Detailed description of GSM

Cellular radio systems

PART 1 (47.1 – 47.3.5)

Introduction

Cellular radio systems are by far the most common of all public mobile telephone networks, the earlier (pre-cellular) networks now all being in decline. The basic principles of cellular systems were established by Bell Laboratories in 1949, but it was not until the early 1980s that technology allowed real commercial networks to be built and service offered to the public.

Systems were developed at different times in different countries and subject to a variety ofdifferent constraints such as frequency band, channel spacing etc. As a result, a number of different and incompatible cellular standards are in use throughout the world, and the more important standards are summarised later in this chapter.

Work is already well in hand to specify and develop second generation cellular systems, forwhich the opportunity is being taken to develop common standards and systems across several countries. One such notable system, GSM, has been developed in Europe, and is described in some detail later in this chapter.

Principles of operation

Network configuration

In a cellular radio system, the area to be covered is divided up into a number of small areas called cells, with one radio base station (BS) positioned to give radio coverage of each cell. Each base station is connected by a fixed link to a mobile services switching centre (MSC), which is generally a digital telephone exchange with special software to handle the mobility aspects of its users. Most cellular networks consist of a number of MSCs each with their own BSs, and interconnected by means of fixed links. The MSCs interconnect to the public switched telephone network (PSTN) for both outgoing calls to, and incoming calls from fixed telephones. Figure 47.1 shows a typical network arrangement.

Figure 47.1Cellular network configuration

A cellular network will be allocated a number of radio frequen­cies, or channels, for use across its coverage area, this number being dependent upon the amount of spectrum made available by the licensing authority and the channel spacing ofthe technical standard used by the network. The radio channels are grouped together into a number ofchannel sets, and these sets are then allocated to the cells, one set per cell, on a regular basis across the whole coverage area. Each channel will therefore be re-used many times by the network. The method ofradio planning and allocation ofchannels to the cells is described later in this chapter.

Signalling

Generally, one radio channel is set aside in each cell to carry signalling information between the network and mobile stations. In the land to mobile (L-M) direction, overhead information about the operating parameters of the network, including an area identifier code, is broadcast to all mobiles located in the cell's coverage area. In addition specific commands are transmitted to individual mobiles in order to control call setup and mobiles' location updating.

In the mobile to land (M-L) direction, the signalling channel is used by the mobiles to carry location updating information, mobile originated call setup requests, and responses to land originated call setup requests.

Location registration

When a mobile is not engaged in a call, it tunes to the signalling channel of the cell in which it is located and monitors the L-M signalling information. As the mobile moves around the network, from time to time it will need to retune to the signalling channel of another cell when the signal from the current cell falls below an acceptable threshold.

When the mobile retunes in this way, it reads the overhead infor­mation broadcast by the new cell and updates the operating par­ameters as necessary. It also checks the location information being broadcast by the new cell and, if this differs from the previous cell, the mobile automatically informs the network of its new location by means of an interchange on the signalling channel (Figure 47.2). By means of this location registration procedure, the network is able to keep updated a database of the location area of all mobiles. This information is used in the call setup procedure for land to mobile calls.

Figure 47.2Mobile location registration

Call set up

The signalling procedures for mobile to land (M-L) and land to mobile (L-M) call set up depend upon the technical standard of the particular network. However the general procedure described below holds true for many networks. When the user wishes to make a call, the telephone number to be called is entered followed by a 'call initiation' key (eg pressing the SEND button).

The mobile will transmit an access request to the network on the M-L signalling channel; this may be preceded by the mobile rescan­ning to ensure it is operating on the signalling channel of the nearest base station. If the network can process the call the base station will send a voice channel allocation message which commands the mobile to switch to a designated voice channel, namely one of the channels allocated to that cell. The mobile retunes to the channel indicated and the network proceeds to set up the call to the desired number. As part of the call set up procedure, the network will validate the mobile requesting the call to ensure that it is a legitimate customer. Many networks incorporate specific security features to carry out this validation.

