Prospective Performance of LEO Satellite Systems

Employing Spread-Spectrum Techniques

Abbas Jamalipour and Akira Ogawa
Department of Information Electronics, Nagoya University, Nagoya, 464-01, Japan
Phone: +81-52-789-2743
Fax: +81-52-789-3173
E-mail: abbas@nuee.nagoya-u.ac.jp

AbstractIn this paper, we discuss on two important issues through the realization of global personal communication
networks with low earth orbital (LEO) satellites. The first issue is the selection of an effective multiple access scheme
which can utilize the limited spectrum allocated to these systems. In this way, we focus on the schemes based on spread
spectrum techniques and explain their unique characteristics. The other issue is partitioning of footprint of satellites
into smaller cells in order to improve the capacity of the system. This latter issue has been considered in satellite
systems of other orbit altitudes such as geostationary orbit and also in terrestrial cellular systems, and recently, in the
proposals of LEO satellite systems it has been explained as an essential requirement. We explain the issue when spread
spectrum is used and discussed on the factors which should be considered in the case of LEO satellite systems in order
to expect similar capacity improvement.

1.

Introduction

Future global personal communications networks

(PCNs) promise to be far more ambitious in terms of
the number and category of user terminals.
Considering the current interests in personal
communications services, it is expected that there will
be an increase in the order of millions of subscribers,
as mobile and personal service may be provided by
satellites. The service to subscribers of the future
satellite-based PCNs will not be limited to
metropolitan areas and will be spread out in all parts
of the world, including developing countries and even
unpopulated areas.
When thinking of the satellite-based personal
communications networks, we cannot ignore the low
earth orbital (LEO) satellite systems as the most
promising candidate, since these systems can provide
global coverage to small hand-held, low-power
terminals without essential need of any existing
terrestrial facilities. The LEO satellite systems, while
having most important features of commonly used
geostationary orbit (GEO) satellite communications
systems, such as wide coverage area, direct radio path,
and flexibility of the network architecture, provide
additional advantages for the global communications
networks, e.g. low propagation delay and low
propagation loss, and high elevation angle at high

latitudes resulting from non-equatorial orbits

constellation. The LEO satellite systems, which may
be considered as an extension of terrestrial cellular
systems, can provide all the necessities for the
tomorrow's personal communications networks [1].
In a global communication network with large
number of users, there is the necessity of a proper
multiple access scheme in cooperation with a
sophisticated protocol which can manage the flow of
information packets on such a manner that their
collisions have the lowest effect on the performance of
the system. In this way, spread-spectrum based
schemes and specially CDMA with their unique
characteristics can be assumed as appropriate ones
among other multiple access schemes. Spread
spectrum signals are effective in mitigating multipath
because their wide bandwidth introduces frequency
diversity [2]. They are also useful in mitigating
interference, again because of their widely spread
bandwidth. The result of these effects is a higher
capacity potential by employing voice activity and
frequency reuse compared to that of non-spread access
methods [3-4]. Moreover, in CDMA, in contrast to
FDMA and TDMA, integration of circuit-mode and
packet-mode traffics requires no special protocol,
make the integrated voice/data system easier to
realize.

In this paper, we discuss on different multiple

access alternatives for LEO satellite systems, and after
that we explain the issue of having spot beams within
the footprint of satellites, which has been recently
considered in the commercial proposals of LEO
satellite systems. We will discuss on the possibility of
capacity improvement when spot beams are
employed. Especially, we will explain that in the case
of LEO satellite systems much more considerations
are involved compared with terrestrial cellular
systems.

2.

