GSM is the name of the new cellular phone system in Europe. Compared to earlier analouge systems the main difference is that speech is converted to digital signals before transmission. The benefit of digital transmission is that the latest developments in telecommunications can be used. The user gets more services and better sound, and the operator increased channel capacity per base station.
Speech is converted to digital signals by `sampling' (taking samples) frequently. The quality depends on how frequent sampling is done, the so called sampling frequency. More frequent sampling produces a larger amount of digital data. You usually want to limit the amount of digital data, so a compromise sampling frequency is choosen, where the speech sounds good but not perfect. Ordinary digital telephony (ISDN) uses 8 kHz sampling frequency. The analogue wave is coded to a digital value of 8 bits. This is performed 8000 times per second, which gives a resulting 64 kbit/s data flow. The same principle is used in GSM, but the result is compressed to 13 kbit/s. The procedure is actually slightly more complicated, both speech and wave form is coded, and there are more parameters, to give a better sound quality.The procedure of drafting GSM began already in 1982, while many contries did not yet have cellular networks. Gradually it became clear that the system would be digital and to a large extent build on ISDN, and in 1987 realisation of the plans started.
All of the west european countries have agreed on the same GSM standard, and many other countries have adopted it. USA and Japan have however choosen to develop their own systems.
GSM is a step towards personal communication, where you communicate with a certain person, instead of only a telephone number. There are two varieties of GSM, one for 900 MHz and one for 1800 MHz, the latter is called DCS1800. The USA uses 1900 MHz. A SIM card contains all information regarding the subscription for all systems.
GSM should according to the specifications give users at least as good speech connections as existing cellular networks, most services in ISDN, encryption of communications, very efficient use of allocated radio spectrum, an internationally standardised signalling system, and protection of radio system signals.
The most significant features of GSM is digitally coded speech which is transferred in pulses at a 13 kbit/s tranfer rate between mobile phone and radio base station. The transfer method is a combination of frequency multiplexing (frequency sharing) and time multiplexing (time sharing) in 8 time slices. To increase radio efficiency the transmitter power can be controlled, transmission only takes place during speech, and slow frequency hopping is used (the frequency is changed during a connection). The carrier frequencies are separated by 200 kHz, a medium bandwidth.
A cellular phone is a small radio transmitter, which is in connection with a base station. To get reception (coverage) over a larger land surface several base stations are used, and each station is called a 'cell'. The concept is that each cell covers a certain area, and the neighbouring cell covers next area.
The power of radio waves decrease in proportion to the distance from the transmitter. Not only does distance influence reception, houses and trees usually obstruct the line of vision between transmitter and receiver. Neighbouring cells must use different frequencies in order not to disturb each other. On the other hand the number of available frequencies is limited, so one frequency has to be re-used, but sufficiently far from the first base station so not to cause to much interference.
To keep the connection the system must be able to switch the connection from one base station to the next one when the telephone moves. The GSM system has been developed to control the connection with sophisticated methods. Here follows some examples of what takes place during an actual GSM connection.
When the GSM phone is switched on it has to get in touch with a network. The phone first listens to several different frequencies to find some base stations. In the worst case it will have to listen to all frequencies, 124 for GSM, or 374 for DCS1800. Finally the phone has list over base stations.
Now the phone tries to register the user at the most powerful base station in the area of the cell. If the operation was permitted the phone will show ok, contact with a network, or else it will show 'forbidden network', and the user can only dial the emergency number.
The phone is now silent, but in contact with the network. However it is by no means inactive. It constantly listens to the base station, in case there is an incoming call. At the same time it measures the contact quality with neighbouring cells, to be able to switch to a better cell if necessary.
To make a call the telephone needs a channel to the base station. If the cellular phone initiates the call, it sends a channel request on a common channel, without identification. In case of a incoming call a page message is broadcast on the paging channel.
Transmission is never continious, it is made during short time periods (bursts), of 0,577 ms. This corresponds to 217 pulses per second. A channel (connection) never uses all these bursts in sequence, but at the most every eighth. Consequenly a GSM telephone transmits really just an eighth part of connected time. The time interval between succesive bursts is carefully specified, and the phone has to keep the same pace as the base station. When transmitter and receiver keep the same pace it is called synchronous transmission.
There are separate frequencies for transmission and reception. When the phone has received a burst of information it waits three burst-periods to transmit, and the frequency is shifted 45 MHz lower. By keeping transmission and reception in separate time intervals the transmitter part does not interfere with the receiver.
Information about the free channels frequency and time dependency, timing advance, and maximum allowed transmission power is sent to the telephone. Note that a channel is a conceptual channel for information transmission, it is not a continious connection with a single frequency. The connection is sliced into short time periods, and their frequency is shifted up and down.
