ACP WGF16-WP04
AERONAUTICAL COMMUNICATIONS PANEL (ACP)
SIXTEENTH MEETING OF WORKING GROUP F
Montreal, Canada 11-15 December, 2006
Agenda Item 5: / Review of CPM text for WRC-07Potential Technical Issues with Aviation Use of
RNSS Service Links in the 5 GHzBand
(Presented by Mike Biggs)
Introduction
As part of its work on Agenda item 1.6 of the 2007 World Radiocommunication Conference (WRC07), the International Telecommunication Union Radiocommunication Sector (ITU-R) Working Party 8B (WP 8B) has been considering an allocation to the aeronautical mobile (R) service (AM(R)S) in the 5 010-5 030 MHz band. Toward that end, studies have been completed on the possibility of sharing between AM(R)S and the radionavigation satellite service (RNSS) already allocated to the band (Doc. 8D/413).
At WRC-03 an allocation was made to the 5010-5030 MHz band for RNSS in the space-Earth direction. To date no systems have been implemented using that band, however ITU-R Working Party 8D (WP 8D) has initiated development of a Working Document toward a draft new Recommendation (WDDNR) containing characteristics of RNSS systems planned for that band. That WDDNR includes planned characteristics both for feeder links and for service links.
Existing RNSS systems operate service links in portions of the 1 164-1 215 MHz, 1 215-1 300 MHz and/or 1 559-1 610 MHz bands (collectively termed “L-Band”). The purpose of this paper is to outline some of the potential technical issues with implementing aeronautical service links in the 5GHz band – including impacts on RNSS satellites to transmit, and possible avionics to use, those service links. These preliminary analyses will need to be revisited as AM(R)S and RNSS system and signal characteristics mature.
Potential technical issues with 5 GHz RNSS service links
Spacecraft implications
The first major implication of adding a 5 GHz service link to the RNSS spacecraft is the 13-17 dB increase in power that would be required to overcome propagation and atmospheric effects and to provide signal levels comparable to those provided in lower frequency bands. For radio links with fixed gain on both ends, the received power falls off in proportion to frequency squared in accordance with:
where PR is the transmitted power, GT is the transmit antenna gain, GR is the receive antenna gain, d is the distance between antennas, and is the wavelength. An RNSS service link is such a link: the spacecraft antenna gain is dictated by the requirement for earth coverage whereas the mobile user’s antenna gain is driven by the requirement to receive satellite signals above the horizon at any azimuth angle. Therefore the only way to compensate for the increased free space path loss at 5GHz vice L-Band, is to increase radiated power by:
However in (Doc. 8D/403), under the assumptions provided in their link budget (including a minimum receiver gain of 3 dBi), the total transmitted power needed is less than 16 Watts in the estimate of one RNSS system provider. Alternatively, since antenna physical aperture is related to antenna gain also by the ratio of the square of the wavelengths, i.e.:
,
an antenna implemented at 5 GHz could be designed to provide 10.1 dB more gain than a similar-sized antenna at L-Band. Unfortunately this increased gain would come at the expense of loss of hemispherical coverage. Another suggested approach involves more-complex user antennas that would track the RNSS satellites thus providing the directive gain. Many questions exist with such antennas, including possible impacts on total system availability (i.e., would the antenna need to track several satellites simultaneously), possible impacts on sharing studies (i.e., if the antenna is providing gain toward the desired/satellite signal, what rejection is it providing toward interfering signals), and technology standardization concerns. Further studies on this approach and associated technology should also be accomplished as plans for RNSS system/signals develop.
Atmospheric loss is also larger at 5 GHz than at L-Band. Whereas an RNSS link budget could include 0.6 dB for atmospheric loss for L-Band signals [1], up to 5dB may be encountered for 5GHz links in tropical regions [2]. Figure 1 shows the worst case 5 GHz rain attenuation in a wet weather region at 20o latitude. Note that, although approximately 5 dB attenuation may be observed at 99.99% availability in tropical regions, the typical 5 GHz rain attenuation in the continental region at 40o latitude is 3 dB (see Figures 1 and 3). In contrast, Figures 2 and 4 include the worst-case and typical L-band rain attenuation to validate the existing rain attenuation loss figure.