Selection of Multiple Access Scheme

The selection of an efficient multiple access scheme is

an important issue in design of any communications
system in which large number of users may
independently access through a common channel with
a limited frequency spectrum. For satellite
communication systems of any types of orbit (GEO,
MEO, or LEO), multiple access techniques are
essential, and specially the problem is much more
serious when they are employed for personal
communications. This might be evidenced by the
following two facts: 1) The users (including users with
either direct access to satellites by portable terminals
or indirect access through gateways from public
telephony networks) of mobile satellite systems are
characterized by huge numbers; and 2) Communications channels in mobile satellite systems suffer by
many imperfections such as thermal noise,
interference, fading, shadowing, and signal deflection.
To accompany with these facts, satellite access
techniques have been evolved from contentionless
protocols to contention (random access) protocols and
CDMA which may be located between the two former
protocols [5].
In this section, we first review the issue of multiple
access scheme and give a classification for it. After
that, we discuss the multiple access for satellite
communication systems.
2.1 Classification of Multiple Access Schemes
Before introducing the promising multiple access
candidate for the future LEO satellite systems, let us
have an overview on the topic of multiple access and
its main alternatives. Multiple access refers to as the
transmission of information data, which might be or
not inserted into packets, by numerous users to or
through a common receiving point at the same time. A
multiple access protocol is a strategy to control such
transmissions so that the probability of collision is
maintained at an enough low level. The issue of

multiple access with the definition given here, may be

included not only the communication systems but also
the computer networks where computer terminals
access to a common processor of a mainframe or a
workstation. In this way, different protocols have been
invented.
As a general categorization to multiple access
protocols, we may divide them into contentionless and
contention protocols [5]. In a contentionless protocol,
the transmission of users are scheduled in either time
or frequency domain. Therefore, sometimes the
contentionless protocols are referred to as scheduling
protocols. By assigning a specific duration of time or
frequency band to each user, it is possible to avoid the
collision between their signals. In this manner, fixed
assignment schemes, such as time division multiple
access (TDMA) and frequency division multiple
access (FDMA), and demand assignment protocols
have been proposed. In a demand assignment
protocol, prior to the transmission, a user should ask
for a channel and after receiving the permission it can
transmit its signal. In the fixed assignment method, a
user is allocated a part of channel capacity, where in
the demand assignment one the scheduling only takes
place whenever the user has something to transmit,
which may improve the spectrum efficiency of the
system.
In a TDMA scenario, each user is apportioned the
entire transmission resource periodically for a fraction
of time; thus, when ten users having equal
requirements, for example, 1 ms in every 10 ms frame
might be assigned to each user. Each user's
transmission is therefore intermittent, a condition
which can only be tolerated by a digital transmitter
which can store its source bits and then burst them out
at a transmission speed at which they are generated.
On the other hand, in a FDMA system, the allocated
frequency spectrum is divided to subbands, and each
user is apportioned a subband at the whole time
domain. Note that in either case there is the necessity
for inserting some part of time or frequency domain
between adjacent channels in order to separate them
from each another. The need for guard time in TDMA
and guard band in FDMA schemes results in waste of
some part of the precious resource. It is important to
note that FDMA can be used with analog transmission
as well as digital, whereas the TDMA can be
employed in digital one.
Different from the contentionless protocols, we
may think the group of contention protocols. Here, a
user may transmit his packet (almost) whenever he has
something to transmit. This type of protocols are
sometimes referred to as random access protocols,
since a user randomly access the channel. Obviously

