7 Limitations of Mobile Satellite Systems

7 Limitations of Mobile Satellite Systems

7.1 Delay and Doppler

The delay and Doppler effects associated with satellite links are due entirely to the mechanical laws governing the satellite orbit. Any system design must take full account of these effects. For example, simple delay has an impact on speech quality that will require echo cancellers to be used at interfaces with the analogue network. Delay also requires allowances to be made in signalling protocols and power level control.
Changes in delay are the result of integrated Doppler shifts on the bit data rate and are significant for all orbits except GSO during a call and particularly during satellite handover. Such changes are likely to require a data buffer to maintain the delay at a constant maximum value. The data buffer can reside in either the LES or the mobile terminal between the two echo control devices and is required for both receive and transmit. Doppler shift itself complicates signal acquisition and spectrum management. The Doppler shift will not be identical for the in-bound and out-bound links due to the different feeder and mobile link microwave frequencies. Furthermore, the shift is in different directions if corrected at the mobile terminal. For LEO and MEO orbits, the shift may need to be individually corrected for each mobile; for HEO, common Doppler compensation can be incorporated in the LES or onboard the satellite.





7.2 Low link margins

Emphasis has already been made on the importance of keeping impairment margins low. An illustration can be based on the calculation of Carrier-to-Noise (C/N) ratios for uplink, downlink, and the total link. Assuming the downlink has C/Nd = 10 dB and is near the performance threshold, the feeder will need a 13 dB margin (C/Nu = 23 dB) to maintain
the degradation to less than 0,2 dB (i.e. C/Nt = 9,8 dB). Operation at levels just above threshold are only feasible for satellite because of the stable propagation path and because most impairments (including the large noise contribution) can be considered to be random. These low margins, compared to the terrestrial environment, result in longer signal acquisition times.

All impairments must be carefully analysed and include: imperfect in-band filtering, group delays, out-of-channel emissions (which demand very tight power amplifier linearity requirements, carrier to interference ratios, etc.) Multi-path also requires especial attention. In the terrestrial environment, multipath propagation normally results in inter-symbol interference that can be compensated with equalisers. The effects of multipath fading itself are often negligible within the main service areas because the detected signal level is sufficiently above threshold. In the satellite path, multipath delays are often short enough to be ignored (except for aircraft and ships) due to the comparatively high elevation angle of the radio path. However multi-path fading, in which a multipath signal partially cancels the main signal, can reduce the final signal below the modem operating threshold. Hand-offs between successive LEO, MEO, or HEO satellites will be more complex because of the small operating margins which makes it difficult to promptly detect signal disappearance. Satellite signal qualities often cannot be assessed from signal level (which is swamped by thermal noise) but are often estimated from the activity within the forward error correction algorithm. This requires time averaging and cannot be an instantaneous measurement. Satellite diversity reception
might alleviate some of these issues. Mobile terminals with high gain (directive) antennas have further problems with signal acquisition as the antenna may need to be mechanically or electronically steered towards the satellite before the signal rises above the detection
threshold.

7.3 Spectrum and orbit matters

Limited spectrum availability will constrain the potential capacity of the satellite component and hence will orientate personal satellite services towards low bit rate voice and data. Spectrum issues are very complex but can be broadly classified into three areas:
- feeder link planning;
- mobile frequency co-ordination;
- mobile frequency re-use and spectrum efficiency.

Global agreements exist for planning GSO systems via the ITU RS (formerly IFRB) for designated frequency bands. Feeder links are normally in one of the established Fixed Satellite Service (FSS) bands and are straightforward except for the large bandwidths required to support peak traffic on each satellite. Mobile frequency co-ordination is not simple however, particularly as their antenna patterns are near omni-directional and any mobile system is likely to require exclusive access to a frequency band. The next problem, that of re-using the frequencies as frequently as possible, is very similar in concept to terrestrial cellular planing except that isolation is provided by satellite antenna beam shaping rather instead of geographical spacing. Feeder links for non-GSO satellites are more complex, particularly because of the lack of established procedures for the
many possible orbits. Furthermore, there is no orbital registration akin to that in the GSO orbit where orbital positions are assigned to particular operators and countries. LEO and MEO may require several widely spaced feeder LESs per satellite sector or inter-satellite links to prevent the feeder link interfering with the geostationary orbit. In either case,
there will be additional delay and Doppler jumps. For HEO orbits where the satellites appear to operate at the same part of the celestial sphere, feeder link planning may not be difficult as GSO-type procedures could be applied. The magnitude of the orbit and spectrum planning problems is partly illustrated by figure 2 which shows an azimuth -
elevation diagram for a fixed land earth station site at a latitude of approximately 50° North (It is not computed from simulated systems but shows only the principle. Therefore slight differences to simulated orbit constellations may exist)


