Practical design and deployment of communications networks lags far behind, however. Simple point-to-point communications over UWB radios in the military go back to the 1970s, but applications were confined to voice and have not involved extended high-capacity networks or high throughput rates and mixed traffic. Even within the ranks of its supporters, UWB?s potential is subject to disagreement.
Martin Rofheart, CEO at XTremeSpectrum (Greenbelt, Md.), believes UWB is "a poor choice for wide area networks. We see it succeeding in the[personal area network/local area network] space. We don?t believe that effective directional antennas can be constructed." This statement is disputed by others.
"Probably the home network market is the best niche for the technology," adds Rofheart.
Bob Fleming and Cherie Kushner, co-founders of Aether Wire Technologies (Nicasio, Calif.), a UWB equipment manufacturer emphasizing radiolocation, agree: "We think ultrawideband will perform much better indoors and will find greater acceptance in short-range applications."
Hendricks and Petroff say otherwise. "I don?t think ultra-wideband is necessarily the best wireless technique for very long-range transmissions," acknowledges Hendricks, "but ultrawideband can operate over distances of several miles."
Petroff is optimistic about metropolitan area implementations and believes that in some instances wide area connections might be feasible. "We?ve been able to transmit at distances of 15 km at radiated power levels of 50 millionths of a watt," he says. "That?s within the Part 15 Rules. Longer distances would require more power. What we?d like to see is a different set of rules for rural areas. That way ultrawideband could provide low-cost access to people who aren?t being served now."
Obscure Origins
The origins of ultrawideband are a matter of some dispute. Such researchers as Henning Harmuth, Jerry Ross, Robert Fontana and John McCorkle claim theoretical and developmental work that predates Fullerton?s involvement, and published articles go back to the early 1960s.
While Petroff acknowledged the work of others in an interview for Wireless Americas, he makes little mention of it in his numerous statements to the popular press, including a piece in USA Today which granted him a cover shot in 1999 and brought ultrawideband to the public?s attention. According to the newspaper, UWB is, at least in practical civilian embodiments, the intellectual property of Time Domain Inc., which will pioneer the technology in broadband communications and other applications.
But several other, less-known entities are racing to develop practical high-speed data communications equipment, including XTremeSpectrum, Multispectral Solutions (Gaithersburg, Md.), Lawrence Livermore National Laboratory (Livermore, Calif.), Aether Wire Technologies, the Dandin Group, UltraPulse (Vienna, Va.), and Interval Research (Palo Alto, Calif.). Firms in Israel and Scandinavia are said to be at work as well. XTremeSpectrum, Multispectral Solutions and Livermore Labs claim extensive experience in execution of special projects for the military.
But none of these companies has succeeded in developing commercial markets. Aside from some research projects, the impetus has come almost exclusively from the US government, and UWB?s techniques until recently have been shrouded in secrecy.
The Nature of Its Novelty
Ultrawideband consists of a large family of radio transmission techniques with certain core attributes. Signals are non-continuous ?that is, non-sinusoidal ? and occupy extremely wide bandwidths, generally in excess of a gigahertz. As with its distant cousin, CDMA, separate transmissions simultaneously occupy the same bandwidth, meaning that there is no attempt to assign transmissions to discrete spectral channels. Also, virtual channels are defined by orthogonal codes.
UWB transmissions for the most part consist of a series of extremely short bursts of energy, having the appearance of broadband noise. The bursts are non-continuous, with intervals of silence separating them. There is a conceptual and physical resemblance to early 20th century radiotelegraphy in which spark gap transmitters were used to blast out messages in Morse code; however, in modern UWB transmissions, code sequences occur billions of times faster and are automatically decoded by digital circuits.
"Ultrawideband is the next level of spreading," explains Hatim Zaghloul, president of Wi-LAN (Calgary, Alberta), manufacturer of high-speed wireless bridges (and a company which has no plans to implement UWB). "It is, in a sense, a progression from CDMA."