When the network receives a call for a mobile (eg from the PSTN) it will first check the location database to determine in which location area the mobile last registered. Paging calls to the mobile are transmitted on the L-M signalling channels of all the base stations in the identified location area and a response from the mobile awaited. If the mobile is turned on and receives the paging call it will acknowledge to its nearest base station on the M-L signalling channel. The base station receiving the acknow­ledgement sends a voice channel allocation message to the mobile and informs the network so that the two halves of the call can be connected.

Figure 47.3In-call handover

In-call handover

At all times during a call (whether L-M or M-L) the base station currently serving the mobile monitors the signal (strength and/or quality) from the mobile. If the signal falls below a predesignated threshold, the network will command neighbouring base stations to measure the signal from the mobile (Figure 47.3(a)). If another base station is receiving the mobile with a stronger signal than the current base station, a signalling message is sent to the mobile on the voice channel from the current base station commanding the mobile to a new voice channel, namely a free voice channel from those allo­cated to the neighbouring cell. The mobile changes frequency (and thereby the serving base station) and simultaneously the network connects the call to the new base station (Figure 47.3(b)).

The measuring process and new cell selection may take several seconds, but the user will only be aware of a brief break in trans­mission as the mobile tunes to the new voice channel.

Power control

Since the size of a cell may be anything from one kilometre to tens of kilometres across, it is not necessary for a mobile to transmit on full power at all times in order to maintain a satisfactory signal level at the base station's receiver. Most cellular standards therefore incorporate mobile power control, the base station commanding the mobile to transmit at one of a number of power levels. As the mobile moves closer to or further from the base station, further commands are issued to keep the received signal level to prescribed limits. By reducing the average mobile power level, co-channel interference is reduced, improving overall system quality.

Radio planning

As previously described, cellular radio re-uses the same radio chan­nels in different cells and because of this re-use, two mobiles using the same channel in different cells may interfere with each other, a phenomenon known as co-channel interference. The key objective of planning a cellular radio system is to design the cell repeat pattern and frequency allocation in order to maximise the capacity of the system whilst controlling co-channel interference to within accept­able limits.

Figure 47.4Cell repeat patterns: (a) four cell repeat; (b) seven cell repeat; (c) twelve cell repeat

Cell repeat patterns

The cell plan has to be chosen such that the number of channel sets (N) fit together in a regular fashion without gaps or overlaps. Only certain values of N achieve this, and typical arrangements of interest to cellular radio are N = 4, 7 and 12 as shown in Figure 47.4. The value of N has a major effect on the capacity of the cellular system.

As the number of channels sets is decreased, the number of channels per cell increases, hence the system capacity increases. For example, if there are a total of 140 channels available, a 4 cell repeat pattern would provide 35 channels per cell, whilst a 7 cell would provide 20 channels per cell.

On this basis, the smallest possible value of N seems desirable. However, as N decreases, so the distance between cells using the same channels reduces, which in turn increases the level of co-chan­nel interference.

The repeat distance D and the cell radius R are both related by the geometry of the cell pattern. These are shown in Figure 47.5 and Equation 47.1.

In practice, in a real network, it is not possible to achieve a regular cell pattern. This is because radio propagation at the frequencies used by cellular radio systems is affected by the terrain and by buildings, trees and other features of the landscape.

Figure 47.5Frequency re-use - D/R ratio

Co-channel interference

Generally, a mobile will receive a wanted carrier signal (C) from the base station serving the cell in which it is located, and in addition, interfering signals (1) from other cells. The carrier to interference ratio C/I is related to the re-use ratio D/R. Cellular radio systems are designed to tolerate a certain amount of interference, but beyond this, speech quality will be severely degraded. The TACS cellular system, forexample, will work with a C/I down to around 17dB. This lower limit on C/I effectively sets the minimum D/R ratio that can be used.

The two key factors in ensuring that good quality transmission can occur between a mobile and base station are that the wanted signal strength is sufficiently large, that is, above the receiver threshold sensitivity, and that the interference level is low enough to give an adequate C/I ratio. Both ofthese factors depend on the radio propa­gation between the mobile and base stations.

