In the responses to NTIA's Inquiry, few comments were submitted on the impact of technology on spectrum management.[EN640] Nearly all comments were on particular technologies affecting their own requirements. Implicit in most comments was the sense that technology, in the aggregate, is not solving the problems of spectrum scarcity. Commenters noted that the aggregate needs of existing and new services are increasing at a rate such that, despite the infusion of new technology that increases spectrum efficiency, additional spectrum allocations for these services are still being requested. Nonetheless, the application of new technology to telecommunications systems has reduced spectrum needs, and continues to hold the promise of satisfying certain radiocommunications demands without additional spectrum. The reduction of spectrum needs is typified by the trend toward optical fiber replacing radio links; the trend of digital modulation replacing analog modulation makes spectrum use more efficient.
The New Technology Directions Committee (NTDC) of the IEEE Technical Activities Board is a group whose views on future trends proposed several technological challenges, and merit serious consideration. Of the seven "grand challenges" facing electrotechnology in 1993 evinced by the NTDC, at least four involve radiocommunications,[EN641] and thereby have spectrum management implications.
These challenges are familiar ones ranging from instant global communications for all to attaining a cheap, clean, and safe source of energy. The communication-related systems that would meet or support these challenges have familiar acronyms, e.g., ISDN - integrated services digital network, HDTV - high definition television, and CDMA - code division multiple access, many of which were proven experimentally feasible decades ago.[EN642] As with large-scale machine computation, application of these systems became practical with the introduction of high-speed analog-to-digital conversion, embedded micro-processors, and the rapid advance of solid-state electronics.
These challenges, along with other applications of technologies, are examined in this Part. Emphasis is placed on technology's ability to offer improved spectrum efficiency while permitting expanded service offerings, and the resultant impact on spectrum management.
During the past decades, the Federal Government has always kept some level of activity in the band above 20 GHz, mainly in the area of military R&D for high-resolution radiolocation and short-range radiocommunications. The national laboratories have been mandated to transfer this technology to the public sector where possible. Among the civilian federal agencies, there is the NSF-supported National Radio Astronomy Observatory that develops passive receiving systems operating well into the millimeter wavelengths.[EN643] NASA has recently launched the ACTS satellite that has the capability for testing ISDN and other new technologies at 20 GHz and 30 GHz.[EN644]
From the commercial sector, additional impetus to utilize the higher frequencies may come from wideband spread spectrum and other advanced megahertz-wide systems because they may face opposition to band sharing from the present users of the more congested lower bands.
Today, despite the fact that radio began at microwave frequencies over 100 years ago,[EN645] the spectrum above 20 GHz is still relatively underutilized. The general reasons for this condition, as found in scientific literature and echoed in the comments to our Inquiry, are
The Impact of Atmospheric Constituents on Radio Propagation
Beginning at about 10 GHz, absorption, scattering and refraction by atmospheric gases and hydrometeors (the various forms of precipitated water vapor such as rain, fog, sleet and snow) become the important limiting factors for electromagnetic wave propagation.[EN647] Figure 8-1 shows the specific attenuation in dB/km as a function of frequency for atmospheric gases (oxygen and water vapor) and liquid hydrometeors (rain, drizzle and fog).[EN648]Figure 8-1. Attenuation Due to Atmospheric Gases and Liquid Hydrometeors.
The atmospheric gases assume importance for telecommunications starting at frequencies greater than 30 GHz, with attenuation generally increasing with frequency (Figure 8-1). This high attenuation appears ideal for short distance line-of-sight communication (e.g., cellular radio), but rain which is even higher in attenuation (see Figure 8-1) and less predictable, would be a major consideration. For the space services operating outside of the atmosphere, negligible path losses and the narrow beamwidths and wide bandwidths possible in the higher, uncrowded frequencies are attractive at 100 GHz and above.[EN649]
Molecular absorption at particular frequencies (shown as peaks in Figure 8-1) are of interest to spectrum management for special applications, such as space operations. Satellite-to-satellite links operating in these bands would be relatively free of interference from co-channel terrestrial stations limited by the atmosphere.