Comparisons to CDMA and other spread-spectrum techniques hold up only to a point. "Unlike all conventional radio signals, ultrawideband is non-sinusoidal," says Fleming. "You can?t truly resolve the signals into a Fourier series, and that has led to all sorts of statements from individuals grounded in traditional radio engineering to the effect that ultrawideband isn?t possible because a radio transmitter can only output sine waves."
"Ultrawideband is fundamentally different," says Petroff. "Spread-spectrum techniques still make use of carrier frequencies. Ultrawideband is really best understood as a carrier-less or baseband transmission. There is no carrier frequency as such."
Petroff alludes to the fact that, unlike CDMA, in which a fairly conventional, phase-modulated signal carrier is subsequently remodulated by a "chipper" that buries the original signal in a layer of broadband noise, a UWB signal has no underlying single frequency signal. Rather, only the noise bursts are transmitted; the temporal sequence of bursts determines the information that they will convey, not some underlying carrier in their midst. Ultrawideband is a pure digital radio system in the strictest sense.
No attempt is made to filter the output and restrict energy to a certain portion of the spectrum, and there is no tuning at either the transmitter or receiver. The element used to generate the signal, such as an avalanche transistor or a step-recovery diode, does not resonate at a center frequency, nor is it followed by reactive circuit elements to remove out-of-band energy, although the antenna will impose a filtering effect. The transmitter is in no sense a tuned circuit, and the concepts of in-band and out-of-band energy have little relevance except insofar as the UWB transmission occurs in the midst of narrowband licensed transmissions. In that sense, UWB could be considered out of band, because it has no spectrum assigned to it in the first place!
Because UWB systems lack an underlying carrier, the coded sequence defines the virtual channel and the information conveyed within it. So to receive a UWB transmission, the receiver has to lock onto a code rather than tuning to a center frequency. What occurs is a sort of sampling process in which an electrical gate opens and shuts at the coded intervals and registers all of the broadband energy in the individual bursts. In some systems, less than a single bit of information is contained in a single pulse; therefore, several pulses have to be registered and summed to get a detectable signal at the output of the receiver?s front end. Given the extremely low energy levels (averaged over the band) in the transmissions, separating the signal from background electrical noise constitutes a fundamental engineering problem when designing the receivers.
Bursty Nature
The basic code defines the channel, not the message. So to impress information on the pulse train, the pulses must be modulated. The most common way is to advance or retard pulses slightly in relationship to the underlying code (pulse modulation). In theory, amplitude modulation could be used as well; however, advanced modulation techniques employed in conventional bandwidth-limited radio transmissions, such as 64 QAM (quadrature amplitude modulation) of 16 PSK (phase shift keying), are not practical in ultrawideband. For these reasons, achieving high throughputs becomes a matter of sending out a lot of bursts very quickly. The bursts must be very short ? two nanoseconds is common. Since the briefer a pulse the wider its bandwidth, throughput speed is directly related to its bandwidth.
"In either case, you?re talking about occupying 1 GHz or 2 GHz of spectrum to get the high data rates," says Hendricks, "but the question is, Where would you rather be? In the millimeter wave region? Or in a region centered around 2 GHz, which in radio spectrum is beachfront property? At the lower end of the spectrum you utilize far less power, achieve excellent in-building penetration and avoid rain fade. Furthermore, the radios are cheaper to manufacture."
sidebar
__________________________________________________
Opposition Camps
Parties opposed to ultrawideband use in the public safety spectrum (764?776 MHz and 794-806 MHz) are proposing one of two solutions.
Solution 1 ? Conduct thorough testing before UWB is allowed into commercial deployment. "There is no objection to allowing UWB companies to thoroughly test and review potential interference to the Safety Spectrum in a controlled and non-safety-threatening way," claims the Air Travelers Association. "Any rulemaking or continued waiver for UWB that could? endanger the Safety Spectrum must first be supported by comprehensive, unbiased evaluation and testing of potential harmful effects of UWB on GPS operations." This proposal is being supported by Lawrence Livermore National Laboratory, the Air Transport Association (which represents the major U.S. airlines), United Airlines, Multispectral Solutions Inc. and Fantasma Networks Inc. (Palo Alto, Calif.).