with this random access a user cannot be sure that his

transmission will not be collided with others;
however, in this protocol the synchronization control
process will be much simpler than the one required,
for example, in a TDMA system. The most note
worthy of such protocols is the Aloha scheme
invented in 1970 [6], in which a user without taking
attention to the transmitting status of other users may
send his packets completely randomly. Several
protocols included in contention protocols group have
been proposed since then, in order to improve
relatively low performance of the Aloha scheme. The
Aloha scheme and its modified version, the slottedAloha (s-Aloha), are included in repeated random
access protocols subgroup, as in these methods it is
necessary to retransmit all packets collided with each
other. Another subgroup for the random access
protocols is the ones with reservation, in which a kind
of reservation is necessary before transmitting the
packets.
CDMA (code division multiple access) scheme
may be considered as either a contentionless or
contention protocol, depends on the situation of the
channel [5]. Simply saying, it is a contentionless
protocol if the number of simultaneous transmissions
on the channel or the level of multiple access
interference is under a given threshold in which all
transmissions can be handled successfully, and it is a
contention protocol if the level of interference is
above the threshold which results in contention and
loss of all simultaneous packets.
In a CDMA system, the signals of users occupy the
whole frequency spectrum at the total period of
transmission, but those signals are distinguished from
one another according to the specific code assigned to
each user. At any given time a subset of the users in
the system may transmit information simultaneously
over the common channel to corresponding receivers.
The transmitted signals in this common spectrum may
be distinguished from one another by superimposing a
different pseudo-random (or pseudo-noise, PN)
pattern, called a code, in each transmitted signal. Thus
a particular receiver can recover the transmitted
information intended for it by knowing the pseudorandom pattern, i.e., the key, used by the
corresponding transmitter. The most popular form of
CDMA is DS/CDMA (direct sequence CDMA), in
which DS spread spectrum signals occupy the same
channel bandwidth provided that each signal has its
own distinct PN sequence. As in this form of CDMA
spread spectrum techniques and their performance
enhancement through the processing gain are used for
multiple accessing purpose, it is sometimes referred to
as spread spectrum multiple access (SSMA).

When several users transmit their packets to a

common receiver in a DS/CDMA system, the
difference in codes of the signals of each user makes it
possible for the receiver to distinguish those signals
from one another, and extracting the individual
information. In a spread spectrum system the
narrowband signal of each user is multiplied by its PN
sequence to form a wideband signal. If we assume that
the information rate at the input to the encoder is R b
bps and that the modulation is BPSK, then the
duration of a rectangular pulse corresponding to the
transmission time of an information bit will be
Tb = 1 / Rb , and then the bandwidth expansion factor
or processing gain of our spread spectrum system will
be

Bc = Wss = Tb
Rb
Tc
where W ss is the bandwidth of the spread signal and
Tc is the reciprocal of W ss which is called a chip
interval. A chip interval is defined as the time duration
of the rectangular pulse (chip) of the spread spectrum
signal. The duration of a single bit in a DS spread
spectrum system then will be equal to B c times of the
duration a chip.
2.2 Multiple Access Schemes in Satellite Systems
After the above brief explanation of multiple access
schemes, now we discuss on this issue for mobile
satellite communications and specially for LEO
satellite systems.
The issue of multiple access first became important
in communication satellites because of the need to
efficiently utilize a precious spectrum resource. The
first comparative study of three main multiple access
schemes; i.e., FDMA, TDMA, and CDMA, can be
seen in a paper published in 1966 [7], based on a
digital communication satellite study performed for
the U.S. Department of Defense.
Considering the very limited applications of digital
satellite communications through isolated-orbit
geostationary satellites in 1966 and their large earth
terminals, that study proposed TDMA as the most
appropriate scheme among others. They mentioned
that for the current applications of digital satellite
communications, the most valuable commodity is the
transmitted power of satellite transponder since this is
proportional to payload of satellite. It is well-known
that in order to utilize the power of transmitters most
efficiently, we must drive them into saturation area,
where the amplifier operates as a nonlinear device. In
this
situation, if the signals of users are
simultaneously being received by the satellite on the
uplinks, the nonlinear amplifier generates undesired

