An artificial satellite is an object placed into orbit around the Earth for some special purpose (scientific, communication, military).

Most satellites are lifted into orbit by multistage rockets, or space shuttle (NASA). A satellite that is to orbit the Earth is positioned at least at 160 kilometres (100 miles) above the Earth's surface so that atmospheric drag will not slow the satellite. At this height, once it is moving fast enough, the satellite's motion is governed by the same laws that govern the motion of natural satellites and it travels around the planet in a nearly circular orbit.

The time required for a satellite to complete one revolution around the Earth depends on its altitude. An orbit is said to be synchronous, because it is synchronised with the Earth's rotation. So a satellite located 35,900 kilometres (22,300 miles) above the Earth's surface, for example, takes exactly 24 hours or one Earth day to circle the planet. If the satellite moves in the same direction as the Earth's rotation and its orbit lies over the equator, it will appear from the Earth to be positioned at a fixed point in the sky. The satellite is said to be in geostationary orbit. Geostationary orbits are often used for communications satellites. A geo-synchronous (from geo = Earth + synchronous = moving at the same rate) orbit has an orbital period that is an integer multiple or sub-multiple of the Earth's rotation rate, resulting in a repeating groundtrack. The satellite track in orbit, traced on the surface of the Earth is termed the satellite groundtrack. The rotation sense of the satellite in the orbit is referenced with respect to the rotation of the Earth. Satellites with inclination angles up to 90 degrees are termed prograde, and those with larger inclination angles retrograde. Launchers make use of the rotational speed of the Earth when launching into prograde orbits, and for retrograde missions more fuel must be carried to counter this initial velocity. An orbit with a low inclination, or angle between the orbital plane and equatorial plane. This type of orbit is ideal for frequenct coverage of the equatorial regions, but for Low Earth Orbit equatorial orbits the satellite coverage circle is limited, and higher latitudes can not be covered. A polar orbit is inclined at about 90 degrees to the equatorial plane, covering both poles. The orbit is inertially fixed in space, and the Earth rotates underneath. For a Low Earth Orbit, the entire globe is covered. Consequently this orbit is ideal for communications and mapping missions.

If a satellite is placed in a circular prograde equatorial orbit, and the height is chosen such that the orbital period is equal to that of the Earth, then the satellite will appear fixed above the Earth's surface. The required orbital height is about 36786km, which gives a footprint covering almost 1/3 of the Earth's surface. The area on Earth that the satellite can "see" is called the satellite's "footprint." In practice a Geostationary orbit has small non-zero inclination and eccentricity, causing the satellite to trace out a small figure of eight in the sky.

The geostationary orbit belongs to the family of geosynchronous orbits. Satellites in geosynchronous orbits describe a single or double loop around a point on the equator once every 24 hours. Polar orbits follow a path perpendicular to the Earth's equator, in the plane that passes through both the north and south poles. In one day, the Earth will rotate once under a polar orbital path. This makes most polar orbits useful for satellites designed to study the entire surface of the Earth or the planet's weather.

Throughout the life of the satellite, the satellite's orbit and attitude (the direction it points) must be adjusted. Generally, satellites have on-board rockets for this purpose. The rocket may be fired to accelerate the satellite and move it to a higher orbit or to decelerate it and move it lower. Firing the rocket to the side changes the direction in which the satellite points.

The power to run satellites usually comes from solar cells. When the satellite is in the Earth's shadow, auxiliary batteries provide power. Some experimental satellites are powered by thermoelectric generators that change heat from radioactive material into electricity.

The information gathered by a satellite is often stored in the form of electronic signals that are sent by radio to ground stations. Many satellites have on-board computers that not only receive, store, and transmit information, but also control the satellite's operation and orbit.

Satellites are put to a wide variety of uses. Applications range from scientific research to military reconnaissance, but we will focus on communication now.

The most important technological application of artificial satellites has been to provide long-distance communication links, to relay radio signals. Telephone companies, cable television stations, newspapers and magazines use communications satellites to transmit data to various parts of the globe. A group of satellites used in communication among Earth stations forms a satellite communication system. Such systems may provide international communication, as does Intelsat, which includes some 400 Earth stations located in 150 countries, or they provide domestic communication only, as does France's Télécom-1 A, B and C system. By the late 1980s satellites had a greatly increased capacity to handle telecommunications signals, and integrated services digital networks were being developed to create a global voice, data, text, and video system. Today, there are more than 100 communications satellites orbiting Earth.

Satellite links:

Two terms are used in satellite communication to describe the direction of transmission. Downlink is transmission from satellite to earth station and the uplink in the opposite direction. In the uplink there is no constraint about the power to send an information because the transmitter is ground based. In the downlink, the information being send by the satellite, a low power source must be used to transmit information (and a highly sensitive receiver in the earth station). In some application, like telephony the high sensitivity constraint of the earth station is not limited by the cost factor (due to a large number of users), unlike others application like television broadcasting where every single user has to use its own receiver. There receiver's cost should be reduced.