Figure 1. 5 GHz Rain AttenuationFigure 2. L-Band Rain
in wet weather region (20o Latitude)in wet weather region (20o Latitude)
Figure 3. 5 GHz Rain AttenuationFigure 4. L-Band Rain Attenuation
in continental weather region (40o Latitude)in continental weather region (20o Latitude)
Another potential satellite implication deals with the antenna array required to provide the 5 GHz signals. Current RNSS systems – with service links concentrated in the L-Band – use a single array. Upgrading such array designs to also provide 5GHz signals might involve increasing the number of array elements (to maintain the sub-wavelength spacing) and adding complexity to the phase networks. Alternatively thenew antenna may be deployed separately, adding the complexity of requiring control or estimate of the phase center variation between the L-band and 5 GHz band antennas. This may complicate the use of dual-frequency (i.e., L-Band and 5 GHz) measurements. Addition of a separate 5 GHz band aperture on the satellite may also complicate the inter-frequency group delay bias. In current RNSS satellite designs, coherence between L1 and L2 is facilitated by sharing common radio frequency (RF) paths as much as possible. Introduction of a separate transmitter aperture may introduce more variability in this bias.
Finally, to achieve the same contribution to phase tracking jitter in user equipment specified at L band (0.1 radians in a 10 Hz loop), the short-term stability requirement of the satellite oscillators would need to be significantly increased (see discussion in avionics section).
All of the above potential complications in the RNSS satellite design (vice the current “L-Band service links only” approach), could result in increased cost and schedule risk. These factors should be considered further as plans for RNSS system/signals develop.
Avionics considerations
The following sections address implications of 5 GHz band service links on RNSS user avionics equipment. It is important to note that final conclusions regarding the acceptability of such service links for aviation/avionics cannot be made until the possible RNSS service links and AM(R)S implementation are better defined. Further studies should be performed as more details become available.
Antenna
Microstrip patch antennas (see Figure 5) are often used for mobile – including aeronautical -- RNSS applications. Dual patch antennas consist of two patches tuned to receive two different L-band frequencies. The two patches may be side-by-side, or stacked. The radius of the patch is inversely proportional to the frequency [3]. Typical L1 patches are roughly 50-100 mm in diameter. 5 GHz band patches are possible at roughly one-third this size.
Accommodating both an L band patch and a 5 GHz patch in one airborne antenna (e.g., ARINC 743) should be entirely possible, however one factor that will likely complicate the antenna design is the larger cable loss at 5 GHz (see below). Increased cable losses will result in more stringent requirements for low noise amplifier (LNA) gain at 5 GHz as compared to L-band. For example, two LNAs and a diplexer may be required to meet aviation High Intensity Radiated Field (HIRF) and RF interference requirements. As a result it may be difficult to fit the necessary electronics into existing aeronautical antenna packages or footprints. If introducing a larger antenna is necessary, it could present a number of problems such as complicating aircraft retrofit. If the new antenna is larger, there are implications on ice accumulation and issues associated with increased RF loss due to water, snow and de-icing fluid on the antenna, which could affect reception of the RNSS signal. Given this, and since in general, 5 GHz RF equipment is more expensive than similar equipment operating at L-band, it is expected that an integrated aeronautical 5 GHz/L-Band antenna would be more expensive than an integrated multi-frequency L-Band antenna. However, based on L-Band RNSS experience, these costs may be reduced as production increases.
Operation of service links in the 5 GHz band may facilitate widespread equipage of multipleelement antennas for interference suppression for some RNSS applications. The concept is that since multiple-element antenna size is a function of the signal wavelength (see Figure 6), the 5 GHz band will allow for smaller antennas, and thus promote this technology. However, although 5 GHz band multiple-element antennas will be about one-third the size, and one-ninth the area, of similar multiple-element L-Band antennas (13” diameter for a four element L-band antenna is typical, 5” should be possible for 5 GHz), size and area are not the only factors determining the market for multiple-element antennas. Other important factors are the cost of the associated electronics which may be higher at 5 GHz, the perceived need for interference suppression capabilities, and any applicable domestic export restrictions.
Figure 5
Circular microstrip patch antenna
Figure 6
Common configurations of four-element antennas
Cabling and connectors
Antenna-receiver cable losses are greatly increased in the 5GHz band relative to L band. For example, cable available from one manufacturer has L band attenuation ranging from 3.1 to 45 dB per 100 feet, whereas 5 GHz band attenuation for the same products range from 9 to 115 dB per 100 feet – with lower loss generally coming with different technology and/or higher-cost cable. As a result, the use of 5 GHz service links willlikelynecessitate the use of more expensive cabling to meet receiver sensitivity requirements. Furthermore, many large aircraft currently cabled for LBand RNSS would likely need to be re-cabled to be able to utilize a 5 GHz band signal. This retrofitting and associated recertification could be extremely costly, especially for aircraft that are certified to use L-Band RNSS for precision approach (e.g., using the International Civil Aviation Organization standard Satellite-Based Augmentation System (SBAS) or Ground-Based Augmentation System (GBAS)).