Hydrometeors in the form of raindrops are the largest natural objects most frequently encountered in the lower atmosphere, with the larger drops more numerous in the heavier rainfalls.[EN650] Consequently, high attenuation is correlated with high rainfall and is significant even below 20 GHz. With the worldwide adoption of the 20-30 GHz band for the fixed-satellite service, raindrop absorption presents a continuing problem for countries with high rainfall rates. Propagation experiments between NASA's ACTS satellite and these countries will aid in gauging the long-term reliability of the higher frequencies.[EN651]
A spectrum management consideration in the near-future will be the appearance of hundreds, perhaps thousands, of transportable VSAT's with antenna diameters less than one meter using the 20-30 GHz radio links. Compared to terminals with larger antennas, the VSAT's are more susceptible to scintillation and fades because of the small aperture area of their antennas. The relationship among microwave scintillation, clear-air and rain, and antenna aperture smoothing is being actively explored by European researchers.[EN652]
U.S. investigators are looking at rain-induced fading itself, with low-margin earth stations in mind, and have compared a theoretical fade-rate model with 20-GHz Earth-satellite link data.[EN653] Recently, conventional radio-link methods (e.g., diversity reception and forward error-correction) as well as selection of appropriate modulations (TDMA, FDMA, and single-channel) were investigated for their effectiveness in VSAT fade control.[EN654] Further consideration of VSAT's and spectrum management will be taken up later in this chapter and the next one.
The smaller size of drops in fog or clouds compared to rain allows for simpler analysis enabling good agreement between the fog curve in Figure 8-1 and measurements by others.[EN655] Moist clouds are similar to elevated fog and would have a curve congruent to the fog curve in Figure 1.[EN656] Fog or moist cloud attenuation is generally less than for atmospheric absorption.
Millimeter-wave attenuation by falling snowflakes appears to be about the same as for rain for comparable values of equivalent rain rate. But since snowfall rates are usually less than 10 mm/h (a moderate rainfall), the effect is less severe than for rain in most cases.[EN657]
A spectrum management consideration in the near future is the effect of high-altitude cirrus clouds, commonly found in Earth-satellite paths, which depolarize radio waves at frequencies above 10 GHz. Reflections from the flat, horizontally-oriented surfaces of the crystals making up these clouds cause abrupt phase reversals of a radio signal,[EN658] thus degrading the performance of dual-polarization (frequency-reuse) systems. More importantly, digital systems in general are seriously affected because they are more sensitive to abrupt phase changes and reversals than analog systems.
Melting falling snow and ice are difficult to study in the field; nevertheless, knowledge of their effect is important in high-latitude countries along low-elevation radio paths. Towards this end, a theoretical approach using field data and observation has been developed by Canadian researchers.[EN659]
Technologies that Extend the Usable Spectrum
The extension of microwave techniques to millimeter wavelengths poses equipment design problems as severe as those associated with the move from the shortwave to the microwave band earlier in this century. Presently, system development has less problems between 30 to 100 GHz, where old techniques can be extended with some success. New devices and methods come into play mostly above 100 GHz.
The insertion loss of rectangular and circular waveguides, ferrite devices (phase shifters, isolators, etc.), and other standard microwave components become excessively high above 150 GHz because of size reduction and material losses. For example, a standard rectangular waveguide with a cutoff frequency of 100 GHz (at the basic TE10 mode) has an inside width of only about 1 mm, clearly making it a costly item difficult to interconnect mechanically and to maintain electrically pure. Also, magnetic and insulating properties undergo changes with reduction in size.
The dielectric line in the form of the microstrip, on the other hand, works well with solid state devices and is an important part of integrated circuit (IC) technology. The microstrip line is easily formed by printed circuit methods and integrated with solid state devices. But difficulty introduced by conflicting analyses has slowed extension to higher frequencies.[EN660] Nevertheless, technology has progressed to the point where commercial computer-aided-design software has been used to model IC's for field-effect-transistor antenna elements of microstrip arrays.[EN661]
The microstrip transmission line works well in integrated circuits up to about 100 GHz where radiation losses due to its inherent low Q can be tolerated. But other line components, such as the H-guide and groove guide, show more promise for integrated circuits at higher frequencies.[EN662]
With respect to spectrum management, the integration of dielectric line components and solid state active devices introduces flexibility in specifying and changing the parameters of a radio system. Two examples of a change in a system parameter are the bandwidth of a receiver stage and the radiating pattern and gain of an antenna. The action in the latter case could be effected by differentially altering the phase of active devices located at the antenna element of a phase array. A conventional antenna array usually has only one active source per array. An analogous situation exists for active filters that determine a receiver's bandpass characteristics. Moreover, microprocessor control for each active element would add real-time response to spectrum management capability.