intermodulation products, which both interfere with

the desired signals and rob them of downlink
transmitted power. In TDMA since there is only one
user accessing the satellite transponder at any given
time interval this problem can be avoided. However,
there are a number of disadvantages which offset this
advantage. Maybe the most important one is that the
intermittent nature of the signal transmitted by the
earth stations requires a high peak-to-average power
ratio, proportional to the number of users, which
reduces the efficiency of the earth transmitters.
An acceptable tradeoff between these advantages
and disadvantages would be possible with large
antennas and expensive high-power amplifiers on the
ground. Therefore, the large trunking communication
satellites launched since the 1970's and operated by
the U.S. Department of Defense, INTELSAT, and
others, which interconnected very few large earth
stations, each having very high bit rate requirements,
mostly employed TDMA.
After the arrival of digital communication satellites
networking large number of very small aperture
terminals (VSAT) in the 1980's, however, the
important consideration became the economics of the
very many small earth terminals, while satellite
transponder costs could be amortized over a much
larger terminal population. For these applications,
TDMA was no longer employed, and most VSAT
networks employed FDMA, with the more recent
trends to employ CDMA [8].
Since the arrival of subject of satellites on lower
orbit altitudes for commercial purposes, CDMA has
become one of the strongest candidates for
establishing the multiple access in LEO satellite
systems [9]. At the moment, many of the mobile
satellite systems have been considering to employ
CDMA schemes. While there does not appear to be a
single multiple accessing technique that is superior to
others in all situations, there are characteristics of
spread spectrum waveforms that give CDMA certain
distinct advantages. The two basic problems which the
mobile radio system designer is faced with are
multipath fading of the radio link and interference
from other systems. Spread spectrum signals are
effective in mitigating multipath because their wide
bandwidth introduces frequency diversity. They are
also useful in mitigating interference, again because of
their widely spread bandwidth. The result of these
effects is a higher capacity potential by employing
voice activity and frequency reuse compared to that of
non-spread access methods. Moreover, in CDMA, in
contrast to FDMA and TDMA, integration of circuitmode and packet-mode traffics requires no special
protocol, make the integrated voice/data system easy

to realize; a multimedia feature of the future personal

communication networks. In addition, in a packet
radio environment, CDMA can offer uncoordinated
random-access channel sharing with high throughput
and low delay, along with other well-known
advantages of spread spectrum such as immunity to
external interference and jamming, low probability of
intercept, etc.
Another alternative for the multiple access scheme
in LEO satellite systems is spread Aloha (SA). This
scheme is a combination of two well-known multiple
access schemes, CDMA and Aloha, which is also
called CDMA Aloha. Aloha by now is known as the
simplest random access method that can realize the
share of communications channel to large number of
users. In contrast to conventional (unspread) Aloha in
which the collision of the packets results in loss of the
packets, if we spread the packets with pseudo-random
sequences before transmissions, simultaneous multiple
packet transmission only increases the level of
multiple access interference and does not always mean
the loss of the packets: As far as the level of the
interference is small enough compared to the strength
of the power of the desired packet, the packet is
accessible. Say in other words, in the spread Aloha
scheme we can receive a packet successfully while the
level of the signal-to-interference ratio could be kept
enough above a threshold value, thereby achieving a
more improved throughput performance than unspread
Aloha.
As a subsequent scheme, spread-slotted Aloha
(SSA) [10], which is a combination of CDMA and
slotted Aloha providing packet transmissions at
specified time instants, allows us to have further
efficient systems, similar to the relationship between
the conventional (unspread) slotted and unslotted
Aloha schemes. We believe that for the LEO satellite
systems, CDMA or its combined forms with Aloha,
either slotted or unslotted type, are appropriate
schemes which can be flexible enough with the
coming needs of future personal communication
networks and its multimedia applications.

3.

Multi-Beam LEO Satellite Systems

Recent proposals of LEO satellite systems for global

services have been considered multi-beam antennas on
spacecraft in order to partitioning the footprint of
satellites into small cells. In this section, we discuss
on this issue and its role in improving the capacity of
the system. We first explain the idea of having spot
beams within the footprints and then show the relation
between this issue and the multiple access. Finally, we