Solution 2 ? Do not use the safety spectrum for UWB. According to published reports, some companies have determined that UWB does not need the Safety Spectrum for successful communications. Multispectral Solutions and Fantasma Networks also reportedly support this initiative.
____________________________________________________
end sidebar
Individual bursts can take a number of different forms, but they are almost always impulsive ? that is, the burst occupies the entire spectrum bounded by its highest and lowest frequencies. So unlike the pseudo-random noise produced by a CDMA spreading code, the UWB burst is unstructured. The code is present in the sequence of bursts ? not in the distribution of energy within the burst.
In most systems, the bursts are symmetrical with a peak of energy at the center frequency. In many systems the center frequency also defines the bandwidth ? in other words, an impulse centered at 2 GHz will also be 2 GHz wide, and thus will extend from 1 GHz to 2 GHz. Theory states that only by making bandwidth at least equivalent to center frequency will maximum resistance to fades be achieved.
As with CDMA transmissions, the bandwidth of the impulse also defines its spreading gain (its immunity to interference). In existing systems, that may range as high as 50dB, which means that UWB signal may be detected at levels some 50dB below an interfering signal.
Because the impulses encompass enormous numbers of contiguous frequencies, Rayleigh fading (multipath fading) is not a significant problem because it cannot affect all frequencies simultaneously ? the phase of reflections never being uniform for different frequencies simultaneously propagated through the same space. Moreover, an impulse will still be detectable after sharp attenuation of its higher frequencies ? a characteristic which permits transmission of UWB signals through walls. Fractional wavelengths of the lower frequencies are sufficient for the detector circuits to register the impulse.
The Interference Issue
The nature of the burst ? short, discontinuous and dispersive ? makes determination of acceptable power limits difficult when conventional standards are invoked. Debate rages on the likelihood that UWB will interfere with licensed carriers and operators.
Proponents claim that existing systems easily meet FCC standards for spurious emissions (an interesting notion when one considers that UWB transmissions are intentional) while opponents reply that such standards were never intended to apply to transmitters and that a proliferation of UWB equipment would sharply raise the general level of RF interference and play hob with licensed radios.
The issue is further confused by the way radiated power is defined. Proponents like Petroff who cite power levels in the microwatts discuss power only at single frequencies considered individually. But aggregate power across the entire spectrum occupied by the impulse is orders of magnitude higher, at least momentarily.
While proponents insist on power measurements of discrete frequencies in the spectral domain, they generally want averaging in the time domain, arguing that since the UWB transmitter broadcasts intermittently rather than continuously, the peaks representing the impulses should be averaged, with zero output levels representing the intervals between the pulses. Since the narrower the pulse the higher its momentary energy level, any transient peak energy measurement is going to be fairly high and will show the system in an unfavorable light ? so it?s to be avoided if possible. Many UWB proponents want it both ways so that the measurement method will always support those reassuring figures of microwatts or even nanowatts.
If, on the other hand, time domain averaging is not permitted and power levels across the spectrum are aggregated, UWB transmissions within a metropolitan area will require power outputs similar to those of conventional high-speed wireless data systems: one to several watts. The issue then boils down how existing data and communications equipment will function in the presence of pervasive electrical interference that is dispersive in the frequency domain but concentrated in the time domain.
How will they weather a bombardment of nanosecond spikes? Proponents cite private tests that indicate the harmlessness of the transmissions at real-world power levels. The FCC has scheduled field trials to try and resolve the issue.
Many Questions
On an intellectual level, UWB exerts immediate appeal. It is a fresh, divergent approach to radio communications. Although the mathematical theories behind it are complex, the principle of operation is simple and straightforward.
But does it work? Is it superior to conventional techniques? Will it scale to large, heavily trafficked networks?
The ability to penetrate walls has been decisively demonstrated, and high immunity to multipath seems likely in theory. The enormous number of codes made possible by the bandwidth may support the claim that the systems would greatly exceed the capacity of conventional wireless networks. The real issue is interference, for the basic technology depends on peaceful coexistence with existing licensed wireless services. Until that issue is resolved, ultrawideband will remain experimental in the United States.
americasnetwork.com |