The dominant role of silicon technology in low-cost integrated circuitry presently extends to about 2.5 GHz, and with improved fabrication may go above 10 GHz.[EN663] Presently, gallium-arsenide devices are extending monolithic microwave integrated circuit (MMIC) technology to higher frequencies despite the higher cost of the starting up material.[EN664] Meanwhile, a low-cost silicon MMIC with cutoff frequency approaching 20 GHz has been proposed for wireless systems.[EN665] At least for the next decade, silicon and galium-arsenide are the semiconductor materials on which MMIC technology is being extending into the higher frequencies.
Upgraded microwave power devices appear to be adequate as power sources for most millimeter wave applications at least to 300 GHz. Sources can be divided into vacuum tube and solid state devices, with the former being the primary means of generating power greater than about 20 W. Slow-wave devices (klystrons, magnetrons and traveling wave tubes) have a mature technology compared to the gyrotrons and the newer fast-wave devices for power applications. However, the requirement for scaling down by a factor inversely proportional to frequency, imposes limits on the amount of power that can be handled by a miniaturized device; hence, power generation at higher frequencies may depend on more efficient designs such as the fast-wave devices. Presently, traveling wave tubes for satellite broadcasting at 12 GHz are commercially available at 100-W continuous wave output,[EN666] and the capability to go into the EHF range has been demonstrated experimentally.
The trend toward more efficient high-power satellite transmitters at higher frequencies is enabling earth stations to function with smaller, more portable antennas; but proliferation of these VSAT-like stations may pose spectrum management problems in the future. For any given frequency, a VSAT antenna is subject to greater interference than a larger antenna because of its poorer directional discrimination.
For low power generation, such as in receiver circuits, improved versions of solid-state devices appear adequate for the millimeter band at least to the year 2000. Specific examples are the impact avalanche and transit time (IMPATT) diode, a variation of the avalanche oscillator, which can operate across the band at pulsed power of 20 W or better, and the transferred electron oscillator (familiarly called the GUNN diode), with somewhat less output and a more restricted frequency range.
Millimeter receiver circuits still largely depend on solid state technology. The familiar gallium arsenide (GaAs) field-effect transistor (FET), improved by better production techniques, can perform well into the lower millimeter band. Better-designed Schottky-barrier diodes are still usable as detectors and mixers. Research to increase receiver sensitivity by cooling circuits to reduce the intrinsic temperature noise is being carried out on Schottky diodes and Josephson junctions.[EN667] In digital circuitry, reduced bit-error-rates were reported for cooled GaAs metal-semiconductor (MES) FET's at 38 GHz.[EN668]
Some of the latest advances in broadband millimeter wave technology have been in low-noise high electron mobility transistor (HEMT) technology. At Comsat Laboratories, Clarksville, MD, a HEMT amplifier successfully generated a broadband (60 GHz) signal for an optical phase shifter designed to control transmission line phase variations.[EN669] In another report, a HEMT GaAs MMIC amplifier with a gain greater than 10 dB from 14 to 44 GHz had a noise figure of only 4 dB at 35 GHz.[EN670] MMIC technology also brings in the microprocessor control of millimeter components and devices. The subsequent control of telecommunication systems at all levels of complexity is the next step in the management of the spectrum.
Optical components and devices are attractive supplements to solid-state electronic circuitry because of their extremely wideband characteristics. Although usually associated with fiber optic transmission lines, they are also being integrated with electronic components and devices. Some quasi-optical components developed for the millimeter wavelengths are beam waveguides, attenuators, and polarization rotators. In small scale systems, optical switches have had definite success due to their high throughput, high-speed response, and immunity to electromagnetic interference. Incorporation into large-scale systems, such as those in commercial telephone central offices, will require careful analysis at all system levels.[EN671]
The technology of components and devices extending into the millimeter band (30-300 GHz) appears to satisfy spectrum needs at least to the year 2000. The uncrowded conditions in these higher frequencies have furthered the development of broadband systems using solid-state as well as optical techniques.