discuss on performance of multi-beam configuration

from different viewpoints.
3.1 The Idea of Spot Beams
One of great advantages of the satellite systems in
communications, which has been mentioned for a long
time in the literature, is the wide area that can be
serviced by a single satellite. In the case of
conventional GEO satellite systems, this feature can
provide long distance intercontinental calls as well as
real time broadcasting through (almost) the whole
parts of the globe. Even in the case of LEO satellite
systems, in which the footprint of a single satellite is
much smaller than that of a GEO satellite, it is still
several times to the coverage area of a terrestrial base
station. This means that with a satellite we can provide
communications between two locations on the earth
enough far from each other, even when they are not in
the line of sight of each other due to spherical shape of
the globe or geographical obstacles.
Although the wide coverage feature is very
attractive, there are some objections with it. As an
example, we can say that the satellite signals ignore
political boundaries as well as geographical ones,
which may or may not be a desirable feature. Some
countries do not like to receive special programs
broadcasting via satellite signals (at least for public
receptions), and then the wide coverage of a satellite is
an opposite to this desire. Moreover, transmitting the
signals to all parts of the footprint of a satellite
requires high level of power necessary to be provided
by satellites batteries, which would be wasted in some
area on the earth without any population. In addition,
many countries have their own satellites for national
purposes, which do not require whole coverage in the
footprint of satellites.
The concept of partitioning of the footprint of
satellites to more specific areas has been used for a
long time in GEO satellite systems. In the case of
GEO satellite systems, it is only necessary to cover the
desirable areas for the purpose of a given satellite
system and that enough number of gateway stations to
be included in those areas. This can be done by
arranging the projection of antennas on the satellite
transponders to the desired areas.
In the case of LEO satellite systems besides to
those requirements, since they are basically designed
for personal communications with hand-held terminals
in addition to the communications of satellites with
large gateway stations, it is more preferable to have
very small areas within the footprint of each satellite.
Especially, we can imagine the LEO satellite systems
as an extension of the existing terrestrial systems, in
which cellular-type base stations are replaced by LEO

satellites. In the case of terrestrial cellular systems,

now there are many activities to reduce the size of the
cell into macro-, micro-, and pico-cells, which may
consider in LEO satellite systems as well. In this
manner, many Big-LEO satellite system proposals
consider the partitioning of the footprint of satellites
into small areas, named cell, by very sharp spot beam
antennas on spacecraft. This is the concept of multicell LEO satellite systems.
By using multiple spot beam antennas the footprint
of each satellite is divided into smaller areas, called
cells; a borrowed name from the terrestrial systems.
As mentioned, the basic concept of partitioning the
footprint of a satellite to small cells came from the
same idea using in current terrestrial cellular systems.
Generally, as an advantage, with multiple cells
configuration we can reuse the frequency bands in
sufficiently separated cells or improving the
bandwidth efficiency [11]. Another advantage is that
by diving a wide service area into small cells the RF
(radio frequency) power of users can be reduced due
to the concentration of power into smaller areas,
equivalent to a higher satellite antenna gain. With
concentrating the power into small cells also the
requirements for the variance in power of the user
transmitters (which should be changed according to
the location of user in order to ensure an acceptable
received power level at the satellites) can be more
limited. By multiple spot beam antennas the power is
reduced approximately by the number of cells in each
satellite footprint. The issue of having lower
transmission power is a key issue for realizing a
global personal communication network with handheld terminals.
Similar to terrestrial cellular systems, in LEO
satellite systems there is the necessity of changing the
connection of a given user from one satellite to
another one when the user leaves the service area of
the first satellite and enters to the service area of the
second one. This issue is referred to as hand-off
between satellites. When the footprint of LEO
satellites are divided into smaller cells, another type of
hand-off becomes necessary. In this case, when a user
leaves a cell and enter into another cell within the
same service area a new communication should be
arranged for him and the past connection should be
canceled. Therefore, in a multi-beam LEO satellite
system we have to arrange hand-off mechanism
between the cells within every satellite's footprint in
addition to the hand-off required between the footprint
of satellites, which makes the hand-off procedure
more complex.