Radio propagation

There are a number of elements which contribute to the received signal strength at a mobile. Firstly, fora line of sight path, there is a free space path loss which is related to the radial distance between base station and mobile.

In addition to this loss, where there is no direct line of sight path, there will be a diffraction loss resulting from obstructions in the path. In general there will also be an effect due to multiple signals arriving at the mobile due to reflections from buildings and other terrain features. This multi-path effect will result in signals either adding constructively or destructively.

As a mobile moves around within a cell it will experience varying signal, as shown in Figure 47.6, due to these factors. Fast fading is caused by the multipath effect, and occurs with only a small move­ment of the mobile. This is also known as Rayleigh fading. Slow fading is mainly caused by terrain features and occurs over large distances ofhundreds of metres.

In addition, the path loss is also dependent upon the type of terrain, for example, urban with dense buildings, or rural with trees, or even over water. The height of the mobile and base station above ground level also affects the propagation, although the mobile height is not usually a variable.

Predicting path loss is an essential part of radio planning, and because of the large number of contributing factors, empirically based formulae are used. The most widely used formula is the Hata model (Hata, 1980) which is based on the propagation measurement results of Okumura et al. (1968).

Hata's basic formula forthe total path loss, Lp, is given by Equation 47.2 where f is the carrier frequency in MHz, k is the base station antenna height, h is the mobile antenna height, R is the radial distance in kilometres, and a(h ) is the mobile antenna height correction factor.

Lp (dB) = 69.55 + 26.16log(f_c) – 13.82log(h_b) – a(h_m) + (44.9 – 6.55log(h_b))logR (47.2)

Correction factors can be used to take into account the type of terrain

Figure 47.6Fading effects

Practical radio planning

Armed with a propagation model it is possible to calculate both the wanted signal strength and the interference level for all locations in a cell. Generally this is done using a computer based tool which can draw upon a database of cell site information and terrain data. Some advanced tools can also take account of diffraction losses. For practical purposes a planner will aim to achieve the required signal strength and C/I ratio over 90% of the cell coverage area, by varying antenna heights, transmitter powers, frequency allocations and other factors as appropriate.

To simplify calculations, an allowance for Rayleigh fading and shadow fading is usually made within the system power budget. A typical power budget is shown in Table 47.1.

Adding capacity

Once a cellular network has been planned to provide overall cover­age, there are a' number of ways of adding additional capacity. A simple and cost effective option is to allocate further radio channels to existing cells. However, this can only be done by an extension band, for example the ETACS allocation in the UK. Other alterna­tives involve rearranging the cellular plan, either by cell splitting or by sectorisation.

Cell splitting is achieved by dividing an existing cell up into a number of smaller cells, by adding additional base stations as shown in Figure 47.7; it is then necessary to reallocate the radio channels. By repeatedly splitting cells; the cell size, and hence the system capacity, can be tailored to meet the traffic capacity requirements demanded by customer behaviour in all areas.

Table 47.1 Typical power budget (TACS) (1 = Key planning parameters)

Figure 47.7Cell splitting

In rural areas, cells may be 20km to 30km in radius. In practice, as cell sizes decrease, propagation effects, particularly in city areas, cause an increase in co-channel interference, even if the repeat pattern is maintained. Also, as cell sizes decrease, it becomes increasingly difficult to find suitable base station sites, which need to be accurately positioned in order to keep to a regular pattern.

The cost of providing and maintaining a large number of individ­ual base stations is also a factor, such that in addition to cell splitting, sectorisation of cells is commonly used in urban areas.

In a regular cellular layout, co-channel interference will be re­ceived from six surrounding cells which all use the same channel set. One way of cutting significantly the level of interference is to use several directional antennas at the base stations, with each antenna illuminating a sector of the cell, and with a separate channel set allocated to each sector.

There are two commonly used methods of sectorisation, using three 120 degree sectors or six 60 degree sectors as shown in Figure 47.8, both of which reduce the number of prime interference sources to one. This is because, of the six surrounding co-channel cells, only one will be directed at the wanted cell.