Summary of Technologies That Extend the Usable Spectrum to Higher Frequencies
The salient points of this chapter are summarized as follows:
Advances in semiconductor technology are major forces behind innovations in radiocommunications. New types of modulation, new uses for microprocessors, and other digital applications have, in turn, led to spectrally efficient radio systems.[EN672]
The distinction between modulation and microprocessor-based technologies is blurring with the trend to all-digital systems and microminiaturization of electronic devices and components. This is most apparent in the wideband systems conceived in the vacuum-tube era (e.g., spread spectrum), whose commercial realization came only with digitization and integrated circuitry. Microprocessors are a necessity rather than an adjunct for digital systems. For instance, central processing units (CPU's) control the spreading and despreading of a coded bit stream, a function that is not usually required in simpler modulation schemes.[EN673]
Spectrum-efficient technologies useful for mobile communications were reviewed in a recent NTIA report.[EN674] Many of the same technologies mentioned in that report are discussed herein. The first, Modulation Technologies, begins with narrowband methods and ends with ultra-wideband techniques. Then, under Microprocessor-Based Technologies, the focus will be on topics in the NOI comments, starting from baseband processing and ending with whole systems.
From the sprectrum management viewpoint, the sharing of spectrum between wideband and narrowband systems in the decade ahead is probably the most serious issue identified in this chapter.
Large gains in spectrum efficiency are possible by digital speech within the limits of a conventional 25-kHz channel. One example is the application of quadrature phase-shift keying compatible (QPSK-C) modulation to reduce the bandwidth requirements by one-half.[EN675]
The Enhanced Specialized Mobile Radio (ESMR) system recently fielded in several metropolitan areas shows promise as a very efficient spectrum saver.[EN676] A TDMA transmission will create six digitized voice channels in place of a single 25 kHz channel. Within a multiple low-power base station configuration, approximately six times the customer capacity of existing systems is claimed by the proponents of the ESMR system.
The demand for more information transmittal can be technically translated into more bits per second within a prescribed bandwidth. For the fixed service where the challenge of spectral efficiency and cost has been great, quadrature amplitude modulation (QAM) is the apparent choice over other alternatives with 256 QAM level commercially available now.[EN677]
The lack of higher-power linear amplifiers is limiting the move to 512 QAM and above for terrestrial fixed service since forward error correction and other techniques to overcome non-linearity are now optimized for 256 QAM.[EN678] This is a common problem for all high-level digital modulation requiring linearity across megahertz-wide channels. In satellite broadcasting, digital compression allows several TV signals in place of one analog channel; however, distortion may result when more than one carrier is processed by a satellite multiplexer. Improved traveling wave tubes and feed-forward, feed-back, and predistortion linearizers are promising solutions to high-level linear amplifiers, at least up to 20 GHz.[EN679]
Spread spectrum (CDMA) modulation spreads a signal across a band that is considerably wider than the information bandwidth. This results in a signal having a low-power spectral density in any narrow portion of the band.[EN680] Nevertheless, commenters in NTIA's Inquiry are concerned about such a wide signal sharing spectrum with non-CDMA spectrum users.[EN681] One developer of TDMA systems comments that a spectrum saving of up to 20 times, as claimed by spread spectrum proponents, needs validation, and the feasibility of sharing with non-CDMA systems awaits further field testing.[EN682] One commenter indicated that the optimum bandwidth of 8 kb/s coded voice traffic ranges from 1 to 3 MHz, so wider-range CDMA would not be able to accommodate more users on a per megahertz basis.[EN683]
Developers of spread spectrum equipment are also among the commenters.[EN684] One typical CDMA system soon to be field demonstrated is designed to locate land mobile radio vehicles in a band shared with other LMS systems.[EN685] Aside from the signal-to-noise advantage of high processing gain,[EN686] CDMA has good rejection of unwanted signals and a high degree of security.[EN687]
Other multiple access systems conserve spectrum within the bandwidth limits imposed on narrowband modulation and require no spectrum sharing to be implemented. From the comments to the NOI and the open literature, digital TDMA appears to be a favorite replacement for time-honored analog FM modulation in mobile radio.[EN688] Frequency division multiple access (FDMA) is also a contender in the mobile and mobile-satellite service.