3.2 Multiple-Access and Multi-Beam Relation

Some Big-LEO satellite systems like IRIDIUM of
Motorola proposed TDMA as the multiple access in
their system, and many others like GLOBALSTAR
considers CDMA, as the CDMA exhibits attractive
features in cellular systems which may be true for the
LEO satellite systems as well. The selection of
multiple access scheme has a tight relation with the
availability of advantages of partitioning the satellites
footprint to small cells mentioned above, which we
will discuss in the followings. We start the discussion
with a review on the frequency considerations in these
two systems.
In the IRIDIUM system proposal, the available
1616.0-1626.5 MHz band is divided into 31.5 kHz
subbands with guard bands enough to space the
individual carriers 41.67 kHz apart in a FDM
(frequency division multiplexing) structure. In each
31.5 kHz subband, 90 msec TDMA frames are
transmitted using QPSK (quadrature phase shift
keying) modulation at 50 kbps rate to form 4 uplink
and 4 downlink channels.
In the case of GLOBALSTAR; a CDMA-based
LEO satellite system example, the basic structure is
based on the one used in terrestrial CDMA system of
the standard IS-95. The frame structures on uplinks
and downlinks of the GLOBALSTAR system consists
of 1.25 MHz FDM subbands with FDD (frequency
division duplexing) so that the 1.25 MHz subbands
required per traffic channel are paired. The 1600 MHz
band is used for the transmitting each 1.25 MHz
uplink and the 2.4 GHz band is used for the
corresponding 1.25 MHz in the downlinks. In this
system, there are also links between the satellites and
its earth gateways in C-band at the frequency 50915250 MHz and 6875-7055 MHz for uplinks and
downlinks, respectively.
As mentioned above, one advantage of partitioning
the footprint of the satellites to smaller cells is the
availability of frequency reuse in sufficiently
separated cells, in order to improve the bandwidth
efficiency of the system. In the case of TDMA
systems, this can limit the level of interference from
adjacent cells. Typically, a seven-frequency reuse
pattern is used whereby the central spot is surrounded
by 6 spots whose FDM frequencies are different for
each spot in any cluster of seven spots.
For CDMA, it is possible to reuse the same
frequency in spots. This is because of the feature of
CDMA, in which the wideband signal mitigates
multiple access interference by using a correlation
receiver that discriminates between desired and
undesired signals. Therefore, it is often said that for

the CDMA the frequency reuse pattern is one, and

then, the advantage of spot beams in LEO satellite
systems utilizing CDMA is not as strong as in the
systems employing TDMA or FDMA schemes. As an
evidence to this issue, we may look at the parameters
in proposals for the IRIDIUM and the GLOBASTAR
systems. The former one uses satellites at the altitude
of 780 km and the latter one at 1400 km. This means
that the footprint of a GLOBALSTAR satellite is
much larger than that of the IRIDIUM. However, the
GLOBALSTAR considers only 16 spot beams in each
satellite footprint compared to 48 of IRIDIUM. One
reason for this fact may be the less expected
advantage due to spot beams in the case of CDMA
and other spread spectrum based schemes.
The above example exhibits the tight relation
between multiple access scheme and availability of
advantages due to spot beams in LEO satellite
systems. Although, as will be discussed in the
following, the issue of having small cells within
footprints is an unavoidable requirement for the future
LEO satellite-based personal communications
systems, its advantages should not over-estimated by
referring it to terrestrial cellular systems.
3.3 Performance of Spot Beam Antennas
Let us now discuss on the effect of spot beam
antennas on the performance of the system. The idea
of having multiple cells within the footprint of a
satellite in a CDMA based system is to increase the
total capacity of the system by limiting the number of
interfering users to the one of which are inside each
cell. In this manner, the approach of spot beam
antennas can be assumed as an interference
cancellation method, which is widely studied in the
literature. Generally, we can assume two kinds of
frequency allocation in a multi-cell CDMA system.
The first one is to assign same frequency band in all
cells and different sets of orthogonal codes in different
cells. In this method, the available number of
orthogonal codes will be divided by the number of
cells, and hence we have to remain the number of
simultaneous users in each cell within the number of
codes. Obviously with this method the hand-off
process would be simpler, as the users have not to
change their frequency when leaving a cell and
entering an adjacent cell and soft hand-off becomes
possible.
The second possibility is to divide the available
frequency band into a number of subbands with some
guard bands between each pair of adjacent subbands.
This number is determined according to the fact that
how much the cells with same frequency bands should
be separated in order to reuse the frequency band. For