A disadvantage of sectorisation is that the channel sets are divided between the sectors such that there are fewer channels per sector, and thus a reduction in trunking efficiency. This means that the total traffic which can be carried for a given level of blocking is reduced. However, this effect is offset by the ability to use smaller cells, such that the end result is a significant increase in total capacity.

Figure 47.8Sectorisation

Exercise 1 Learn the words and word combinations

Exercise 2 Read the text

Exercise 3 Find the Russian equivalents for the following English words and word combinations

Exercise 4 Answer the following questions:

Part II (47.4 – 47.4.5)

Overview of systems

TACS

TACS stands for Total Access Communications System, and was adapted from the AMPS standard by the UK when cellular radio was licensed for operation from 1985. The adaptation was necessary to suit European frequency allocations which were at 900MHz, with 25kHz channel spacing. This meant a reduction in frequency devi­ation and signalling speed was necessary (BS, 1990).

The signalling scheme ofAMPS was retained largely unchanged, but some enhancements were introduced, particularly in the proce­dures for location registration, to make the standard more suitable for deployment in systems offering contiguous nationwide cover­age. The opportunity was also taken to introduce extra features, such as signalling of charge rate information (e.g. for payphones).

TACS was originally specified to use the full 1(X)O channels (2 x 25MHz) allocated to mobile services in Europe. However in the UK, only 600 channels (2 x 15MHz) were released by the licensing authority, the remainder being reserved for GSM. Subsequently an additional allocation of channels below the existing TACS channels was made, namely the Extended TACS (ETACS) channels, and the standard was modified accordingly.

TACS equipment availability and cost have both benefitted from the standard's similarity to AMPS, and TACS systems have been adopted by several European countries (UK, Eire, Spain, Italy, Austria and Malta), in the middle east (Kuwait, UAE and Bahrain) and the far east (Hong Kong, Singapore, Malaysia and China). In Europe, TACS is on an equal footing with NMT in terms of installed customer base. A variant of TACS (called J-TACS) has also been adopted in Japan.

NMT

NMT stands for Nordic Mobile Telephone (system), and was de­veloped jointly by the PTTs of Sweden, Norway, Denmark and Finland during the late 1970's/early 1980's.The system was de­signed to operate in the 450MHz band, and was later adapted to also use the 900MHz band. Although NMT was developed after AMPS, it saw commercial service before it, opening in late 1981.

NMT450 uses a channel spacing of 25kHz, speech modulation being analogue FM with a peak frequency deviation of 5kHz, the same as standard PMR practice. NMT900 also uses a frequency deviation of 5kHz, but with a 12.5kHz channel spacing to double the number of available channels, albeit with a degraded adjacent channel rejection performance which must be taken into account during frequency planning. Signalling is at 12(K) bit/s using audio fast frequency shift keying (FFSK). Error protection of the signall­ing information is by means of a Hagelbarger convolutional forward error correcting code.

NMT was designed from the outset to support international roam­ing and was first implemented with full four nation roaming in the four participating countries (Norway, Sweden, Finland and Den­mark). Since then NMT450 has been deployed in many other European countries (Austria, Spain, Netherlands, Belgium, Luxem­bourg, France, Iceland, Faroe Is., Turkey and Hungary) but due to differences in the frequency allocations in the 450MHz band be­tween countries, not all networks are fully compatible to allow roaming.

NMT900 was developed as a necessity as capacity became ex­hausted on the NMT450 networks, and has been deployed since 1987 as an overlay network in several countries, and in Switzerland as their main network.

C450

C450 (also known as Netz-C) was developed by Siemens during the early 1980's under the direction of the (West) German PTT, Deut­sche Bundespost. Commercial service opened in 1985 following a trial period.

C450 has a channel spacing of 20kHz, in common with other mobile services in Germany at 450MHz and speech modulation is analogue FM with a frequency deviation of 4.0kHz. Signalling for call control is transmitted at 5.28kbit/s by direct FSK. Error protec­tion of the signalling is by bit interleaving with a BCH block code backed by an acknowledgement protocol.

