Charles W. Bostian
William T. Brandon
Vincent W.S. Chan
Raymond D. Jennings
Joseph N. Pelton
Propagation is the study of how radio waves travel from one point to another. Its most important practical results for telecommunications are predictions of the transmission impairment characteristics (loss, fading, interference, dispersion, distortion, etc.) of radio links. These strongly influence the choice of transmitting and receiving antennas, transmitter powers, and modulation techniques. The atmospheric propagation characteristics of radio signals at a given frequency must be understood before communications links using that frequency can be designed. Once understood, further research is necessary only if a new service is to be introduced that makes unusual demands on the radio channel.
Propagation research has typically preceded commercial exploitation by many years. Ka-band propagation studies in the U.S. provide good examples. Terrestrial experiments from the late 1940s through the early 1970s established the potential viability of this band for satellite communications. The COMSTAR family of spacecraft, launched in the late 1970s, carried 19 and 28 GHz propagation beacons and helped develop a good understanding of Ka-band propagation in the U.S. The ACTS program, nearly 20 years later, will provide additional propagation data needed to implement and test active techniques for overcoming rain fades. But to date no U.S. commercial satellite using Ka-band transponders has been planned or built.
Europe has a large and well regarded propagation research community, most of whom work for universities, posts, telephone and telegraph entities (PTTs), and government laboratories. Research in satellite path propagation is coordinated and funded by the European Space Agency in a highly effective program that has been operating since the early 1970s. Table 4.1 summarizes ESA's work since 1977 and its planned activities through 1994. Strong national programs exist in Italy, Germany, France, and the United Kingdom. European propagation research is also conducted through COST projects; COST 205, for example, was a European joint research program titled "Influence of the Atmosphere on Earth-Satellite Radio Propagation at Frequencies Above 10 GHz" (Fedi 1985).
European researchers currently have access to sophisticated propagation beacon packages carried by OLYMPUS and ITALSAT F1. The ESA satellite OLYMPUS provides coherent signals at 12.5, 20, and 30 GHz; the 20 GHz signal is switches between horizontal and vertical polarization. The Italian satellite ITALSAT F1, built by Alenia Spazio, carries a 20 GHz telemetry beacon and sophisticated propagation beacons at 40 and 50 GHz. The 50 GHz beacon is polarization-switched. For detailed descriptions see (Revista Tecnica Selenia 1989) for OLYMPUS and (Revista Tecnica Selenia 1990) for ITALSAT.
Japanese researchers have pioneered much of our understanding of satellite path propagation at frequencies above 10 GHz, particularly through the work of T. Oguchi at the Ministry of Posts and Telecommunication's Communications Research Laboratory (CRL). While the panel is unaware of a central body in Japan with a coordinating role comparable to ESA's, Japanese research has been comprehensive and thorough. (For a summary of Japanese propagation activities and an excellent bibliography, see Kagoshima and Shiokawa 1992.)
Early Japanese use of Ku-band, and the availability of beacons and transponders on the CS series and BSE spacecraft, provided opportunities for rain attenuation and depolarization measurements and uplink power control experiments. Recent work has focused on L-band satellite-mobile propagation, particularly with ETS-V. ETS-VI and COMETS will provide opportunities for millimeter wave (43 to 47 GHz) propagation experiments that are unavailable to researchers in the U.S.
Propagation Research at ESA
The Clear Atmosphere. At 4 GHz and above, satellite path signals in the clear atmosphere experience rapid, usually shallow fades called "scintillations." These are caused by atmospheric turbulence and are the radio equivalent of the twinkling of stars. Scintillations are stronger on low-elevation-angle paths. The clear cloudless atmosphere also absorbs radio signals, but below 30 GHz absorption losses are small (less than 1 to 2 dB) except around the 22 GHz resonance frequencies of water vapor, and they vary appreciably with water vapor content.