example, if a seven-frequency reuse pattern is used

whereby the central cell is surrounded by 6 cells, we
should divide the total frequency band by seven. With
this method, it is possible to reuse the same code in
adjacent cells and then total number of available codes
can be used in each cell, thereby larger number of
users can access simultaneously to the CDMA
channel. Here, users have to change their frequency
when changing their host cell.
As the first method has wider bandwidth and
smaller number of codes, and the second method has
narrower bandwidth and larger number of codes, there
should be possible to make a tradeoff between these
two methods. Usually a seven-frequency reuse pattern
is used, though sometimes a three-frequency pattern is
also employed. In practical systems it is assumed that
signals of users are perfectly orthogonal within a beam
but signals of different beams are not orthogonal.
Therefore signals of other beams appear as pseudonoise in the frequency reused beams, which their
summation can be assumed as a Gaussian noise.
Frequency reuse in a satellite channel depends
critically on the spacecraft antenna, which is the
primary source of isolation between users with the
same frequency band. In particular, it depends on the
number of spot beams and their size. In this manner,
maybe one spot beam for each user would be the ideal
case, though it is impractical. Actually, a study
presented in [12] has shown that either in a narrowband or in a wide-band satellite system increasing the
number of spot beams more than 37 in a threefrequency reuse pattern only slightly changes the
interference distribution. Therefore, in a real situation
there is an asymptotic gain due to having spot beam
antennas.
In the case of GEO satellite systems employing
spot beam antennas, maybe the simplest configuration
is that we have spot beams pointed permanently in a
given direction. However, a study in [13] has shown
that in certain situations, scanning spot beams can
make optimal use of satellite power. In particular, in
areas where there are low traffic density, dedicated
spot beams would be underutilized, and hence
represent uneconomical satellite design. In such cases,
scanning spot beams could be used to advantage. A
side benefit would be the reduction of interference
because at any time there would be less beams than
service areas. A similar conclusion would be true for
satellite systems in low earth orbits.
It is interesting to consider the availability of
advantages of spot beam antennas, discussed here, in
the case of nonuniform traffic distribution of users in
LEO satellite systems. The reason is that the
advantages of utilizing spot beam antennas in LEO

satellite systems, is very much effective to the traffic

distribution of the system. Consider a satellite system
in which the number of spot beams in a satellite
footprint and available frequency bands are N c and
N f , respectively. Then, in a narrowband system and
uniform traffic assumption the required system
bandwidth will be reduced approximately by the
factor Nc / Nf (an advantage for spot beam
architecture). However, in the case of nonuniform
traffic distribution, we cannot expect the same
advantageous effect. Specially, if we assume the case
in which all traffic channels are concentrated in a
single spot beam, the gain in bandwidth efficiency is
completely lost. On the other hand, in the case of
spread spectrum systems, after dividing the service
area to multiple cells, the interference within the
heavily loaded cell rises by a factor of N c , and then
with respect to the fact that the spread spectrum
systems are interference limited, we should expect a
link degradation of 10log N c . This means that in the
case of nonuniform traffic distribution, we cannot
expect the gain in bandwidth efficiency by
partitioning the service area into cells at all, although
the advantage of the reduction in necessary transmit
power of the satellite and users remains.
From the above discussions we conclude that the
issue of multi-beam antenna for the LEO satellite
system is a serious requirement in the future personal
communication networks in order to increase their
total capacity. However, the implementation of it in
LEO satellite systems requires many considerations,
much more than the ones usually considered in GEO
satellite systems; maybe the most important one is the
traffic distribution of users in the system. An effective
LEO satellite system may be the one which has
intelligent enough on-board processors of satellites
that can manage the system resources; that is the
available bandwidth and codes, in such a manner that
more resources are to be given to the areas with higher
traffic loads without wasting them in areas with low
traffics carelessly. The bounds of the intelligence of
the system would be very wide. A system with the
ability of changing the number of beams per satellite;
a system in which the size of every spot beam can be
changed according to the traffic load of the area it is
projected in; a system with the capability of assigning
different number of codes in different spot beams; a
system which can change the number of frequency
reuse pattern between spot beams; a system that can
assign different frequency bandwidth within the
available spectrum to different cells according to the
requirements of the quality of service in those cells,
would be some examples for high efficiency LEO
satellite systems.