The current Multi-Mode Receiver (MMR) being developed for commercial air transport uses bayonet style connectors. This type of connector is needed for fast simple replacement of a Line Replaceable Unit (LRU). The Airline Electronic Engineering Committee (AEEC) developing MMR requirements has found it difficult to develop a 5 GHz bayonet style connector with acceptable loss. Instead, to achieve acceptable loss requires specially manufactured connectors with high mechanical tolerances. As a result, MMRs delivered with the current bayonet style connectors on the GNSS port that work fine for L-band will not have connectors that would support 5 GHz operation, requiring installation of different connectors on the rack to use such a signal. On some aircraft, the entire avionics rack may need to be removed. In the next several years there are expected to be over 1000 such aircraft configurations in service.
Radio frequency (RF) circuitry
Most RNSS receivermanufacturers employ designs based around Silicon (Si) technology. For 5 GHz links, Gallium Arsenide (GaAs) technology may be needed, resulting in potentially large non-recurring and recurring engineering costs to receiver manufacturers.
Phase noise
Almost all civil aviation RNSS applications require the use of carrier phase tracking to achieve required levels of accuracy (e.g., carrier aiding, smoothing, or kinematic carrier phase positioning) or to demodulate the RNSS navigation message. One of the impediments to reliable carrier phase tracking is the phase noise of both the spacecraft oscillator and the receiver oscillator. The variance of the error in tracking the carrier of the RNSS signal increases proportionally to the square of the signal frequency [6]. Thus, phase noise performance of both the receiver and spacecraft oscillators will need to be improved by 10 dB in order to maintain the same receiver carrier tracking performance at 5 GHz as compared to L band. This performance enhancement may come at some cost, although experience shows that current L-Band oscillators greatly exceed their L-Band phase noise requirements. Alternatively, if the same quality oscillators are used, the loss-of-lock threshold of the receiver in terms of input signal-to-noise ratio will be higher (i.e., receiver will be more susceptible to interference).
Dynamics
Signal dynamics due to vehicle-satellite motion will be higher by a factor equal to the ratio of the frequencies, requiring loop bandwidths that are also higher. This further increases susceptibility to interference (and thermal noise) and could increase software throughput requirements.
Interchannel bias
Integrated receivers incorporating 5 GHz operations would require the use of a completely separate front-end from that used to process L-Band signals. Any group and phase delay biases introduced between L-Band and 5 GHz would complicate the processing of dual-frequency L-Band/5 GHz measurements. Although biases would be present no matter what two frequencies are processed, the bias problem could be more manageable with two L-band signals since most of the front-end circuitry can be shared for both.
Summary
In summary, the following potential technical issues have been raised with regard to the use of the 5010-5030 MHz band for aeronautical service links:
–Spacecraft implications
–The ability to overcome 13 - 17 dB increased propagation (10.1 dB) plus rain/atmospheric (3-7 dB) losses.
–Depending on the implementation, possible increases in complexity of the frequency synthesizer and signal generation.
–Requirement for a multi-band (L-Band/5 GHz), or additional 5 GHz antenna.
–Possible requirement for lower phase-noise oscillators to meet current RNSS requirements.
–User avionics equipment implications
–Increased antenna-to-receiver cable losses may make meeting receiver sensitivity requirements more difficult or costly. Furthermore, aircraft with cabling for L-Band RNSS may require costly replacement of existing cabling and recertification.
–Oscillator phase noise is increased. Either new types of oscillators will be required, or carrier phase tracking performance of user equipment will be degraded.
–RF circuitry may need to be based on GaAs technology, representing a change from current designs.
–Interchannel biases could make the processing of dual-frequency L-Band/5 GHz measurements more challenging than the current processing of dual L-Band frequency measurements.
As a result, further study is required as plans for 5 GHz service links develop to determine if they would be utilized by aviation.
References
[1]Department of Defense, December 8, 1993, Global Positioning System Standard Positioning Service Signal Specification.
[2]Roger L. Freeman, 1994, Reference Manual for Telecommunications Engineering, 2nd Edition, New York. John Wiley & Sons, Inc.
[3]Ohmori, S., H. Wakana, and S. Kawase, 1998, Mobile Satellite Communications, Boston: Artech House Publishers.
[4]Moelker, D.J., Y. Bar-Ness, October 1996, “Adaptive Antenna Arrays for Interference Cancellation in GPS and GLONASS Receivers,” IEEE Position, Location, and Navigation Symposium.
[5]Pasternack Enterprises Catalog #1997A.
[6]Hegarty, C., Spring 1997, “Analytical Derivation of Maximum Tolerable In-Band Interference Levels for Aviation Applications of GNSS,” NAVIGATION: Journal of the Institute of Navigation.
______
Page 1 of 7