Rain. The propagation characteristics of the atmosphere that most strongly influence the fixed satellite service (FSS) systems are associated with rain. Rain on a satellite radio path causes fading, or "rain attenuation." The attenuation in dB is roughly proportional to the square of the operating frequency, and rain that will cause a 2 dB fade on a 6 GHz uplink will wipe out a 30 GHz channel with about 50 dB of attenuation.
Rain also depolarizes satellite signals, converting energy from one polarization to another, and causes interference between channels that depend on orthogonal polarization for frequency re-use. At C- and K-band rain depolarization must be taken into account in dual-polarized systems.
The severity of rain attenuation and depolarization depends on how hard it is raining (described by the "rain rate" in millimeters of accumulation per hour), not on the total rain accumulation. Thus, areas subject to intense thunderstorms (like the Gulf Coast of the U.S.A.) experience more severe propagation problems from rain than do areas with a high average rainfall but few thunderstorms (like the Pacific Northwest of North America).
Ice Depolarization. Satellite path signals are depolarized, but not attenuated, by clouds of ice crystals in the upper atmosphere and by the ice particles which are present at the top of rain clouds. The former is particularly difficult to predict and to model because it is not accompanied by anything analogous to rain rate that can be measured on the ground.
Line-of-sight satellite path propagation to or from a mobile station is no different from propagation for fixed sites. Thus propagation data collected at, say, Ku-band fixed sites has some application to Ku-band mobile system design as well.
But the radio paths joining terrestrial mobile stations with satellites are frequently blocked or obscured by terrain, trees, or buildings. This causes fading and, frequently, multipath propagation (usually just called "multipath") in which signals reach the receiver by several independent paths. Additional multipath can occur when a vehicle is on or near a smooth, flat surface which is a good reflector of radio signals. Thus the receiver must acquire the wanted signal in the presence of several interfering time-shifted replicas of itself.
Research in satellite mobile propagation has involved extensive measurement campaigns in the United States, Europe, and Japan. Statistical models have been developed which provide reasonably good guidelines for system design but not necessarily much useful information about the exact characteristics of the signal that will be received at a particular location. Work now in progress, aimed at more site-specific propagation models, will import data from maps and aerial photographs and make more accurate predictions based on the actual environment.
Rain propagation research for satellite applications until recently focused on developing techniques for predicting fading which might occur 0.01 to 0.001 percent of the time, since these values corresponded to the reliabilities that a telephone system had to deliver. Predicting the statistics of rain attenuation from local rain rate statistics is fairly well understood. The rain rate statistics are themselves reasonably well characterized for most of Europe, North America, and Japan, less so for the tropical/equatorial regions. Small-scale terrain effects, localized climates, and the normal year-to-year fluctuations in rainfall all combine to introduce some irreducible variability in the phenomena.
More recent research efforts have focused on determining attenuation statistics at higher percentages of time, say 10 percent to 0.1 percent, since these correspond to the reliabilities and the 1 to 3 dB fade margins that very small aperture terminal (VSAT) systems and other low margin services provide.
While it is useful to know what kind of a fade margin a satellite link must provide to achieve a stated reliability, for most Ka-band terminals of any size and for some Ku-band terminals, the margin may be too large to implement in practice. In this case, systems may be designed to overcome rain fading by using site diversity, UPC, variable rate encoding, or combinations of these techniques.
A site diversity system uses two or more earth stations in a redundancy configuration on the usually valid assumption that attenuation will not be as great at two stations simultaneously as it is at either one of them. Diversity systems have not generally been found to be cost effective.
In systems using UPC and/or variable rate encoding, the quality of the received signal is monitored and, during a fade, the transmitting station either increases power to compensate for the fade or changes the encoding rate to maintain an acceptable bit error rate on the attenuated signal. For satellite systems, this means that the satellite must have reserve capacity (power or bandwidth or both) which can be dynamically allocated to those links which are experiencing fades, and that the earth stations must be able to sense fades and either compensate for them themselves or request the satellite or the earth station at the other end of the link to do so. The round-trip delays involved in geostationary orbit satellite paths complicate this last part of the process.