The preoccupation in satellite link design is the power. The problem is to ensure that the signal level is large enough to produce a satisfactory output for a large part of the year. The distance between satellite and earth station being constant (geostationary satellite), the attenuation due to the distance is also known, and cannot be changed. Improvement must then occur at either the transmitter or the receiver.

Error coding

Looking to the future it seems that all satellite systems will adopt digital transmission methods because of their efficiency and flexibility. Nearly all digital transmissions via communication satellites are modulated onto carriers using phase shift keying (PSK) or methods which are closely related to PSK. FDMA can be used with digital transmissions just as with analogue transmissions. However, the ease with which digital signals can be stored and manipulated makes it practicable for a number of earth stations to use the same frequency by interpolating their transmission in time; this is time-division multiple access (TDMA).

A communication system is used for transferring information. The communication channel is typically imperfect and may distort information, causing the received data to be an erroneous version of the information which was to have been transmitted. The errors in the received information are undesirable and so error-control techniques must be adopted to transfer information accurately.

The two broad types of error-control techniques are forward-error-correction (FEC) and automatic-repeat-request (ARQ). In FEC, k information symbols are mapped to a larger number of symbols, n, which are sent through the channel.

The mapping |the FEC code| is devised so that there is some redundancy in the symbols and so, upon receipt, the redundancy can be exploited to detect and correct errors which may have developed in transport. Thus error control is achieved by sending more symbols than would have been sent if no error control were needed. In general, codes with lower code rates (n=k) or complex codes can protect against more errors. Hence, higher fidelity transmission can be achieved.

Both forward-error-correction (FEC) and automatic-repeat-request (ARQ) at the expense of sending more symbols in the channel and/or complexity. In ARQ-based error control, the information is segmented into packets, to which are attached check symbols calculated from the packet in a manner similar to the mapping in FEC. These symbols are used [generally] only to detect errors, and so n is usually only slightly greater than k. If an error is detected in a received packet, the receiver sends to the transmitter a request for the packet to be retransmitted: a negative acknowledgement. If no error is detected, a positive acknowledgement of the received packet is sent instead. The code used for error detection is usually a cyclic redundancy check (CRC), and such codes provide excellent error detection capabilities while admitting simple implementations.

In some cases, FEC may be preferred for error control, while ARQ may be better in others. If the channel is very noisy, information sent through it will be severely corrupted. If ARQ is used in this case, many retransmissions will be required. A retransmission sent through the same channel as the original transmission will also likely be corrupted. Hence, if channel conditions are poor, information delivery may be slow or non-existent if ARQ error control is employed. A FEC code which can correct many errors would be a better error control choice in this case. Such a code would have many check symbols although such overhead may be tolerable if high-fidelity communication is required. the channel symbol rate is fixed and errors are unlikely. Of course, if no channel is available for feedback from the receiver to the transmitter, ARQ is not a feasible error control option, and FEC must be employed.

It is clear, then, that selecting an error-control scheme entails considering the interrelationships of several factors which include required fidelity of information delivered, amount of acceptable overhead, error characteristics of the channel, implementation cost and the availability of a feedback channel.

If the channel is fairly noise-free, ARQ may be a better choice. Usually, less overhead is required for to provide excellent error detection capability than to achieve a comparable degree of error correction. Hence, less overhead is required for high-fidelity communication if ARQ is employed than if FEC is used. With less overhead, information can be delivered more quickly by ARQ than FEC if error control schemes are used for assuring the accuracy of information transferred through imperfect channels. In satellite systems, in which propagation times are typically large, ARQ error control can result in poor throughput to the destination. Also, an ARQ protocol for satellite multicast communication must be carefully crafted to assure good throughput to all destinations regardless of which stations require retransmissions.

Supplementing a satellite link with a parallel terrestrial link may allow mitigating some problems of using ARQ in satellite communication systems. ARQ acknowledgements, and possibly retransmissions as well, can be sent terrestrially in such a hybrid network, and so avoid the large satellite propagation delay.

The satellite transmission capability of a receiving station which communicates with the transmitter exclusively by terrestrial means can be eliminated and the system cost correspondingly reduced. Further, multicasting with a hybrid network may allow retransmissions to be conducted without interrupting the flow of new information to all destinations, so throughput need not drastically suffer if retransmissions are required. The degree to which throughput can be improved by adopting a hybrid network is not clear. A hybrid network's effect on the fidelity of information delivered to the destination(s) is also not clear. An experiment is presented for investigating such error control issues of hybrid networking.

Sources:

Telecommunications Engineering third edition J. Dunlop and D.G. Smith

1995 Encyclopædia Universalis France

Grolhier 1995

Encarta 1997

An introduction to satellite communication by D.I. Dalgleish

Center for satellite and hybrid communication networks (sponsored by Nasa)

http://www.isr.umd.edu/chscn/

Master's thesis : Error control for satellite and hybrid communication by D.E. Friedman

http://www.cs.ucl.ac.uk/staff/S.Bhatti/D51-notes/node33.html

http://www.TBS-satellite.com/tse/

http://www.thetech.org/hyper/satellite/

http://www.ee.surrey.ac.uk/SSC/SSHP/sshp_intro_comms.html