The telephone confirms the channel allocation, and reports it's class. The class contains information about revision (01 for all telephones in phase one), maximum power, encryption algorithm, frequency capacity, and Short Message capacity. By using this information the system can avoid requesting new functions from old telehones.
Speech is converted to digital signals, and transmitted to the base station at 13 kbit/s. In the future a more efficient speech coding method will cut the data rate in half, and then twice as many channels will be available. The base station 'unpacks' from 13 kbit/s to 64 kbit/s according to PCM-coded ISDN standard, and the speech is still digitally coded. The signals are passed on to the public telephone network and the called part.
Transmission of speech is in groups of digital ones and zeroes. Several groups are coded together in order to allow for error correction methods. If at least five out of eight groups are received the speech can be regenerated.
Reception and transmission is performed in accurately specified time slots, the telephone must keep it's allocated time slot to transmit. Radio waves propagate swiftly, but a certain delay is caused if the phone is a bit from the base station. This delay must be compensated, and the base station specifies how much earlier the telephone has to begin to transmit. The timing advance is computed from measurements of the reply from the phone. It is necessary for the phone to keep the same pace as the base station. The timing advance compensation limits each GSM cell to maximum 35 km radius.
Discontinous transmission (DTX) is applied when the user is not talking, mainly to save battery capacity. The background noise is then coded at 500 bits/second, instead of 13 kbit/s for speech. Another effect of this is that the interference is reduced for other connections, by allowing the phone to be silent for a longer period. The background noise is necessary so the other part will not think that the connection has been cut off. The coded background noise is transmitted twice a second. If one or more background noise transmissions are lost, then the previous is used again. Occasionally this causes amusing effects! Sounds which only occured for a short period when the background noise was coded may be repeated, and the other part can hear clear sounds which really did not exist.
Another metod used to reduce the interference is to reduce the transmitted power. The base station can tell the phone the maximum allowed transmitter power, and can of course also vary it's own effect.
Frequency hopping is another method which contributes to reduced interference. The frequency is changed regularly. Two transmitters may of course interfere, but only during the short time they transmit on the same frequency. Then they change to two different frequencies, and no longer interfere. Yet an advantage of frequency hopping is that reception no longer depends only on one frequency, where conditions may be unsuitable. Reception conditions are spread over several frequencies, and on the average there is an improvement.
During the connection the phone constantly measures the contact with other base stations. Measurements are performed in the pause between transmission and reception. Measurements of up to six stations are reported to the base station at least once a second.
One of the most important functions is 'handover' (handoff). The connection is switched from one base station to the next one, and this is a prerequisite for mobile telephony. Handover can have three reasons; radio contact can become too bad in one cell, a change of cell in order to reduce the transmission power and thus interference of other cells, or to move connections from a congested cell. Handover decisions are based on the measurements which the phone have reported to the base station.
The GSM specifications does not exactly tell when a handover should be performed, the operator can decide. A good indicator of a bad connection is increased error frequency, and also lower signal strength is of interest. Handover errors are easily perceived as a bad connection.
Security is an important aspect in the GSM system. The primary importance is to identify the user uniquely without any possibility of forgery, but also to protect the connection against tapping. Many cellular phone systems have had 'forged users', wher the cost of call has been shifted over to other unsuspecting subscribers.
This is impossible in GSM. The SIM card contains a secret key, which is used with a mathematical operation to change a random number which the system sends over to the telephone. Only the systems operations centre and the SIM card knows the key, and the result from the telephone is compared to the value computed in the systems operations centre. If the values match it must be the correct user.
The same random number is used in another method to calculate an encryption key. Encryption is the process of transposing the contents of a message so an intruder cannot understand the information without knowing the key and the method used. Signalling and speech is transferred in encrypted form, so any intruders who might listen in to the traffic cannot understand the communication.
Even the telephone is checked. Each telephone has a number, a so called IMEI-number, which is controlled against a list of stolen telephones. If a stolen telephone is detected, it will be blocked, even if the subscription is valid. Control of the IMEI-number is not activated yet, and will be introduced in a later phase. The telephone manufacturers believe that the IMEI number in the future will be an integrated part of the telephone chip, so it will be impossible to change IMEI-number just by changing an integrated circuit. The idea is that crooks will know that it is useless to steal GSM telephones. We will have to se how it works in practise.
One of the disadvantages of the digital transmission is interference with other equipment. Because the transmission is split in short pulses a 217 kHz buzzing sound is caused in radios or ordinary telephones. It is not the radio transmission that causes the noise, but the pulsing of the transmission.
Surprisingly few products seem protected against such noise. One of the largest problems is hearing aids, many models pick up the buzz easily. It has however not been shown that GSM phones causes disfunctions in for example cars or other equipment.