Ka-Band (And Above) Fixed-Satellite Services. ESA's Ka-band propagation research related to FSS supports a broad range of applications ranging from "microterminals" for personal communications to High Definition TV (HDTV) distribution, to data relay systems for the European Data Relay Satellite (DRS). It includes projects involving scintillation, depolarization, fade duration, and interfade intervals; typically these involve measurements (with the OLYMPUS beacons), model development, and theory. Both rain effects and cloud attenuation (important for VSATs and other low-margin systems) are included.
Longer-term and more science-based ESA projects include a three-year study incorporating simultaneous measurements by OLYMPUS beacon receivers, radiometers, and dual-polarized radar and focus on basic questions of atmospheric structure, characteristics of the melting layer, and spatial inhomogeneity of precipitation. This work will contribute to a planned ESA global radio meteorology relational database that will include a library of propagation prediction methods.
By the late 1970s Deutsche Telecom Laboratories (DTL) had good single-site statistics down to 0.1 percent of the time. Rain fades were not as big a problem as had been anticipated. OLYMPUS measurements confirm these trends. Attenuation at DTL is not as bad as predicted by the common models. Ice depolarization may turn out to be the critical problem. Frequency dependence of ice depolarization is as expected, but ice depolarization is not linked to rain depolarization, i.e., ice depolarization at Ka-band is severe and not predictable.
From a propagation standpoint, DTL is confident that Ka-band is a good band to use. DTL feels that the real problem is the cost of the radio frequency components; it is too high to make Ka-band operation commercially viable.
Alenia Spazio is building the terminals for the ITALSAT F1 propagation experiment to support attenuation and depolarization measurements at 20, 40, and 50 GHz, as well as measurement of phase and amplitude dispersion. Since the 50 GHz beacon switches between two orthogonal polarizations, data collected with it will fully characterize depolarization at this frequency. The terminals were not yet operational, but they should provide a wealth of propagation data in coming years.
Routine Ka-band propagation measurements for FSS were not of great interest at the Japanese sites visited, probably because satellite path propagation in the Japanese climate at these frequencies is already well characterized. Japanese propagation researchers will benefit from O-band (38 to 43 GHz) transponders on ETS-VI and similar equipment on COMETS. These will provide opportunities to collect propagation data concurrently with link performance measurements.
Fade Mitigation Techniques for Ka-Band and Ku-Band. ESA researchers feel that site diversity is the only viable alternative for systems which must cope with fades exceeding 10 dB. Large-scale diversity (diversity with widely separated sites) might be employed effectively with an on-board processing system that would shift traffic from an attenuated downlink to a nearby station that was not experiencing a fade. Small-scale diversity (diversity with sites that are only a few kilometers apart) could provide a few dB fade restoration for terminals that were connected by a terrestrial metropolitan area network (MAN). Traffic for an earth station experiencing heavy rain could be routed to a nearby station experiencing less rain for delivery by the MAN.
DTL reports that the optimum site diversity separation for Germany is 15 km. They are looking at diversity-like operation with baselines on the order of 500 km and having an uplink in eastern Germany and a downlink in western Germany (or vice versa). The goal is to establish the statistics of simultaneous uplink and downlink rain fades on satellite circuits that begin and terminate in Germany.
DTL is performing satellite switched time division multiple access (SS-TDMA) Ku-band experiments with OLYMPUS and plans to test adaptive fade countermeasures using digital TDMA service. They plan to try both UPC and coding rate changes to combat rain fades and to determine which or what combination is best, starting in 1993.
Current Japanese efforts focus on satellite broadcasting and the problem of adjusting the power on a one-way link to overcome downlink fades. Thus Space Communications Research Corporation (SCR) is using a "power upon demand" technique for broadcast applications in beams whose coverage areas are characterized by high rainfall rates. A terrestrial network of some 3,000 meteorological stations will provide data that will determine the power levels for each beam. NHK is working on methods for implementing power control in spacecraft hardware. These include varying the transponder power output by changing the traveling wave tube cathode current and by paralleling TWTs (see Figure 4.1).