Ultra-wideband radar and radio systems transmit one to a few extremely short duration pulses of a nanosecond or slightly longer. These pulses require a bandwidth of 50-100 percent or more of the center frequency and may span across bands allocated to several services.[EN689] By contrast, conventional radio has a bandwidth of 10 percent or less. Two recent examples of ultra-wideband systems are a synthetic aperture radar operating at the relatively low frequency band of 20-90 MHz useful for ground and foliage penetration,[EN690] and a commercial communications radio that transmits a single non-sinusoidal nanosecond pulse with a bandwidth of 1 GHz (see Figure 9-1, where peak energy is at fc).[EN691]Figure 9-1. A gaussian monocycle in the time and frequency domain
By spreading signal energy broadly across the spectrum, as in wideband spread spectrum emission, impulse radio has the same advantages of extremely high processing gain and relative immunity to fading. The implementation of pseudorandom coding (also common in spread spectrum applications) can add the desirable features of easy encoding and near-orthogonal discrimination of transmitted signals.[EN692]
However, at frequencies below several gigahertz, the ultra-wide bandwidth (on the order of a gigahertz) requires sharing the spectrum with many services using narrow and wideband systems. At this time, field tests against conventional narrowband and other new technology wideband systems would help to assess the spectrum sharing capabilities of ultra-wideband radar and radio. Also, tutorials published in widely-read professional journals can help spectrum managers understand this relatively little-known technology.[EN693]
Digital compression is a technique to reduce the bandwidth of an existing analog or digital channel by eliminating redundant information. It has significant implications for spectrum management because the technique works within the requirements of existing bandwidths. Compression is done on raw digitized information within a piece of equipment, and output can be to many types of digital processing equipment.
Comments to our Inquiry on compression technology came from several sources. Most were favorable, but none saw it as a panacea for spectrum sharing. In fact, a few cautioned that it would not lessen the demand for spectrum.[EN694] Compression removes audio or visual information that is redundant or irrelevant to the user's perception of quality reception. In digital audio broadcasting, as much as 80 percent of the raw data may be removed and still maintain a CD-quality performance.[EN695]
The greatest savings in spectrum is in digital video compression. Generally, the raw digitized data (bits or pixels) are partitioned into blocks which are periodically tested for redundancy; then bits are kept or discarded as necessary. Using video compression, one satellite transponder is said to carry up to 18 conventional TV channels, and requires only a bandwidth of 2 MHz to send a high-quality full-motion NTSC TV signal.[EN696]
Trunking, a concept borrowed from the wire telephony, maximizes the use of radio channels or trunks by assigning a call to an idle channel within a dedicated block of channels. The spectrum efficiency afforded by trunking techniques applied to the land mobile service was reported recently in a NTIA publication.[EN697] There is general agreement among the Notice commenters that mobile assignments not critical to safety or emergency service would benefit from trunking.[EN698]
Spectrum efficiency can be enhanced by digitizing trunked radio systems. For a trunked channel, this enables control and voice information to be sent simultaneously. In one example, half of the channel capacity of 9.6 kbps is required for voice information leaving the rest for control signaling, error correction, and other functions.[EN699]
Also important from the spectrum management viewpoint is the flexibility afforded by digital trunking in accommodating voice and data traffic on a real-time basis. A typical example is a recently introduced high-capacity trunking system that can control a mix of mobile radios, portable sets, and fixed console positions through two computers and a network of radio repeaters.[EN700] This system reportedly can handle up to 50 repeater sites, and existing trunking systems.
Comments to the Inquiry on mobile cellular and portable PCS technology were mainly focused on ways to facilitate band sharing among a mix of wideband and narrowband systems, and also with the fixed and mobile-satellite service. A favorite topic of commenters was the dynamic control of spectrum resources, e.g., transmitter power, channel and frequency switching, and antenna directionality. Two important topics brought up were the control of transmitter power and the lack of data on channel characteristics.