In addition, C450 uses continuous signalling between base station and mobile during a call, achieved by time compressing the speech in bursts of 12.5ms, each burst being compressed into 11.4ms. This process opens up slots of 1.1msduration every 12.5ms and the signalling data is inserted into these slots and extracted by the receiver which also time expands the speech back to its original form.

This continuous signalling serves several purposes:

1. It allows the base station to send power control and handover messages to the mobile without disturbing the voice channel.

2. The data is checked for jitter, and thereby the quality of the channel can be determined in order to indicate the need for a handover.

3. The time delay between a base station transmitting a data burst and receiving the response from the mobile is measured at the base station and used to calculate the distance between them.
This distance is also taken into account in handover determina­tion.

4. The data is used as a timing reference by the mobile to lock its internal clocks.

C450 contains a number of advanced features made possible by the application of current developments in technology. Although speech transmission is analogue, it can be regarded as a hybrid technology system, and several of its characteristics such as time slotted signalling channels and continuous signalling during call have been carried through into the GSM system design.

Coming later to the European scene, C450 has chiefly only served the German market, although systems are also operating in Portugal and South Africa.

GSM

The GSM standard was developed as a joint initiative by the mem­bers of the Conference of European Posts and Telecommunications administrations (CEPT) with the eventual aim of building a unified pan-European network, giving the user a near uniform service throughout all European countries. An added bonus of a common standard should be lower terminal equipment prices through econ­omies of scale.

Work on the standard started in 1982, and by 1987 all the basic architectural features were decided. The full Phase 1 specification was completed in 1990, but work continues on further phases incorporating new features and services. In 1987, the majority of operators participating in GSM signed a Memorandum of Under­standing (MoU) committing them to make GSM a reality by install­ing networks and opening commercial service by 1991. Since that time further operators have signed the MoU, bringing the total to date to 25.

The GSM technical standard makes full use of currently available levels of technology, incorporating features such as low bit rate speech, convolutional channel coding with bit interleaving and frequency hopping. The standard is intended to endure for many years to come.

Exercise 1 Learn the words and word combinations

Exercise 2 Read the text

Exercise 3 Find the Russian equivalents for the following English words and word combinations

Exercise 4 Answer the following questions:

Part III (47.5.1 – 47.5.7)

Detailed description of GSM

GSM architecture

The basic architecture of GSM is not dissimilar to other cellular radio systems and comprises base transceiver stations (BTS), Base Station Controllers (BSC), Mobile Switching Centres (MSC), a variety of registers and a network management system, as shown in Figure 47.9. The mobile station comprises a mobile equipment and a subscriber identity module (SIM). In addition to these functional entities, GSM also defines several interfaces, the Radio Interface (Um), the interface between the MSC and BSC (A interface) and the signalling interface which allows roaming between networks. This is based on the CCITT No.7 signalling standard and is defined as a Mobile Application Part (MAP).

The BTS and BSC together form the Base Station Subsystem (BSS) and carry out all the functions related to the radio channel management. This includes the management of the radio channel configurations, allocating radio channels for speech, data and signalling purposes, and controlling frequency hopping and power control. The BSS also includes, as does the MS, the speech encoding and decoding, and channel coding and decoding.

The MSC, VLR and HLR are concerned with mobility manage­ment functions. These include authentication and registration of the mobile customer, location updating, and call set up and release. The HLR is the master subscriber database and carries information about individual subscribers numbers, subscription levels, call restriction-s, supplementary services and the current location (or most recent location) of subscribers. The VLR acts as a temporary subscriber database forall subscribers within its coverage area, and contains similar information to that in the HLR. The provision of a VLR means that the MSC does not need to access the HLR forevery transaction.

The authentication centre (AUC) works closely with the HLR and provides information to authenticate all calls in order to guard against fraud. The equipment identity register (EIR) is used for equipment security and validation of different types of mobile equipment. This information can be used to screen mobile types from accessing the system, for example if a mobile equipment is stolen, not type approved, or has a fault which could disturb the network.