Figure 4.1. System for Controlling Beam Power to Compensate for Rain Fading (Courtesy SCR)
Propagation Research for Mobile Satellite Services. ESA has been involved in MSS propagation research since the PRODAT project measurements of the mid 1980s using airborne L-band and S-band transmitters to simulate satellites. Current work includes construction of a K-band mobile propagation terminal for 20 GHz satellite beacon measurements. ESA also has theoretical work in progress on MSS propagation modeling. A long-term goal is the development of techniques for predicting signal strength in urban areas from maps and incorporating the electromagnetic characteristics of buildings.
DTL is not now actively working on mobile propagation but is planning a mobile experiment at Ka-band. They have no interest in VHF or UHF measurements for LEOSATs because no significant spectrum is available in Europe. They feel that satellite mobile services will have to employ higher frequencies -- hence the interest in Ka-band.
The ETS-V program provided Japanese researchers opportunities to characterize L-band MSS propagation in Japan. This was not an active research area at the sites. Antennas previously used in propagation experiments were displayed, but interest had shifted from measurements to techniques for compensating or eliminating propagation impairments at intermediate frequency or RF.
Looking to future hand-held terminal development, Matsushita has just completed a series of propagation measurements at L- and C-bands using ETS-V. Outdoor results were characterized as "better than we expected." Indoor reception was impossible. These measurements were conducted with a variety of modulation types, including Gaussian minimum shift keying (GMSK) and single side band (SSB).
In contrast to the U.S., which flew no NASA propagation packages between CTS (launched in 1976) and ACTS, Europe has had a continuing and well-planned series of propagation experiments supported by ESA's OTS and OLYMPUS spacecraft and by national satellites like Italy's SIRIO and ITALSAT F1. Japan has had the CS and ETS series and BSE. Thus there exists a wealth of Ku- and Ka-band European and Japanese propagation data, and more are being collected. And the European and Japanese experiments have been of higher quality and involved many more sites in more diverse geographic locations than has recent or planned work in the U.S. For example, while many European researchers were making sophisticated measurements with OTS, only one or two sites in the U.S. were able to make Ku- or Ka-band measurements at all, and these used the SIRIO or OLYMPUS propagation beacons at the edge of coverage or the tracking beacons of INTELSAT spacecraft. The ACTS propagation experiment is comparable with what the Europeans have done with OLYMPUS, but the OLYMPUS experiment will have yielded several years of data and be over before ACTS is launched. And OLYMPUS carries a much more sophisticated propagation package (coherent beacons at 12, 20, and 30 GHz with polarization switching on the 20 GHz unit) than ACTS (incoherent, fixed-polarization beacons at 20 and 30 GHz with poor polarization purity).
Whether or not the small U.S. effort poses a problem depends on how extensively Ka-band will be used and on how much margin will be allocated to rain fades, scintillations, and other propagation impairments. Propagation research tends to be a rather academic discipline, conducted primarily by universities, government laboratories, and telephone companies or PTTs, and the results are disseminated widely in the open literature and through the Consultative Committees on International Radio (CCIR). Thus the results of the European and Japanese experiments are well known in the U.S. and available to U.S. system designers. Since local climate has an important effect on Ka-band propagation, there is some inherent uncertainty and potential for error in designing U.S. links from overseas propagation data, but an analysis of the old COMSTAR measurements and the new ACTS propagation data should reveal most of the discrepancies.
With the ITALSAT F1 40/50 GHz propagation package and the ETS-VI O-band experiment, European and Japanese investigators will be able to measure propagation effects at frequencies which have not been studied on earth-space paths in the U.S. or have not been reported in the open literature. The implications for commercial satellite communications technology are unclear, although the 40 GHz band has military applications. Certainly Japan and Europe will have some advantages in exploiting these frequencies.