The comments received on controlling power include one on mobile operations near fixed microwave receivers which will be discussed in detail later.[EN701] In anticipating the overlaying of fixed and mobile channels, developers of cellular CDMA equipment have added power control to their mobile sets.[EN702] In one "open-loop" design, the mobile set adjusts its transmitted power only according to the received signal strength.[EN703] A more complex design features a closed loop for adjusting a mobile unit's power to the minimum for acceptable signal quality within a cell, and an open loop for smoothing the transition from cell to cell.[EN704]
One of the more pressing technical problems facing the cellular and PCS sector is the lack of information on the channel characteristics required for delivering high-traffic continuous service approaching wire-telephone quality.[EN705] The transmission environment encountered by users in motion when passing into and around large obstructions is a demanding one. High propagation loss, multipath from many directions, and the fluctuations caused by motion of the user are but a few characteristics that determine such factors as the choice of modulation (narrowband or wideband), location of fixed antennas (utility pole or mast), and type of service (video, voice, or data). Some trends and challenges brought out in a review of wireless communication were as follows:[EN706]
Upgrading the ALE radios used on HF frequencies should lead to substantial spectrum savings according to commenters.[EN707] The present ALE radios have spectrum-efficient selective calling and frequency selection features. By adding chirp capability, frequency channels may become even more efficiently selected[EN708] and the idle time on a channel reduced. Chirp is the periodic linear sweeping of the HF band for a fixed point-to-point HF link to establish the maximum usable frequency (MUF) for the path. Other dynamic control systems use different methods to monitor the HF path.[EN709]
The Intelligent Multiple Access Spectrum Sharing (IMASS) is an example of an active avoidance technique which reclaims underutilized frequencies assigned to fixed microwave services for use by new mobile services on a shared basis.[EN710] This "active avoidance" technique employs a special receiver to scan the microwave channels in an area of fixed microwave systems. Actual measurements are made to determine the amount of RF isolation between mobile and fixed radios, and any available channels are temporarily assigned to low-power mobile networks. IMASS is designed to accept only certain FDMA, TDMA, and narrowband CDMA protocols. On the basis of current information, broadband overlay service (wideband CDMA and impulse radio) is not generally recommended for operation in high-concentration fixed microwave areas.[EN711]
Summary of Modulation & Microprocessor-Based Technologies
The salient points of this chapter are summarized as follows:
Antenna design, long thought of as a static area of engineering research, has undergone accelerated growth in recent times through the use of computer-aided design (CAD). Antennas can now be designed as self-optimizing systems that respond to changes imposed by man and nature, rather than simply acting as passive couplers between the external environment and electronic equipment.
According to several comments to our Inquiry, an antenna's ability (or inability) to alter its radiation pattern is probably its most important characteristic from a spectrum management viewpoint.[EN712]
With the trend toward higher frequencies (hence antennas of smaller physical size), adaptive antenna designs are becoming easier to implement, since antennas will interact less in an unpredictable way with their immediate environment. Progress in simple applications is reaching a point where no clear distinction can be made between the antenna and its surroundings; for example, slot antennas can be designed flush with the surfaces of personal communicator cases (Figure 10-1).
In the future, the space-time alteration of an antenna's radiation pattern, both in transmission and reception, will become a powerful tool for spectrum management. The emphasis of this section is on electronically controlled antennas where most of the new technological work is being done, and ends with developments in extending the bandwidth of antennas to accommodate emerging wideband systems.
The greatest activity in antenna technology appears to be in arrays, especially electronically-steered phased arrays.[EN713] This trend is readily seen in the sophisticated conformal designs developed for aircraft and land vehicles communicating with satellites moving in non-synchronous orbits.[EN714] An array is defined as a group of identical elements activated to produce (or receive) a prescribed overall field pattern.[EN715]
Several comments stressed the importance of developing arrays that can maximize or minimize gain in different directions in real time.[EN716] The following discussion on emerging array technology generally supports the view of the commenters.
Adaptive Antenna Arrays
The topic of adaptive arrays goes back at least 30 years.[EN717] Today, after a long period of development, adaptive arrays are meeting the requirements of the burgeoning mobile-satellite services. An adaptive array is defined as a collection of antenna elements, each connected in a feedback loop to a weighting and summing network. The network algorithms automatically reduce the effect of an unwanted signal or enhance a desired one. These algorithms still occupy much of the research effort in adaptive arrays, but work is being done on hardware as well.[EN718]
Scanning Phased Arrays
A phased array generally has one control element for each array element making the cost relatively high compared to other beamforming techniques.[EN719] The phased array appears to have an advantage over the mechanically steered reflector antenna due to its ability to beamshape electronically. This is generally true if the performance specifications are modest, as in INMARSAT applications.[EN720]
Accurate electronic control at each antenna element is the key to reconfiguring the radiation pattern of an array. The next-generation communication satellites are designed with steerable phased arrays having beamforming networks digitally switched by phase-shifters and attenuators.[EN721] Phase and amplitude controls have reached a level where MMIC chips have switching accuracies better than 5o and 0.5 dB across a 8 percent bandwidth.
The implementation of phase-only control for telecommunications and EMC is significant.[EN722] Since phase is easier to control electronically than amplitude (where power is handled), circuit miniaturization is greatly simplified, notably for multi-element arrays. The ability of one antenna to produce two or more radiation patterns (e.g., cosecant and pencil-beam) on demand and in prescribed directions has obvious application in the mobile and satellite services. For example, an earth station could avoid interference from an orbiting satellite by switching from an omnidirectional to a cardioid pattern when the satellite is in view.