The dotted line, extending from East to West in the shape of an arc, represents the geostationary orbit with two fixed GSO satellites designated 1 and 2. The three LEO tracks belong to one system of approximately polar orbits. The LEO satellites designated 1 and 2 travel North-South. LEO satellite 1 is about to hand over to LEO satellite 2. The LEO satellite 3 travels South-North, but this satellite No. 3 appears here at this track, due to the Earth's revolution, only at a time shift of half a day with respect to the satellites travelling North-South. The slight drift of the three LEO satellites towards West is caused by the Earth's rotation, and hence the rotation of the earth station site, towards East.
At the north-north-western horizon and in the East of the zenith there are two loop-shaped tracks of the HEO satellites designated 1, 2 and 3. The dotted lines extending from the loop near the zenith show the branches of the track where the communication payloads are inactive, as is here the case for the HEO satellite 2. From the diagram one can conclude that a fixed earth station (according to CCIR Recommendation 465 [1]) at this site can communicate with GSO satellite 1 and HEO satellite 1, even when the LEO system is in operation. On the contrary, the links with GSO satellite 2 and HEO satellite 3 could not co-exist with LEO satellite 3, since it passes both the other satellite positions.
Assuming, the LEO satellites' orbit period were not adjusted with the Earth's rotation, then the LEO satellite tracks would scan across the sky like the lines on a television screen, and co-existence with neither GSO nor HEO satellites on the same frequencies would be possible.




7.4 Scope for technical developments

7.4.1 Signal to Noise (S/N) levels
Satellite systems operate very close to theoretical signal to noise demodulation thresholds. There is virtually no scope for reduction in receive thermal noise levels at the satellite or at the mobile terminal as noise levels are dominated by the Earth's background thermal noise (290 K). The noise performance of modern amplifiers is almost
insignificant against this background. The only scope to improve signal to noise margins (for example to provide shadow or in-building operation) is to improve satellite antenna gain.

7.4.2 Hand-held terminal antennas
Present operational mobile satellite systems provide voice services with medium gain steered antennas in the gain range 8 dBi to 15 dBi. Low data rate services can use unsteered lower gain antennas with gains between 0 dBi and 4 dBi.The challenge for UMTS is to provide voice telephony to hand-held terminals using unsteered low gain antennas. The hand-held target imposes practical limitations on the form of antenna and it is unlikely that antennas will have usable gains greater than 0 dBi. However this does not prohibit the use of higher gain antennas for particular applications or circumstances.

7.4.3 Satellites
Satellite technology and commercial launcher capabilities have matured over the past ten years allowing systems planners to design complex systems with confidence. However, reliability is paramount for commercial satellite services and therefore only well established technology, proven in space, is normally considered for major projects.
The satellite antenna is a critical system element. In order to allow operation with low-performance hand-held PES's, the satellite antenna must provide a high gain. This can only be achieved by using advanced array-type antenna technology, including electronic beam forming and beam steering. The resulting spot (cell) diameters on the Earth's surface are typically in the range 1 000 km to 3 000 km.

7.4.4 Digital modulation techniques
The potential capacity of any satellite system is limited essentially by the availability of frequency spectrum and onboard satellite DC power. Hence, for most cost effective operation, it is of paramount importance that power and spectrally efficient transmission schemes are employed. Current research is continuing to make worthwhile progress in
this area.

7.4.5 Voice coding
Lower bit rate voice codecs have been widely used in mobile satellite systems compared to terrestrial systems to reduce power and spectrum requirements. Continuing codec development, coupled with advances in semiconductor integration, is likely to yield improved speech quality and some reduction in overall power/spectrum demands. Target performances for UMTS speech codecs have been set for both terrestrial and satellite components, taking into account the progress that is expected to be made by the time UMTS is introduced.