4.

Conclusions

In this paper, we have discussed on two important

issues for LEO satellite systems realizing the future
personal communication networks. These two issues
were the selection of multiple access and partitioning
of satellites footprints into small cells by spot beam
antennas. Regarding the former, we have presented
that spread spectrum based schemes and specially
CDMA are strong enough candidates because of their
unique characteristics resulting from their widely
spread bandwidth.
In the discussion of LEO satellite systems using
spot beam antennas, we concluded that the issue of
multi-beam antennas for the LEO satellite system is a
serious requirement in the future personal
communication networks in order to reuse the limited
frequency spectrum and hence increase their total
capacity. However, the implementation of spot beams
in LEO satellite systems requires additional
considerations than the ones usually considered in
GEO satellite systems because of traffic distribution
of users in the system. Maybe a LEO satellite system
with intelligent beam configuration so that the
interference from other cells can be reduced to its
minimum value, would be the final goal in design of
spread spectrum-based LEO satellite systems.

References
[1]

A. Jamalipour and A. Ogawa, Traffic characteristics

of LEOS-based global personal communications
networks, IEEE Communication Mag., vol. 35, no.
2, 1997, pp. 118-122.

[2]

R. Kohno, R. Median, and L. B. Milstein, Spread

spectrum access methods for wireless
communications, IEEE Communication Mag., vol.
33, no. 1, 1995, pp. 58-67.

[3]

K. S. Gilhousen, et al., Increased capacity using

CDMA for mobile satellite communication, IEEE
Jour. Select. Areas Commun., vol. 8, no. 4, 1990, pp.
503-514.

[4]

K. S. Gilhousen, et al., On the capacity of a cellular

CDMA system, IEEE Trans. Vehic. Technol., vol.
40, no. 2, 1991, pp. 303-312.

[5]

R. Prasad, CDMA for wireless personal communications. Norwood, MA: Artech House, 1996.

[6]

N. Abramson, The ALOHA systemAnother

alternative for computer communications, Proc.
1970 Fall Joint Comp. Conf., 1970, pp. 281-285.

[7]

J. W. Schwartz, J. M. Aein, and J. Kaiser,

Modulation techniques for multiple-access to a hard

limiting repeater, Proceedings IEEE, vol. 54, 1966,

pp. 763-777.
[8]

A. J. Viterbi, A perspective on the evaluation of

multiple access satellite communication, IEEE
Jour. Select. Areas Commun., vol. 10, no. 6, 1992,
pp. 980-983.

[9]

N. Abramson, Fundamentals of packet multiple

access for satellite networks, IEEE Jour. Select.
Areas Commun., vol. 10, no. 2, 1992, pp. 309-316.

[10]

D. Makrakis and K. M. Sundaru Murthy, Spread

slotted ALOHA techniques for mobile and personal
satellite communication systems, IEEE Jour. Select.
Areas Commun., vol. 10, no. 6, 1992, pp. 985-1002.

[11]

R. L. Pickholtz, Communications by means of low

earth orbiting satellites, Modern Radio Science
1 9 9 6, Edited by J. Hamelin, Oxford Science
Publications, U.R.S.I, pp. 133-149.

[12]

F. Vatalaro and F. Ananasso (editors), Mobile and

Personal Satellite Communications 2. London:
Springer, 1996, M. Moher, et al., Interference
statistics for multibeam satellites, pp. 366-384.

[13]

D. O. Reudink, A. S. Acampora, and Y. S. Yeh The

transmission capacity of multibeam communication
satellites, Proc. IEEE, vol. 69, no. 2, 1981, pp. 209225.