Historically, multiple-beam antennas and scanning phased arrays have been competing approaches to pattern forming. A multiple-beam antenna differs in requiring one port for each beam and a reflector or lens aperture. With good isolation between ports, the multiple-beam antenna (being based on time-proven design) is probably a better choice when the requirement is only for a few independently switchable beams, as with the ACTS space platform and most Earth-based systems. But tens of beams are planned for future satellites, and the added weight and volume of hardware required for a multiple-beam antenna makes the active phased array a very viable alternative.[EN723]
Presently, the most sophisticated designs in beamformers combine a scanning reflector with a phased-array feed in a hybrid arrangement. Broad, instantaneously tunable bandwidths, and good control of amplitude and phase of the aperture illumination, have recently been reported in the literature.[EN724] But there are still problems common to most hybrid antennas, e.g., the dynamic control of aperture amplitude. However, these appear tractable in the near future.
In the 1950's, it was shown that an antenna could be relatively frequency independent (i.e., broadband) if its shape could be specified by smoothly changing physical dimensions.[EN725] This led to the invention of broadband angle-dependent antennas such as the omnidirectional spiral and cone antennas, and the directional multi-wire log-periodic and planar-dipole arrays. These appear satisfactory for the relatively wide bandwidth requirements of spread spectrum and multiple access systems.
In the near future, mechanically-steerable log periodic and dipole arrays may successfully compete with electronically steerable phased arrays for tracking LEO satellites. Low-cost rotatable Yagi arrays have been developed for mobile satellite use.[EN726] But Yagis are inherently narrowband and may not be suitable for the wideband requirements of some proposed systems. Instead, mechanically steerable wideband wire antennas may be a better choice until electronically steerable phased arrays are developed for widespread use.
For a bandwidth ratio of three-to-one or more, conventional antennas, such as discones and spiral cones, can meet an omnidirectional criterion for signal magnitude. For the additional control of phase across the band needed for digital systems, the conical antenna appears to be a more suitable choice.[EN727]
Directive antennas, on the other hand, have the problems of poor compatibility with the feedline, high cross-polarization, and poor control of beamwidth when operated over a wide band. One solution uses microwave components combined in a novel way.[EN728] Since this design is based on conventional microwave technology, extension to 18 GHz and beyond requires close attention to construction techniques and dielectric characteristics.
The self-complementary antenna holds some promise as a directive wideband radiator. Consisting of two complementary parts fed at a common point, this antenna type has a constant input impedance over a wide band and can be ideally designed for any desirable radiation pattern.[EN729] However, its phase characteristics need more study before the antenna can find acceptance as a distortion-free short-pulse emitter.
Only about a dozen comments to our Inquiry explicitly addressed propagation technology.[EN730] Implicit in these replies was the need for radio systems to attain a real-time response to changing propagation conditions. This section is presented from a spectrum management viewpoint with this need in mind.
Propagation technology is directed at producing better models to explain and predict the behavior of radio waves in the natural and manmade environment. Models are often semiempirical because of the difficulty in describing the environment. In recent years, these models have been improved with numerical computation afforded by the availability of fast computers. Propagation models will be discussed in terms of three broad frequency bands: 10 kHz to 30 MHz, 30 MHz to 10 GHz, and above 10 GHz. Propagation is mainly affected by the ground and ionosphere in the first band and by the atmosphere in the last band. From a spectrum management standpoint, the least affected middle band is the most important because most spectrum usage is occurring in the 30 MHz to 10 GHz region and this will probably continue to be the case past the year 2000.
Radio waves propagate primarily along the Earth's surface in the lower part of this band. A refined empirical expression based on ground permittivity is the accepted surface-wave model today and probably into the near future.[EN731]
In the upper parts of the band, the dominating sky-wave propagation between two points on Earth depends on the launch angle, frequency and ionosphere. Since the angle and frequency can be controlled, knowledge of the spatial and temporal distribution of electrons making up the ionosphere is central to predicting sky-wave propagation. The short-term prediction needed by adaptive radio systems is aided by providing better input data to accepted models such as the Ionosperic Communications Analysis and Prediction (IONCAP) program,[EN732] as well as programs specifically designed for shorter-term prediction, such as PROPHET.[EN733]
The direct monitoring of the short-term behavior of ionospheric propagation paths is accomplished by special ionosondings or by channel "probes" embedded in the signal itself.[EN734] The monitoring data become inputs not only to propagation models, but can be used directly by efficient users of HF channels, such as the ALE system.[EN735]
Generally, as the frequency increases from 30 MHz to 10 GHz, the ionosphere appears more transparent to radio waves. The properties of optic waves (such as reflection, refraction, and diffraction) assume importance at these frequencies, and topographic features affect propagation.