Network management is used to monitor and control the major elements of the GSM network. In particular, it monitors and reports faults and performance data. It can also be used to re-configure the network.

MS Mobile Station

MSC Mobile Switching Centre

BTS Base Transceiver Station

BSC Base Station Controller

VLR Visited Location Register

HLR Home Location Register

EIR Equipment Identity Regsiter

AUC Authentication centre

Figure 47.9GSM architecture

Air interface

The GSM Air Interface (Urn) provides the physical link between the mobile and the network. Some of the key characteristics of the air interface are given in Table 47.3. As already described, GSM is a digital system employing time division multiple access (TDMA) techniques and operating in the 900MHz band. The CEPT have made available two frequency bands to be used throughout Europe by the GSM system, namely;

1. 890MHz to 915MHz forthe mobile to base station (uplink)direction.

2. 935MHz to 960MHz for the base station to mobile (downlink)direction.

These 25MHz bands are divided into 124 pairs of carriers spaced by 200kHz. In addition, consideration is now being given to specif­ying additional carriers in a pair of extension bands 872MHz to 888MHz and 917MHz to 933MHz. Each of the carriers is divided up into eight TDMA timeslots of length 0.577ms such that the frame length is 4.615ms. The recurrence of each timeslot makes up one physical channel, such that each carrier can support eight physical channels, in both the uplink and downlink directions.

The timeslot allocation in either direction is staggered so that the mobile station does not need to transmit and receive at the same time. Data is transmitted in bursts within the timeslots and a number of different types ofburst can be carried as shown in Figure 47.10. The normal burst has a data structure as shown. It consists of148 bits of which 114 are available fordata transmission, 26 are used for a training sequence which allows the receiver to estimate the radio propagation characteristics and to set up a dispersion equaliser, 6 bits as tail bits, and two stealing flags. These physical channels therefore provide a data throughput of 114 bits every 4.615ms or 24.7kbit/s.

The bursts modulate one of the RF carriers using Gaussian Mini­mum Shift Keying (GMSK) modulation with a BT index of 0.3. The allocation ofthe carrier can be such that frequency hopping is achieved, i.e consecutive bursts of a physical channel will be carried by differing RF carriers. This "hopping" is performed every TDMA frame, or every 4.615ms and provides extra protection against channel fading and co-channel interference.

A number oflogical channels can be carried by the physical channels described above. These are summarised in Table 47.4.

There are two categories of traffic channels; speech, whether full rate using 22.8kbit/s or half rate using 11.4kbit/s, and data, provid­ing a variety of data rates. There are four basic categories of control channels, known as the broadcast control channel (BCCH), the common control channel (CCCH), the standalone dedicated control channel (SDCCH) and the associated control channel (ACCH).

These are further divided into channels with specific purposes and for a detailed description of these channels the reader is referred to the GSM Recommendations published by ETS1.

Each of these logical channels is mapped onto the physical chan­nels, using the appropriate burst type as shown in Figure 47.10.

TDMA frames are built up into 26 or 51 frame multiframes, such that individual timeslots can use either of the multiframe types, and then into superframes and hyperframes as shown in Figure 47.10. The TCH and the associated ACCH uses the 26 frame structure, whilst the BCCH and CCCH use the 51 frame structure. The SDCCH may occupy one physical channel, providing 8 SDCCH, or may share a physical channel with the BCCH/CCCH. Typical ar­rangements for allocating the 8 physical channels could be:

1. 7 channels TCH and SACCH + 1 channelBCCH/CCCH/SDCCH

2. 6 channels TCH and SACCH + 1 channel BCCH/CCCH + 1channel SDCCH.

Each cell must have at least one physical channel assigned to the BCCH/CCCH, where there are 2 or more carriers per cell, the non-BCCH carriers may have all 8 channels allocated to TCH.

Table 47-3 GSM air interface parameters

Figure 47.10GSM timeframes, timeslots and bursts (Extract from GSM Recommendation 05.01)

12
• 3
• Следующая ⇒