Terrestrial Radio Models
The demands of the fixed and mobile services have created a need for better terrestrial propagation models for the VHF and SHF bands. NTIA and others have several models available for use in analyzing point-to-point paths.[EN736] But, models that relate more directly to spectrum management need to be developed.
The past few decades has seen the development of a number of prediction models for VHF-UHF land mobile communication. However, the large number of mobile models currently in contention indicates the difficulties of characterizing a typical mobile channel.[EN737] In recognizing the problem, the Scientific Committee of the International Union of Radio Science for Telecommunication is forming a task force whose objective is to seek such a characterization.[EN738]
For spectrum management, the requirements go past the need to simply characterize a channel. The rising complexity of allocation problems in terrestrial communications demand more intensive processing of a wider range of input data for proper scenario representation than can be handled by the present models. Fixed, mobile, and mobile-satellite systems will have to be considered together in many cases.
The beginnings of the sophisticated processing needed for spectrum management can be seen in a recently developed polarimetric model.[EN739] Here, the probability density function (PDF) and field strength delay spectrum are derived for a multipath signal over a 3-D representation of the terrain, and further processed to yield a 2-D receiver field strength PDF. The model's developers point out that the multi-dimensional approach allows better coordination between mobile and fixed systems in local-to-large area coverage and evaluates the spectrum-efficient use of frequency, modulation, coding, and dual-polarization.
Propagation Into and Within Buildings
The advent of cellular and PCS devices requires detailed information on radio propagation through the external walls and within a building.[EN740] For the simpler problem of penetration through the external walls, particularly from 100 MHz to 1 GHz, several sources of information are available for system planning and spectrum management.[EN741]
By contrast, the problems of path loss and time dispersion of propagation within a building are more complex, but a large body of measurements for narrowband propagation for the upper UHF band is presently available.[EN742] For the near future in the UHF band, the potential advent of wideband modulation for PCS devices requires better propagation data.[EN743] For the more distant future, a review of wireless personal communication indicated that propagation modeling for higher bands may be satisfied by scaling data from the work done so far for the UHF band,[EN744] but scaling alone cannot account for scattering from surfaces that appear rough at higher frequencies. A diffuse-scattering model of rms time delay (which affects high data rate digital signals) is available, but the model would need validation by field measurements.[EN745]
Atmospheric factors that affect propagation at frequencies above 10 GHz were discussed in Chapter 8. Here, the propagation models themselves will be discussed with regard to Earth-space and terrestrial paths.
Most Earth-space propagation models span from about the upper VHF to lower SHF (200 MHz to 20 GHz) reflecting the current needs of the satellite services. At the gigahertz frequencies, the models depend heavily on the data and models describing the effects of atmospheric constituents.
The Earth-space models have been extended into the SHF, but are better substantiated below 30 GHz.[EN746] Propagation experiments of ISDN transmissions at 20/30 GHz between Earth and the ACTS satellite are scheduled this year.[EN747] These are vital in determining the reliability of wideband digital transmission at higher frequencies where atmospheric scintillation and ice cloud depolarization take effect.[EN748] As described in Chapter 8, Earth-based VSAT's are especially susceptible to these effects. The deployment of thousands of highly mobile VSAT's in the near future may cause interference problems of importance to spectrum management.
Fixed point-to-point propagation have been treated extensively in CCIR publications and other sources dealing with system studies and spectrum management.[EN749] These nearly always include information on rain conditions, the most significant factor at gigahertz frequencies.
The trend toward using short-haul common carrier links at higher uncrowded frequencies raises the question as to whether the long-haul microwave models are still applicable at higher frequencies. A study at 19 and 23 GHz addressed this question by determining that models used successfully below 10 GHz are overly optimistic and should be replaced by a model having better rain rate statistics.[EN750]
An atmospheric propagation model taking into account dry-air, water vapor, and suspended-particle conditions at gigahertz frequencies has been developed by NTIA's Institute for Telecommunication Sciences.[EN751] System designers and spectrum managers will find the PC-software version of the model useful in assessing propagation under "clear-air" conditions.
The salient points of this chapter are summarized as follows: