Introduction to the GSM System

Chapter 1: Introduction to the GSM System
This chapter provides an overview of the GSM cellular system, with a focus on the radio interface. The purpose is not to give a detailed description of the many features supported by the system, but to summarize the elementary concepts of GSM, as an aid to reader comprehension of the subsequent chapters. In-depth presentation of the GSM system can be found in [1, 2].

1.1 Introduction
1.1.1 Birth of the GSM System
The first step in the history of GSM development was achieved back in 1979, at the World Administrative Radio Conference, with the reservation of the 900-MHz band. In 1982 at the Conference of European Posts and Telegraphs (CEPT) in Stockholm, the Groupe Spécial Mobile was created, to implement a common mobile phone service in Europe on this 900-MHz frequency band. Currently the acronym GSM stands for Global System for Mobile Communication; the term "global" was preferred due to the intended adoption of this standard in every continent of the world.

The proposed system had to meet certain criteria, such as:

Good subjective speech quality (similar to the fixed network);

Affordability of handheld terminals and service;

Adaptability of handsets from country to country;

Support for wide range of new services;

Spectral efficiency improved with respect to the existing first-generation analog systems;

Compatibility with the fixed voice network and the data networks such as ISDN;

Security of transmissions.

Digital technology was chosen to ensure call quality.

The basic design of the system was set by 1987, after numerous discussions led to the choice of key elements such as the narrowband time-division multiple access (TDMA) scheme, or the modulation technique. In 1989 responsibility for the GSM was transferred to the European Telecommunication Standards Institute (ETSI). ETSI was asked by the EEC to unify European regulations in the telecommunications sector and in 1990 published phase I of the GSM system specifications (the phase 2 recommendations were published in 1995).

The first GSM handset prototypes were presented in Geneva for Telecom '91, where a GSM network was also set up. Commercial service had started by the end of 1991, and by 1993 there were 36 GSM networks in 22 countries. The system was standardized in Europe, but is now operational in more than 160 countries all over the world, and was adopted by 436 operators.

The growth of subscribers has been tremendous, reaching 500 million by May 2001.

1.1.2 The Standard Approach
The thousands of pages of GSM recommendations, designed by operators and infrastructure and mobile vendors, provide enough standardization to guarantee proper interworking between the components of the system. This is achieved by means of the functional and interface descriptions for each of the different entities.

The GSM today is still under improvement, with the definition of new features and evolution of existing features. This permanent evolution is reflected in the organization of the recommendations, first published as phase I, then phase II and phase II+, and now published with one release each year (releases 96, 97, 98, 99, and releases 4 and 5 in 2000 and 2001).

As stated, responsibility for the GSM specifications was carried by ETSI up to the end of 1999. During 2000, the responsibility of the GSM recommendations was transferred to the Third Generation Partnership Project (3GPP). This world organization was created to produce the third-generation mobile system specifications and technical reports. The partners have agreed to cooperate in the maintenance and development of GSM technical specifications and technical reports, including evolved radio access technologies [e.g., General Packet Radio Service (GPRS) and Enhanced Data rates for Global Evolution (EDGE)]. The structure and organization of the 3GPP is further described in Section 2.5.



1.2 General Concepts
1.2.1 Analog Versus Digital Telephony Systems
First-generation systems were analog. During the early 1980s these systems underwent rapid development in Europe. Although the NMT system was used by all the Nordic countries, and the TACS system in the United Kingdom and Italy, there was a variety of systems and no compatibility among them. Compared with these systems, the main advantages offered by GSM, which is the most important of the second-generation digital systems, are:

Standardization;

Capacity;

Quality;

Security.

Standardization guarantees compatibility among systems of different countries, allowing subscribers to use their own terminals in those countries that have adopted the digital standard. The lack of standardization in the first-generation system limited service to within the borders of a country. Mobility is improved, since roaming is no longer limited to areas covered by a certain system (see Section 1.2.6). Calls can be charged and handled using the same personal number even when the subscriber moves from one country to another.

Standardization also allows the operator to buy entities of the network from different vendors, since the functional elements of the network and the interfaces between these elements are standardized. This means that a mobile phone from any manufacturer is able to communicate with any network, even if this network is built with entities from different vendors. This leads to a large economy of scale and results in cost reduction for both the operator and the subscriber. Furthermore, the phone cost is also reduced, because as GSM is an international standard, produced quantities are greater and the level of competition is high.

With respect to capacity, the use of the radio resource is much more efficient in a digital system such as GSM than in an analog system. This means that more users can be allocated in the same frequency bandwidth. This is possible with the use of advanced digital techniques, such as voice compression algorithms, channel coding, and multiple access techniques. Note that capacity gains are also achieved with radio frequency reuse, which had also been used in analog systems. Frequency reuse means that a given carrier can be employed in different areas, as explained in Section 1.2.2.

The quality in digital transmission systems is better, thanks to the channel coding schemes that increase the robustness in the face of noise and disturbances such as interference caused by other users or other systems. The quality improvement is also due to the improved control of the radio link, and adaptations to propagation conditions, with advanced techniques such as power control or frequency hopping. This will be explained in greater detail in Section 1.5.6.3.

In terms of security, powerful authentication and encryption techniques for voice and data communications are enabled with GSM, which guarantees protected access to the network, and confidentiality.

1.2.2 Cellular Telephony
In mobile radio systems, one of the most important factors is the frequency spectrum. In order to make the best use of the bandwidth, the system is designed by means of the division of the service area into neighboring zones, or cells, which in theory have a hexagonal shape. Each cell has a base transceiver station (BTS), which to avoid interference operates on a set of radio channels different from those of the adjacent cells. This division allows for the use of the same frequencies in nonadjacent cells. A group of cells that as a whole use the entire radio spectrum available to the operator is referred to as a cluster. The shape of a cell is irregular, depending on the availability of a spot for the BTS, the geography of the terrain, the propagation of the radio signal in the presence of obstacles, and so on.

In dense urban areas, for instance, where the mobile telephony traffic is important, the diameter of the cells is often reduced in order to increase capacity. This is allowed since the same frequency channels are used in a smaller area. On the other hand, reducing the cell diameter leads to a decrease in the distance necessary to reuse the frequencies (that is, the distance between two cochannel cells), increasing cochannel interference. In order to minimize the level of interference, several techniques are used on the radio interface.

A basic example of cluster organization is shown in Figure 1.1. In this example, we see a reuse pattern for seven different frequencies, f1 to f7. These frequencies correspond to the beacon carrier of each cell, on which signaling information about the cell is broadcast (see Section 1.2.7). It can be seen from this figure that a given carrier can be reused in two separate geographical areas, as long as these areas are far enough from each other to reduce the effect of interference. With this technique of dividing the area in cells and clusters, the operator can increase the area it is able to cover with a limited frequency bandwidth.


Figure 1.1: Example of a cell planning.
1.2.3 Public Land Mobile Network
A public land mobile network (PLMN) is a network established for the purpose of providing land mobile telecommunications services to the public. It may be considered as an extension of a fixed network, such as the Public Switched Telephone Network (PSTN), or as an integral part of the PSTN.

1.2.4 Multiband Mobile Phones
Because of the increasing demand on the mobile networks, today the mobile stations (MSs) tend to be multiband. Indeed, to avoid network saturation in densely populated regions, mobile phones capable of supporting different frequency bands have been implemented, to allow for the user making communications in any area, at any time.

A dual-band phone can operate in two different frequency bands of the same technology, for instance in the 900-MHz and 1800-MHz frequency bands of the GSM system. Triple-band mobile phones have also come on the market, with the support of GSM-900 (900-MHz GSM band), DCS-1800 (1800-MHz GSM band), and PCS-1900 (1900-MHz GSM band), for example. Note that DCS-1800 and PCS-1900 are never deployed in the same country, and therefore this kind of phone can be used by travelers who want to have service coverage in a large number of countries.

1.2.5 SIM Card
One of the most interesting innovations of GSM is that the subscriber's data is not maintained in the mobile phone. Rather a "smart card," called a subscriber identity module (SIM) card, is used.

The SIM is inserted in the phone to allow the communications. A user may thus make telephone calls with a mobile phone that is not his own, or have several phones but only one contract. It is for example possible to use a SIM card in a different mobile when traveling to a country that has adopted the GSM on a different frequency band. A European can therefore rent a PCS1900 phone when traveling to the United States, while still using his own SIM card, and thus may receive or send calls. The SIM is used to keep names and phone numbers, in addition to those that are already kept in the phone's memory.

The card is also used for the protection of the subscriber, by means of a ciphering and authentication code.

1.2.6 Mobility
GSM is a cellular telephony system that supports mobility over a large area. Unlike cordless telephony systems, it provides location, roaming, and handover.

1.2.6.1 Location Area
The ability to locate a user is not supported in first-generation cellular systems. This means that when a mobile is called, the network has to broadcast the notification of this call in all the radio coverage. In GSM, however, location areas (LAs), which are groups of cells, are defined by the operator. The system is able to identify the LA in which the subscriber is located. This way, when a user receives a call, the notification (or paging) is only transmitted in this area. This is far more efficient, since the physical resource use is limited.

1.2.6.2 Roaming
In particular, the GSM system has the capability of international roaming, or the ability to make and receive phone calls to and from other nations as if one had never left home. This is possible because bilateral agreements have been signed between the different operators, to allow GSM mobile clients to take advantage of GSM services with the same subscription when traveling to different countries, as if they had a subscription to the local network. To allow this, the SIM card contains a list of the networks with which a roaming agreement exists.

When a user is "roaming" to a foreign country, the mobile phone automatically starts a search for a network stipulated on the SIM card list. The choice of a network is performed automatically, and if more than one network is given in the list, the choice is based on the order in which the operators appear. This order can be changed by the user. The home PLMN is the network in which the user has subscribed, while the visited PLMN often refers to the PLMN in which the user is roaming. When a user receives a call on a visited PLMN, the transfer of the call from the home PLMN to the visited PLMN is charged to the called user by his operator.

1.2.6.3 Handover
When the user is moving from one cell to the other during a call, the radio link between BTS 1 and the MS can be replaced by another link, between BTS 2 and the MS. The continuity of the call can be performed in a seamless way for the user. This is called handover. With respect to dual-band telephones, one interesting feature is called the dual-band handover. It allows the user in an area covered both by the GSM-900 and by the DCS-1800 frequency bands, for instance, to be able to transfer automatically from one system to the other in the middle of a call.

1.2.7 Beacon Channel
For each BTS of a GSM network, one frequency channel is used to broadcast general signaling information about this cell. This particular carrier frequency is called a beacon channel, and it is transmitted by the BTS with the maximum power used in the cell, so that every MS in the cell is able to receive it.

1.2.8 MS Idle Mode
When it is not in communication, but still powered on, the MS is said to be in idle mode. This means that it is in a low consumption mode, but synchronized to the network and able to receive or initiate calls.



1.3 GSM Services
In the specification of a telecommunication standard such as GSM, the first step is of course the definition of the services offered by the system. GSM is a digital cellular system designed to support a wide variety of services, depending on the user contract and the network and mobile equipment capabilities.

In GSM terminology, telecommunication services are divided into two broad categories:

Bearer services are telecommunication services providing the capability of transmission of signals between access points [the user-network interfaces (UNIs) in ISDN]. For instance, synchronous dedicated packet data access is a bearer service.

Teleservices are telecommunication services providing the complete capability, including terminal equipment functions, for communication between users according to protocols established by agreement between network operators.

In addition to these services, supplementary services are defined that modify or supplement a basic telecommunication service.

1.3.1 Bearer Services
There exist several categories of bearer services:

Unrestricted digital information (UDI) is designed to offer a peer-to-peer digital link.

The 3.1 kHz is external to the PLMN and provides a UDI service on the GSM network, interconnected with the ISDN or the PSTN by means of a modem.

PAD allows an asynchronous connection to a packet assembler/disassembler (PAD). This enables the PLMN subscribers to access a packet-switched public data network (PSPDN).

Packet enables a synchronous connection to access a PSPDN network and alternate speech and data, providing the capability to switch between voice and data during a call.

Speech followed by data first provides a speech connection, and then allows to switch during the call for a data connection. The user cannot switch back to speech after the data portion.

1.3.2 Teleservices
In terms of application, teleservices correspond to the association of a particular terminal to one or several bearer services. They provide access to two kinds of applications:

Between two compatible terminals;

From an access point of the PLMN to a system including high-level functions, for example, a server.

Of course, the most basic teleservice supported by GSM is digital voice telephony, based on transmission of the digitally encoded voice over the radio. The voice service also includes emergency calls, for which the nearest emergency-service provider is notified by dialing three digits.

The other teleservices that are defined for a PLMN are:

Data services, with data rates ranging from 2.4 Kbps to 14.4 Kbps. These services are based on circuit-switched technology. Circuit switched means that during the communication, a circuit is established between two entities for the transfer of data. The physical resource is used during the whole duration of the call.

Short message service (SMS), which is a bidirectional service for short alphanumeric (up to 160 bytes) messages.

Access to a voice message service.

Fax transmission.

1.3.3 Supplementary Services
Supplementary services include several forms of call forward (such as call forwarding when the mobile subscriber is unreachable by the network), caller identification, call waiting, multiparty conversations, charging information, and call barring of outgoing or incoming calls. These call-barring features can be used for example when roaming in another country, if the user wants to limit the communication fees.



1.4 Network Architecture
The structure of a GSM network relies on several functional entities, which have been specified in terms of functions and interfaces. It involves three main subsystems, each containing functional units and interconnected with the others through a series of standard interfaces.

The main parts of a GSM network, as shown in Figure 1.2, are listed below. (In the figure, the lines between the entities represent the interfaces.)


Figure 1.2: General architecture of a GSM network.
The MS, the handheld mobile terminal;

The base station subsystem (BSS), which controls the radio link with the MS;

The network and switching subsystem (NSS), which manages the function of connection switching to other fixed public network or mobile network subscribers, and handles the databases required for mobility management and for the subscriber data.

As can be seen from the figure, in a PLMN, the radio access network part (BSS) is logically separated from the core network (NSS), in order to ease the standardization of the different functions.

1.4.1 MS
The MS is made up of the mobile equipment (ME), and a SIM. It performs the following functions:

Radio transmission and reception;

Source and channel coding and decoding, modulation and demodulation;

Audio functions (amplifiers, microphone, earphone);

Protocols to handle radio functions: power control, frequency hopping, rules for access to the radio medium;

Protocol to handle call control and mobility;

Security algorithms (encryption techniques).

As mentioned, the SIM enables the user to have access to subscribed services irrespective of a specific terminal. The insertion of the SIM card into any GSM terminal allows the user to receive calls on that terminal, to make calls from that terminal, and to use the other subscribed services. The ME is identified with an international mobile equipment identity (IMEI).

The SIM card contains, among other information, the international mobile subscriber identity (IMSI) used to identify the subscriber to the system, and a secret key for authentication. The IMEI and the IMSI are independent, thereby allowing personal mobility.

1.4.2 BSS
The BSS is composed of several base station controllers (BSCs) and BTS. These two elements communicate across the Abis interface.

The BTS contains the radio transceivers, responsible for the radio transmissions with the MS. This includes the following functions:

Modulation and demodulation;

Channel coding and decoding;

Encryption process;

RF transmit and receive circuits (power control, frequency hopping, management of antenna diversity, discontinuous transmission).

Several types of BTS exist: the normal BTS, the micro BTS, and the pico BTS. The micro BTS is different from a normal BTS in two ways. First, the range requirements are reduced, and the close proximity requirements are more stringent. Second, the micro BTS is required to be small and affordable in order to allow external street deployment in large numbers. The pico BTS is an extension of the micro BTS concept to the indoor environments. The RF performances of these different BTSs are slightly different.

The BSC manages the radio resources for one or more BTSs. It handles the management of the radio resource, and as such it controls the following functions: allocation and release of radio channels, frequency hopping, power control algorithms, handover management, choice of the encryption algorithm, and monitoring of the radio link.

1.4.3 Network Subsystem
The mobile services switching center (MSC) is the central part of the network subsystem (NSS). It is responsible for the switching of calls between the mobile users (between different BSCs or toward another MSC) and between mobile and fixed network users. It manages outgoing and incoming calls from various types of networks, such as PSTN, ISDN, and PDN. It also handles the functionality required for the registration and authentication of a user, and the mobility operations. This includes location updating, inter-MSC handovers, and call routing.

The BSS communicates with the MSC across the A interface.

Associated with the MSC, two databases, the home location register (HLR) and the visitor location register (VLR), provide the call-routing and roaming capabilities. The HLR contains all the administrative information related to the registered subscribers within the GSM network. This includes the IMSI, which unequivocally identifies the subscriber within any GSM network, the MS ISDN number (MSISDN), and the list of services subscribed by the user (such as voice, data service). The HLR also stores the current location of the MS, by means of the address of the VLR in which it is registered.

The VLR temporarily keeps the administrative data of the subscribers that are currently located in a given geographical area under its control.

Each functional entity may be implemented as an independent unit, but most of the time, the VLR is colocated with the MSC, so that the geographical area controlled by the MSC corresponds to that controlled by the VLR. The MSC contains no information about particular MSs, but rather, the information is stored in the location registers.

Two other registers are used for authentication and security purposes:

The equipment identity register (EIR) is a database that contains a list of all valid ME on the network, where each MS is identified by its IMEI. An IMEI is marked as invalid if it has been reported stolen.

The authentication center (AuC) is a protected database that contains a copy of the secret key stored in each subscriber's SIM card, for authentication and encryption over the radio channel. The AuC verifies if a legitimate subscriber has requested a service. It provides the codes for both authentication and encryption to avoid undesired violations of the system by third parties.

Two other important entities of the NSS are the operations and maintenance center (OMC) and the network management center (NMC). These entities perform the functions relative to the network management (NM), such as the configuration of the system (locally or remotely), maintenance and tests of the pieces of equipment, billing, statistics on the performance, and gathering of all information related to subscriber traffic necessary for invoicing and administration of subscribers.


1.5 Radio Interface
1.5.1 General Characteristics
Currently, there are several types of networks in the world using the GSM standard, but at different frequencies.

The GSM-900 is the most common in Europe and the rest of the world. Its extension is E-GSM.

The DCS-1800 operates in the 1,800-MHz band and is used mainly in Europe, usually to cover urban areas. It was also introduced to avoid saturation problems with the GSM-900.

The PCS-1900 is used primarily in North America.

The GSM-850 is under development in America.

The GSM-400 is intended for deployment in Scandinavian countries in the band previously used for the analog Nordic Mobile Telephony (NMT) system.

The system is based on frequency-division duplex (FDD), which means that the uplink (radio link from the mobile to the network-that is, mobile transmit, base receive), and downlink (from the network to the mobile-that is, base transmit, mobile receive) are transmitted on different frequency bands. For instance, in the 900-MHz E-GSM band, the block 880-915 MHz is used for transmission from mobiles to network, and the block 925-960 MHz is used for the transmission from network to mobiles. Table 1.1 gives a summary of uplink and downlink frequency bands for the different GSM systems.

Table 1.1: GSM System Frequency Bands Uplink Band
Downlink Band

GSM-900
890-915 MHz
935-960 MHz

E-GSM-900
880-915 MHz
925-960 MHz

DCS-1800
1,710-1,785 MHz
1,805-1,880 MHz

PCS-1900
1,850-1,910 MHz
1,930-1,990 MHz

GSM-400
GSM-450
450,4-457,6 MHz
460.4-467.6 MHz

GSM-480
478.8-486 MHz
488.8-496 MHz

GSM-850
824-849 MHz
869-894 MHz


Operators may implement networks that operate on a combination of the frequency bands listed above to support multiband mobile terminals.

There are different ways of sharing the physical resource among all the users in a radio system, and this is called the multiple-access method. The multiple-access scheme defines how simultaneous communications share the GSM radio spectrum. The various multiple-access techniques in use in radio systems are frequency-division multiple access (FDMA), TDMA, and code-division multiple access (CDMA). GSM is based on both FDMA and TDMA techniques (see Figure 1.3).


Figure 1.3: TDMA and FDMA.
FDMA consists in dividing the frequency band of the system into several channels. In GSM, each RF channel has a bandwidth of 200 kHz, which is used to convey radio modulated signals, or carriers. Each pair of uplink/ downlink channels is called an absolute radio frequency channel (ARFC) and is assigned an ARFC number (ARFCN). The mapping of each ARFCN on the corresponding carrier frequency is given in [3].

TDMA is the division of the time into intervals: within a frequency channel, the time is divided into time slots. This division allows several users (eight) to be multiplexed on the same carrier frequency, each user being assigned a single time slot. A packet of data information, called a burst, is transmitted during a time slot. The succession of eight time slots is called a TDMA frame, and each time slot belonging to a TDMA frame is identified by a time slot number (TN), from 0 to 7.

1.5.2 Logical Channels
The association of a radio frequency channel and a time slot-the pair ARFCN and TN-uniquely defines a physical channel on both the uplink and the downlink.

On top of the physical channels, logical channels ar mapped to convey the information of voice, data, and signaling. This signaling information is used for setting up a call, or to adapt the link to rapidly changing radio conditions, or to manage handovers, to give a few examples. Logical channels can be seen as pipes, each one used for a different purpose by the higher layers of the system.

Two types of logical channels exist, traffic channels and control channels. Among the control channels, according to their functions, four classes are defined: broadcast, dedicated, common, and associated. A broadcast channel is used by the network (in downlink only) to send general information to the MSs. A channel is said to be dedicated if only one MS can transmit or receive in the ARFCN-TN defining this channel, and common if it carries information for several mobiles. An associated control channel is allocated to one mobile, in addition to a dedicated channel, and carries signaling for the operation of this channel.

The broadcast channels are transmitted on the beacon carrier frequency presented in Section 1.2.7. The purposes of the beacon are:

To allow a synchronization in time and frequency of the MSs to the BTS. This synchronization is needed by the MS to access the services of a cell. The frequency and time synchronization procedures that are performed by the mobile are explained in Section 1.5.7.

To help the mobile in estimating the quality of the link during a communication, by measurements on the received signal from the BTS it is transmitting to, and from the other BTSs of the geographical area. These measurements are used by the network to determine when a handover is necessary, and to which BTS this handover should apply.

To help the mobile in the selection of a cell when it is in idle mode (that is, not in communication, but still synchronized to the system and able to receive an incoming call or to initiate a call). This selection is performed on the basis of the received power measurements made on the adjacent cells' beacon channels.

To access the general parameters of the cell needed for the procedures applied by the MS, or general information concerning the cell, such as its identification, the beacon frequencies of the surrounding cells, or the option supported by the cell (services).

To allow these various operations, the logical channels transmitted on the beacon are:

The broadcast control channel (BCCH), which continually broadcasts, on the downlink, general information on the cell, including base station identity, frequency allocations, and frequency-hopping sequences. The information is transmitted within system information (SI) blocks, which can be of different types according to the information that is carried out. The frequency with which an SI is retransmitted on the BCCH varies with the type of information.

The frequency control channel (FCCH), used by the MS to adjust its local oscillator (LO) to the BTS oscillator, in order to have a frequency synchronization between the MS and the BTS.

The synchronization channel (SCH), used by the MS to synchronize in time with the BTS, and to identify the cell.

As listed below, four channels comprise the common control channels (CCCH). Among these, the first three are used for the MS-initiated call or for call paging (notification of an incoming call toward the MS):

The random access channel (RACH) is used for the MS access requests to the network, for the establishment of a call, based on a slotted aloha method.

The paging channel (PCH) is defined to inform the MS of an incoming call.

The access grant channel (AGCH) is used to allocate some physical resource to a mobile for signaling, following a request on the RACH.

The cell broadcast channel (CBCH) may be used to broadcast specific news to the mobiles of a cell.

The dedicated control channels are:

The stand-alone dedicated control channel (SDCCH), utilized for registration, authentication, call setup, and location updating.

The slow associated control channel (SACCH), which carries signaling for the TCH or SDCCH with which it corresponds. The information that is transmitted on this channel concerns the radio link control (RLC), such as the power control on the corresponding TCH or SDCCH, or the time synchronization between the MS and the BTS.

The fast associated control channel (FACCH), carries the signaling that must be sent by the network to the MS to notify that a handover is occurring.

The TCHs can be of several types, according to the service that is accessed by the subscribers: voice or data, with various possible data rates.

Table 1.2 summarizes the purpose of the different logical channels. In this table, UL stands for uplink, and DL for downlink.

Table 1.2: The Logical Channels and Their Purpose Logical Channel
Abbreviation
Uplink/ Downlink
Task

Broadcast channel (BCH)
Broadcast control channel
BCCH
DL
System Information broadcast

Frequency correction channel
FCCH
DL
Cell frequency synchronization

Synchronization channel
SCH
DL
Cell time synchronization and identification

Common control channel (CCCH)
Paging channel
PCH
DL
MS paging

Random access channel
RACH
UL
MS random access

Access grant channel
AGCH
DL
Resource allocation

Cell broadcast channel
CBCH
DL
Short messages broad cast

Dedicated control channel
Standalone dedicated control channel
SDCCH
UL/DL
General signaling

Slow associated control channel
SACCH
UL/DL
Signaling associated with the TCH

Fast associated control channel
FACCH
UL/DL
Handover signaling

Traffic channel (TCH)
Full speech
TCH/FS
UL/DL
Full-rate voice channel

Half rate
TCH/HS
UL/DL
Half-rate voice channel

2.4 Kbps, 4.8 Kbps, 9.6 Kbps, and 14.4 Kbps full-rate data channels
TCH/F2.4 TCH/F4.8 TCH/F9.6 TCH/F14.4
UL/DL
Full-rate data channels

2.4-Kbps- and 4.8-Kbps-rate data channels
TCH/H2.4 TCH/H4.8
UL/DL
Half-rate data channels


1.5.3 Mapping of Logical Channels onto Physical Channels
1.5.3.1 TDMA Time Structure
The basic time unit is the time slot. Its duration is 576.9 μs = 15/26 ms, or 156.25 symbol periods (a symbol period is 48/13 μs). The piece of information transmitted during a time slot is called a burst. As we saw in Section 1.5.1, the GSM multiple access scheme is TDMA, with eight time slots per carrier. A sequence of eight time slots is called a TDMA frame, and has a duration of 4.615 ms. The time slots of a TDMA frame are numbered from 0 to 7, as shown in Figure 1.4. Note that the beginning and end of TDMA frames in uplink and downlink are shifted in time: Time slot number 0 on the uplink corresponds to time slot 3 in the downlink. This allows some time for the mobile to switch from one frequency to the other.


Figure 1.4: Slot numbering within the TDMA frame.
As seen earlier, a physical channel is defined as a sequence of TDMA frames, a time slot number (from 0 to 7) and a frequency. It is bidirectional, with the same TN in uplink and in downlink. In order to support cryptographic mechanisms, a long time-structure has been defined. It is called a hyperframe and has a duration of 3 hours, 28 minutes, 53 seconds, and 760 ms (or 12,533.76 seconds). The TDMA frames are numbered within the hyperframe. The numbering is done with the TDMA frame number (FN) from 0 to 2,715,647.

One hyperframe is subdivided into 2,048 superframes, which have a duration of 6.12 seconds. The superframe is itself subdivided into multi-frames. In GSM, there are two types of multiframes defined, containing 26 or 51 TDMA frames.

The 26 multiframe has a duration of 120 ms, and comprises 26 TDMA frames This multiframe is used to carry TCH, SACCH, and FACCH. The 51 multiframe is made up of 51 TDMA frames. Its duration is 235.4 ms (3,060/13 ms). This multiframe is used to carry BCH, CCCH, and SDCCH (with its associated SACCH). Note that a superframe is composed of twenty-six 51-multiframes, or of fifty-one 26-multiframes. This hierarchical time structure is summarized in Figure 1.5.


Figure 1.5: Hierarchical structure of a hyperframe.
1.5.3.2 Mapping of the TCH and SACCH on the 26-Multiframe
The TCHs are bidirectional channels mapped onto the 26-multiframe. Two types of channels must be distinguished: full-rate and half-rate channels, and therefore two different mappings of the TCH on the multiframe are possible:

A full-rate traffic channel (TCH/FS, for full speech) uses one time slot per TDMA frame, for each frame of the multiframe, except the frames 12 and 25 (see Figure 1.6). The TDMA frame 12 is used to carry the SACCH/FS, and the TDMA frame 25 is an idle frame, which means that no channel is transmitted during this entire TDMA frame.


Figure 1.6: Mapping of a TCH/FS and SACCH/FS on the 26-multiframe.

A half-rate traffic channel (TCH/HS) uses one time slot every two TDMA frames, due to the fact that it carries data from a half-rate voice coder. As shown in Figure 1.7, two half-rate channels can be mapped on the same time slot, one using TDMA frames 0, 2, 4, 6, 8, 10, 13, 15, 17, 19, 21, and 23 and the other one using frames 1, 3, 5, 7, 9, 11, 14, 16, 18, 20, 22, and 24. The SACCH/HS channel associated with the first TCH subchannel is transported on TDMA frame 12, and the SACCH/HS associated with the second subchannel is on time slot 25.


Figure 1.7: Mapping of a TCH/HS and SACCH/HS on the 26-multiframe.

1.5.3.3 Mapping of the FACCH on the 26-Multiframe
The FACCH is associated with a TCH, and is required to support the highspeed signaling needed during call establishment, subscriber authentication, and handover management. The occurrence of the FACCH is not fixed in the multiframe, as it is for the SACCH. Rather, the FACCH occurs on a TDMA frame that is reserved for a TCH. The multiplexing of TCH and FACCH is possible by means of the frame stealing. This means that a speech frame carried over a TCH can be replaced by a FACCH frame. This is signaled to the receiver by means of stealing flags, as described in Section 1.5.4. This principle allows for a fast signaling channel without significant loss on the quality of speech, if it is not performed too often.

1.5.3.4 Mapping of the SDCCH on the 51-Multiframe
The SDCCH is a signaling channel that carries the higher layers of control information. A SACCH is associated with a SDCCH.

Two different multiplexings on the 51 multiframe are possible:

SDCCH with SACCH alone. An SDCCH channel is mapped on four TDMA frames of a 51 multiframe. As a result, eight SDCCH channels, dedicated to eight different MSs are mapped onto a 51 multiframe. The TDMA frames not occupied by an SDCCH are used by the eight associated SACCH channels. The mapping of these associated channels is performed on two consecutive 51 multi-frames, as shown in Figure 1.8(b).

SDCCH with SACCH multiplexed together with CCCH, BCCH, SCH. and FCCH. This case is described in the next subsection and in Figure 1.8(c).


Figure 1.8: Channel associations on the 51-multiframe. (a) BCCH + CCCH, (b) 8 SDCCH/8, and (c) BCCH + CCCH 4 SDCCH/4. (From- [4].)
Note that the frame-stealing concept, used on the TCH to allocate FACCH frames, is not in use in the case of an SDCCH. This is due to the fact that sufficiently fast signaling can be transmitted over an SDCCH to carry on a handover procedure.

1.5.3.5 Mapping of the Broadcast Channels and CCCH on the 51 Multiframe
The broadcast channels and the CCCH (i.e., PCH, RACH, AGCH) are all multiplexed on the 51 multiframe, on the beacon carrier frequency. The FCCH is sent on downlink time slot 0, on the TDMA frames 0, 10, 20, 30, and 40 of the 51 multiframe. The SCH is also mapped on slot 0, in the TDMA frames immediately following an FCCH frame (i.e., frames 1, 11, 21, 31, 41).

The BCCH is associated with time slot 0 as well, in the TDMA frames that are not occupied by the FCCH and the SCH, and can also be mapped on time slots 2, 4, and 6 in some configurations.

The AGCH and PCH are dynamically multiplexed on the multiframe according to the network load. They are mapped on time slot 0, together with the FCCH, SCH, and BCCH, and optionally on time slots 2, 4, and 6 (also possibly used for the BCCH). Note that the BCCH and CCCH are multiplexed dynamically. The exact mapping is conveyed to the MS by means of SI blocks, sent over the BCCH.

One of these broadcast parameters also indicates whether or not the CCCH are combined with SDCCH and SACCH onto the same basic physical channel (the second type of mapping of SDCCH/SACCH presented in Section 1.5.3.4).

The mapping of the RACH channel is simple: every uplink slot corresponding to a downlink FCCH, SCH, BCCH, PCH, and AGCH can be used for a RACH. If the SDCCH and SACCH are multiplexed with the CCCH, the number of available random-access channel blocks are reduced. Figure 1.8(a, c) shows possible configurations with the multiplexing BCCH/ CCCH and BCCH/CCCH/SDCCH/SACCH.

1.5.3.6 Summary of Logical Channel Combinations
The different allowed combinations of logical channels on a physical channel are as follows:

TCH/F + FACCH/F + SACCH/TF;

TCH/H+ FACCH/H + SACCH/TH;

FCCH + SCH + BCCH + CCCH;

FCCH + SCH + BCCH + CCCH + SDCCH + SACCH;

BCCH + CCCH;

SDCCH+ SACCH.

Other logical channels exist for packet-switched services (GPRS, EDGE, or DTM), and will be described later.

1.5.4 Voice Digital Communication Chain
The functions performed by the physical layer during transmission are presented in this section, and the TCH/FS voice channel will be used for the sake of example. The steps of the communication chain, as outlined below, are presented in Figure 1.9.


Figure 1.9: The voice transmission chain.
1.5.4.1 Source Coding
First of all, speech must be digitized. On the basis of subjective speech quality and complexity issues, a regular pulse excited-linear predictive coder (RPE-LPC) with a long-term predictor loop was chosen. Basically, information from previous samples, which does not change very quickly, is used to predict the current sample. The coefficients of the linear combination of the previous samples, plus an encoded form of the residual (the difference between the predicted and actual sample), represent the signal. Speech is divided into 20-ms samples, each of which is encoded as 260 bits, giving a total bit rate of 13 Kbps. This is the so-called full-rate speech coding. Several other voice codecs are defined (half rate, enhanced full rate, adaptive multi-rate codecs), but these elements are beyond the scope of this book.

1.5.4.2 Channel Coding
In order to protect the voice against noise, interference, and multipath radio propagation conditions, the encoded speech transmitted over the radio interface is protected from errors. Convolutional encoding and block interleaving are used to achieve this protection. As discussed above, the speech codec produces a 260-bit block for every 20-ms speech sample. From subjective testing, it was found that some bits of this block were more important for perceived speech quality than others. The bits are therefore divided into three classes:

Class Ia: 50 bits, most sensitive to bit errors;

Class Ib: 132 bits, moderately sensitive to bit errors;

Class II: 78 bits, least sensitive to bit errors.

Class Ia bits have a 3-bit cyclic redundancy code added for error detection. If an error is detected, the frame is determined to be incomprehensible, and it is discarded. In such a case it is replaced by a slightly attenuated version of the previous correctly received frame.

These 53 bits, together with the 132 class Ib bits and a 4-bit tail sequence (a total of 189 bits), are input into a 1/2-rate convolutional encoder of constraint length 4. Each input bit is encoded as 2 output bits, based on a combination of the previous 4 input bits. The convolutional encoder thus outputs 378 bits. The 78 remaining class II bits, which are unprotected, are added to these 378 bits. A total of 456 encoded bits are therefore produced for every 20ms speech sample. This represents a bit rate of 22.8 Kbps. In order to protect against burst errors common to the radio interface, each block of 456 bits is interleaved.

1.5.4.3 Interleaving
The 456 coded bits are permutated, and divided into eight blocks of 57 bits, and these blocks are transmitted in eight consecutive bursts, as described in Figure 1.10. Since each burst can carry two 57-bit blocks, each burst carries traffic from two different speech frames. The benefit of doing this is that it provides time diversity. If a sequence of several consecutive bits is corrupted by the degraded propagation conditions (fading, for example) during a given period of time, interleaving ensures that the errors will be randomly distributed over the block of 456 bits. This property is required for a better performance of the decoding algorithm. The decoding of convolutional codes is indeed often performed by the Viterbi algorithm (refer to the case study on this subject in Chapter 4), for which the errors occur in packets. Thus, a random distribution of the corrupted bits over a block at the input of the decoder leads to an increased decoding efficiency, and so to better performance in terms of bit error rate.


Figure 1.10: Interleaving scheme on the TCH.
1.5.4.4 Burst Formatting
We have seen that the coded bits are interleaved and transmitted over bursts. The format of the burst-carrying TCH traffic, or normal burst (NB), is shown in Figure 1.11. NBs are used on most of the logical channels, except the RACH, SCH, and FCCH. The NB is constituted of two data blocks of 57 bits, carrying the coded voice samples, as described above. The middle of the burst contains the training sequence of 26 bits, known by the receiver, used to estimate the distortion introduced by the radio channel and for time synchronization. At the beginning and at the end of the burst, two sequences of 3 bits, known as the tail bits, all equal to 0, are used. Two bits, one before and one after the training sequence, form the stealing flags (SF). In Section 1.5.3.3 we saw that those bits allow the FACCH and TCH to be multiplexed, with the frame-stealing concept. When a speech frame is stolen for the transmission of a FACCH block, this is signaled by the SF bits. When one stealing bit is equal to one, this means that half of the burst is used for FACCH; otherwise it is used for TCH. An example of the multiplexing of FACCH and TCH is shown in Figure 1.12. Note that a FACCH is transported over eight half bursts. The burst duration is 148 bits, or 546.46 μs, and it is followed by a guard period of 8.25 bits, to allow for the burst ramping up and down between the time slots, as represented in Figure 1.11.


Figure 1.11: Structure of an NB.

Figure 1.12: Example of frame stealing- multiplexing of a TCH and an FACCH.
1.5.4.5 Ciphering
For security reasons, ciphering is used to modify the data parts of the burst, with a binary addition between a pseudorandom bit sequence and the 114 data bits. The same operation is performed by the receiver for deciphering. The pseudorandom bit sequence is different in the uplink and in the downlink.

1.5.4.6 Differential Encoding
All the bits di of the burst are differentially encoded. The output of the differential encoder is

(1.1)

where ⊕ denotes modulo 2 addition.

The result is mapped onto +1 and -1 values, to form the modulating data value αi input to the modulator, as follows:

(1.2)

1.5.4.7 GMSK Modulation
The digital signal is modulated onto the analog carrier frequency using Gaussian-filtered minimum shift keying (GMSK), with a symbol period of 48/13 μs (i.e., 270.8333 kHz). This modulation was selected over other modulation schemes as a compromise between spectral efficiency, complexity of the transmitter, power consumption for the MS, and limited out of channels emissions. These radio emissions, outside of the allocated channel, must be strictly controlled so as to limit adjacent channel interference and allow for the coexistence of GSM and the other systems. The spectrum due to the modulation mask requirement is presented in Figure 1.13, along with an ideal GMSK spectrum.


Figure 1.13: Ideal GMSK spectrum and required spectrum mask.
On the receive section, the opposite operations are performed, namely, demodulation, burst deformatting, de-interleaving, channel decoding, and source decoding.

1.5.4.8 Demodulation
Because of the various obstacles in the environment, many reflected signals, each with a different time delay and phase, arrives at the receiver. This is called multipath, and it is time variant, since the terminal is mobile, by definition. To cope with these time-varying propagation conditions, the receiver uses an equalizer. An equalizer is an algorithm that uses the receive sampled symbols to estimate the sequence of bits that was transmitted by the peer entity, by suppressing the intersymbol interference (ISI). Equalization is therefore used to extract the desired signal from the unwanted reflections. To do this, the receiver uses a known sequence of transmitted bits, the training sequence, to estimate the channel impulse response (CIR). This known signal is the 26-bit training sequence transmitted in the middle of every NB. The CIR is then used by the equalizer to retrieve the transmitted symbols. The estimation of the CIR also allows the finest time synchronization of the mobile to the BTS (the mobile can detect and correct a delay of several symbol periods). Note that the actual implementation of the equalizer is not specified in the GSM system, and it is up to the mobile phone or BTS vendor to implement a solution that will achieve the specified receiver performance (see Section 1.5.6.2).

1.5.5 Bursts Format
There are four burst formats the purposes of which are defined as follows:

The NB is used to carry information on traffic and control channels, except for RACH, SCH, and FCCH. It contains 114 encrypted symbols and includes a guard time of 8.25 symbol duration (=30,46 μs). A training sequence of 26 symbols is present in the middle of the burst (see Figure 1.11, Section 1.5.4).

The frequency correction burst (FB), as shown in Figure 1.14(a), contains a sequence of 142 fixed bits. This sequence is made of alternating ones and zeros (1, 0, 1, 0, ... 1, 0), so that after the differential encoding and GMSK modulation, the RF signal is equivalent to an unmodulated carrier, shifted by 67.7 kHz above the carrier frequency. This characteristic is used to help the mobile synchronize in frequency with the BTS, as explained in Section 1.5.7. The FB is transmitted over the FCCH.

The synchronization burst (SB) is needed for time synchronization of the mobile on the SCH. It contains a long training sequence and carries the information of the TDMA FN and base station identity code (BSIC), as can be seen in Figure 1.14(b).

The access burst (AB), presented in Figure 1.14(c), is used for random access (or the RACH) and is characterized by a longer guard period (68.25 bit duration or 252 μs), allowing the estimation of the timing advance (TA) by the BTS (see Section 1.5.6.3).


Figure 1.14: The (a) FB, (b) SB, and (c) AB burst structures.
In Figure 1.14, we see a guard period at the end of a burst. During this period, the transmission is attenuated in several steps, as specified by the power-versus-time mask specification (see the example of NB, Figure 1.11).

1.5.6 RF Characteristics
1.5.6.1 Transmission Characteristics
Several classes of mobiles are defined, according to their maximum output power capability, as shown in Table 1.3. In GSM-900, most of the mobiles available on the market are class 4 handheld terminals, while class 2 terminals are used as vehicle-mounted equipment. The class 4 and 5 MSs are denoted as "small MS." In DCS-1800, the typical class is class 1.

Table 1.3: MS Power Classes Power Class
GSM-400, GSM-900, GSM-850

Nominal Maximum Output Power
DCS-1800

Nominal Maximum Output Power
PCS-1900

Nominal Maximum Output Power

1
-
1W (30 dBm)
1W(30dBm)

2
8W (39 dBm)
0.25W (24 dBm)
0.25W (24 dBm)

3
5W (37 dBm)
4W (36 dBm)
2W (33 dBm)

4
2W (33 dBm)

5
0.8W (29 dBm)


These output power levels are maximum values, and can be reduced according to the commands that are sent by the network to the MSs. With these network commands, the MS operates at the lowest power level that maintains an acceptable signal quality.

These commands are based on the measurements that are performed by the MS and by the BTS. For instance, with a class 4 MS, the range of transmission can be several kilometers, but if the MS is getting closer to the BTS, it may receive a request from the network to decrease its output power level. This procedure, called power control, improves the performance of the system by reducing the interference caused to the other users. Moreover, it is a means of prolonging the battery life of the mobile. The power level can be stepped up or down in steps of 2 dB from the maximum power (depending on the MS class) down to a minimum of 5 dBm in GSM-400/900/850, and 0 dBm in DCS-1800/PCS-1900. The transmission of power control commands by the BTS is explained in Section 1.5.6.3.

For the BTS transceiver (TRX), the power classes are given in Table 1.4.

Table 1.4: TRX Power Classes TRX Power Class
GSM-400, GSM-900, GSM- 850

Maximum Output Power
DCS-1800 and PCS-1900

Maximum Output Power

1
320 (< 640)W
20(<40)W

2
160(<320)W
10(<20)W

3
80(<160)W
5(<10)W

4
40(<80)W
2.5(<5)W

5
20(<40)W

6
10(<20)W

7
5(<10)W

8
2.5(<5)W


As an option, the BSS can utilize downlink RF power control, with up to 15 steps of power control levels with a step size of 2 dB. Note that this power control on the downlink is not used on the beacon frequency, which is always transmitted with constant output power.

Many other requirements on the transmit section are defined in the GSM specifications, such as the spectrum due to modulation constraint (see Figure 1.13), the modulation accuracy, the transmitter frequency error, and the spurious emissions requirements.

1.5.6.2 Reception Characteristics
Several types of propagation models have been defined, in order to measure the mobile and BTS performances. These models represent several environments:

Typical urban (TUx);

Rural area (RAx);

Hilly terrain (HTx).

In the above definitions, the x stands for the velocity of the mobile, in km/h. The various propagation models are represented by a number of taps, each determined by their time delay and average power. The Rayleigh distributed amplitude of each tap varies according to a Doppler spectrum.

In addition to these multipath fading channels, the static channel was defined. This is a simple single-path constant channel. With this channel, the only perturbation comes from the receiver noise of the measured equipment.

One of the most important receiver performances that is specified is the sensitivity level, which determines the minimum level for which the receiver can demodulate a signal correctly. The sensitivity requirement, in GSM, is specified as an input level, in dBm, for which the measured equipment should reach a certain performance, in terms of bit error rate. For instance, for GSM400/900/850 power classes 4 or 5 mobiles and DCS-1800/PCS-1900 classes 1 or 2 mobiles, the sensitivity level is -102 dBm. For a normal BTS (that is not a micro- or a pico-BTS) the sensitivity level is -104 dBm, for all the frequency bands.

At these levels, different performances, according to both the logical channel and the propagation channel used for the measurement, must be met. Table 1.5 shows an example of performances that are reached at the sensitivity level, for GSM-900 and GSM-850. In this table, BER stands for bit error rate, FER for frame erasure ratio (i.e., incorrect-speech-frames ratio), and RBER for residual BER (defined as the ratio of the number of errors detected over the frames defined as "good" to the number of transmitted bits in the good frames). This table is an example; similar tables exist for the other logical channels and for the different frequency bands. Note that frequency hopping may be used for the sensitivity performance measurements.

Table 1.5: Sensitivity-Level Performance Requirements Propagation Conditions

Logical Channel
Static
TU

(no FH)
TU50

(ideal FH)
RA250

(no FH)
HT100

(no FH)

FACCH/H
(FER)
0.1%
6.9%
6.9%
5.7%
10.0%

FACCH/F
(FER)
0.1%
8.0%
3.8%
3.4%
6.3%

SDCCH
(FER)
0.1%
13%
8%
8%
12%

RACH
(FER)
0.5%
13%
13%
12%
13%

SCH
(FER)
1%
16%
16%
15%
16%

TCH/F14.4
(BER)
10.0-5
2.5%
2%
2%
5%

TCH/F9.6
(BER)
10.0-5
0.5%
0.4%
0.1%
0.7%

TCH/FS
(FER)
0.1α%
6α%
3α%
2α%
7α%

Class Ib
(RBER)
0.4/α%
0.4/α%
0.3/α%
0.2/α%
0.5/α%

Class II
(RBER)
2%
8%
8%
7%
9%


Note that in this example, the parameter a is defined as 1 ≤ α ≤ 1.6 and allows a tradeoff between the number of erased speech frames (i.e., decoded as wrong, and therefore not transmitted to the voice decoder) and the quality of the nonerased frames.

Another important characteristic of the receiver concerns its performance in the presence of an interferer. This is specified either for a cochannel interference (i.e., an interference situated at the same frequency as the signal of interest) or an adjacent channel interference (situated at 200 or 400 kHz from the carrier of interest). The level of the useful signal is set 20 dB higher than for the sensitivity evaluation, and a GMSK interfering signal is added, either at the same frequency or with an offset of 200 or 400 kHz from the carrier. For the cochannel test, the carrier to interference ratio C/Ic is set to 9 dB. Under these conditions, the performance of Table 1.6 must be met. Again, this table does not contain all the logical channels, and concerns the GSM-900 and GSM-850 only. Similar performance requirements are defined for the other cases.

Table 1.6: Interference Performance Requirements Propagation Conditions

Type of Channel
TU3

(NoFH)
TU3

(Ideal FH)
TU50

(No FH)
TU50

(Ideal FH)
RA250

(No FH)

FACCH/H
(FER)
22%
6.7%
6.7%
6.7%
5.7%

FACCH/F
(FER)
22%
3.4%
9.5%
3.4%
3.5%

SDCCH
(FER)
22%
9%
13%
9%
8%

RACH
(FER)
15%
15%
16%
16%
13%

SCH
(FER)
17%
17%
17%
17%
18%

TCH/F 14.4
BER)
10%
3%
4.5%
3%
3%

TCH/F 9.6
(BER)
8%
0.3%
0.8%
0.3%
0.2%

TCH/FS
(FER)
21α%
3α%
6α%
3α%
3α%

Class Ib
(RBER)
2/α%
0.2/α%
0.4/α%
0.2/α%
0.2/α%

Class II
(RBER)
4%
8%
8%
8%
8%


This table is also applied in the case of an adjacent channel interference. In this case, the C/I is set to -9 dB if the interferer is 200 kHz from the carrier, and -41 dB if it is 400 kHz from the carrier.

1.5.6.3 Control of the Radio Link
This section describes some of the procedures that are in use to improve the efficiency of the system, by adapting the transmission between the mobile and the BTS to the continuously varying radio environment.

Compensation for the Propagation Delay
Due to the distance between the MS and the BTS, there is a propagation delay that is equal to d/ c seconds, where d is the MS to BTS distance in meters, and c is the speed of light (c = 3 · 108 m.s-1). Without any compensation of this delay, the bursts transmitted by two different MSs, in the same TDMA frame on two consecutive slots, could interfere with one another.

Let us take the example of one MS situated 25 km away from the BTS, transmitting on time slot 0 of a given channel frequency. Another MS is located, say, 1 km away from the BTS, and transmitting on time slot 1 of that same frequency. The second MS transmission will experience a very short delay (around 3.33 μs), but the burst on time slot 0, from MS 1, will be received by the BTS 83.33 μs after it has been transmitted. This means that at the BTS receiver, the burst on time slot 0 will interfere with the beginning of the burst of time slot 1, for a period of about 80 μs. This example is represented in Figure 1.15.


Figure 1.15: Propagation delay difference between two MSs transmitting to the same BTS.
In order to cope with this problem, the network manages a parameter for each mobile called the TA. This parameter represents the transmission delay between the BTS and the MS, added to the delay for the return link.

The estimation of the delay is performed by the BTS upon reception of an AB on the RACH. As described in Section 1.5.5, this burst is characterized by a longer guard period (68.25-bit duration or 252 μs) to allow burst transmission from a mobile that does not know the TA at the first access. The received AB allows the BTS to estimate the delay by means of a correlation with the training sequence.

The TA value, between 0 and 63 symbol periods (i.e., between 0 and 232.615 μs by steps of 48/13 μs), is transmitted on the AGCH. It allows the MS to advance its time base, so that the burst received at the BTS arrives exactly three time slots after the BTS transmit burst, as shown in Figure 1.16. A distance of 35 km between the MS and the BTS is therefore possible. (The 232.675 μs allows to compensatefor a distance of around 70 km, including the forward and return links.)


Figure 1.16: Correction of MS transmission timing to compensate for propagation delay.
After this first propagation delay estimation, the BTS continuously monitors the delay of the NBs sent by from the MS on the other logical channels. If the delay changes by more than one symbol period, a new value of the TA is signaled to the MS on the SACCH.

MS Power Control
As explained in Section 1.5.6.1, the MS can vary its transmit output power from a maximum defined by its power class, by steps of 2 dB. During a communication, the MS and BTS measure the received signal strength and quality (based on the bit error ratio) and pass the information to the BSC, which ultimately decides if and when the power level should be changed. A command is then sent to the MS on the SACCH.

Power control is a difficult mechanism to implement, since there is a possibility of instability. This arises from having MS in cochannel cells, alternatively increasing their power in response to increased cochannel interference. Suppose that mobile A increases its power because the corresponding BTS receives a cochannel interference caused by mobile B, in another cell. Then the BTS receiving the signal from mobile B might request mobile B to increase its power, and so forth. This is the reason why some coordination is required at the BSC level.

Note that for an access request on the RACH, the MS uses the maximum power level defined by the parameter MS_TXPWR_MAX_CCH broadcast by the network.

Frequency Hopping
The radio environment depends on the radio frequency. In order to avoid important differences in the quality of the channels, a feature called slow frequency hopping (FH) was introduced. The slow FH changes the frequency with every TDMA frame, which also has the effect of reducing the cochannel interference. This capability is optionally used by the operator, and is not necessarily implemented in all the cells of the network, but it must be supported by all the MSs. The main advantage of FH is to provide diversity on one transmission link (especially to increase the efficiency of coding and interleaving for slowly moving MSs) and also to average the quality on all the communications through interference diversity.

The principle of slow FH is that every mobile transmits its time slots according to a sequence of frequencies that it derives from an algorithm. The FH sequences are orthogonal inside one cell (i.e., no collisions occur between communications of the same cell) and independent from one cell to a cochannel cell (i.e., a cell using the same set of RF channels or cell allocation). The hopping sequence is derived by the mobile from parameters broadcast at the channel assignment, namely, the mobile allocation (set of N frequencies on which to hop), the hopping sequence number (HSN) of the cell (which allows different sequences on cochannel cells), and the index offset (to distinguish the different mobiles of the cell using the same mobile allocation) or mobile allocation index offset (MAIO). Based on these parameters and on the FN, the MS knows which frequency to hop in each TDMA frame.

It must be noted that the basic physical channel supporting the BCCH does not hop.

1.5.7 MS Cell Synchronization Procedure
In synchronizing to a cell, the MS first searches for the FB on the FACCH. This allows a first timing synchronization, but most of all, it allows the mobile to adjust its oscillator to be synchronized in the frequency domain with the BTS. This is possible because, as described in Section 1.5.6, the fixed sequence of the FB has been chosen so that the modulating bit sequence at the GMSK modulator input is constant. This results in a continuous π/2 phase rotation, which in the frequency domain is equivalent to an unmodulated carrier with a +1 625/24 kHz frequency offset, above the nominal carrier frequency. Once the MS has identified the FB, it uses this property to estimate its frequency drift with regard to the BTS.

In the TDMA frame immediately following the occurrence of an FB, on the same carrier frequency, the SB is transmitted over the air interface, on the SCH. This burst is identified by the MS with its extended 64-bit training sequence code. The MS received samples of the SB, correlated with the known training sequence, allow for the timing of the mobile to be adjusted to the base station with good precision. At this point, the MS and the BTS are synchronized in the time domain, except that the propagation delay between them is not compensated. This is performed with the TA scheme, as discussed in Section 1.5.6.3, when the MS sends an AB on the RACH.

The decoding of the SB enables a logical synchronization of the mobile to the cell, since it gives the elements to estimate the TDMA FN (see Section 1.5.3.1), which allows the MS to determine the position of the SCH in the hyperframe. The SCH also contains the BTS identity code (BSIC): these 6 bits (before channel coding) consist of the PLMN color code with range 0 to 7 and of the BS color code with range 0 to 7 (3 bits each).

1.5.8 Summary of MS Operations in Idle Mode
In idle mode, as opposed to dedicated mode, the mobile has no channel of its own. It is required to

Synchronize in time and frequency to a given cell, selected as the best suitable cell with regard to a set of criteria (based on the beacon received power at the MS). This is termed "camping onto" a cell. This process of evaluating different cells and choosing the best suitable one is called cell selection, or reselection if it is performed again, due to the degradation of the link quality with the previously selected cell. The MS during idle mode continuously measures the radio link quality of the serving cell and the surrounding cells, so that cell reselection criteria are evaluated periodically.

Listen to possible incoming calls from the network. The notification of an incoming call is usually referred to as paging.

1.5.8.1 Selection of the PLMN
When the mobile is switched on, the first operation that it performs is identification or selection of a PLMN. Most of the time, the PLMN will be the home PLMN (i.e., the network to which the user has subscribed). In such a case, no selection is needed, as information about the network is stored in the SIM card. If it is not the case, because the user is traveling in a different area, the MS will scan all the frequencies in order to detect the surrounding beacon channels (detection of FB and SB, as described in Section 1.5.7). The MS is then able to decode the PLMN identifiers, and either choose the first PLMN in the priority ordered list of the SIM card, or ask the user which PLMN is preferred among all the detected PLMNs. The selection is then stored, in order to be used at the next terminal switch on. In any case, the user can explicitly ask for a given PLMN selection.

1.5.8.2 Principles of Cell Selection and Reselection
Once the PLMN is selected, the MS must select a cell. Two scenarios are possible:

The beacon channel frequencies are stored in the MS, because it has already performed a selection in the previous terminal activity. In this case, the MS will perform measurement on these frequencies, to determine which cell is the most suitable with regard to certain criteria. Once the best cell is selected, the MS performs registration and "camps on" this cell. Note that if the stored frequency list of beacon carriers is not detected by the MS, it will perform a PLMN selection as described above.

It is the first time the PLMN is accessed. The carriers of the system are all scanned, in order to detect the beacon channels, and the received signal strength of these channels is added in an ordered list. Once this is achieved, the cell selection can be performed, as in the previous case. In order to speed up the process, a list of the RF channels containing BCCH carriers of the same PLMN is broadcast in the system information messages.

When an MS is camping on a cell, it can receive paging blocks on the PCH, or initiate call setup for outgoing calls by sending an AB on the RACH. It still regularly monitors the signal level on the surrounding beacon carriers, and evaluates the reselection criteria. The reselection is triggered if one of the following events occurs:

The path loss criterion parameter C1 indicates that the path loss to the cell has become too high.

There is a downlink signaling failure (i.e., the success rate of the MS in decoding signaling blocks drops too low).

The cell camped on has become barred (this means that the operator decides not to allow MSs to camp on this cell).

There is a better cell, in terms of the path loss criterion C2 in the same registration area.

A random access attempt is still unsuccessful after a given number of repetitions, specified by a broadcast parameter.

The criteria for cell selection and reselection (path loss criterion C1 and reselection criterion C2) are based on the measurements performed by the MS on the BCCH frequency. (As stated earlier, the beacon frequency is transmitted with its maximum output power by the BTS.) Details on these criteria for cell selection are given in Chapter 5.

1.5.8.3 Monitoring of Paging Blocks
We discussed the mapping of the CCCH on the 51 multiframe in Section 1.5.3.5, and in particular the case of the PCH. This logical channel is used to convey paging blocks on the downlink. These blocks are used to notify the MS of an incoming call. In order to conserve the MS's power, a PCH is divided into subchannels, each corresponding to a group of MSs. Each MS will then only "listen" to its subchannel and will stay in the sleep mode during the other subchannels of the PCH. This is called the discontinuous reception (DRX) mode. The mobile knows in which group it belongs by determining the parameter CCCH_GROUP. It is estimated with an algorithm, which inputs are the mobile IMSI and the parameter BS_CC_CHANS, broadcast on the BCCH. This parameter defines the number of basic physical channels supporting CCCH.

Mobiles in a specific CCCH_GROUP will listen for paging messages and make random accesses only on the specific CCCH to which the CCCH_GROUP belongs. This algorithm is detailed in Chapter 4. Note that the MS is not authorized to use the DRX mode of operation while performing the cell-selection algorithm.

1.5.9 Measurements Performed by MS During Communication
When assigned a TCH or SDCCH, during the time slots that are not used for these channels and for the associated SACCH, the MS performs measurements on all the adjacent BCCH frequencies. These measurements are then sent to the network by means of the SACCH, and are interpreted by the NSS for the power control and handover procedures. Measurements are performed in each TDMA frame, (see Figure 1.17) and are referred to as monitoring, which consists of estimating the receive signal strength on a given frequency. The list of frequencies to be monitored is broadcast on the BCCH, by means of the BCCH allocation (BA) list, which contains up to 32 frequencies. The frequencies are monitored one after the other, and the measured samples are averaged prior to the reporting to the network, on an uplink SACCH block, under form of a value called RXLEV. The MS therefore measures the received signal level from surrounding cells by tuning and listening to their BCCH carriers. This can be achieved without interbase station synchronization. The measurements are reported at every reporting period.


Figure 1.17: Monitoring during a TDMA frame.
For a TCH/FS, the reporting period duration is 104 TDMA frames (480 ms).

It is essential that the MS identify which surrounding BSS is being measured in order to ensure reliable handover. Because of frequency reuse with small cluster sizes, the BCCH carrier frequency may not be sufficient to uniquely identify a surrounding cell. The cell in which the MS is situated may have more than one surrounding cell using the same BCCH frequency. It is therefore necessary for the MS to synchronize to (using the method explained in Section 1.5.7) and demodulate surrounding BCCH carriers to identify the BSIC in the SB. In order to do so, the MS uses the idle frames. These frames are termed "search" frames. Note that a window of nine consecutive slots is needed to find time slot 0 on the BCCH frequency (remember that time slot 0 carries the SCH and FCCH), since the beacon channels are not necessarily synchronized with one another. One important characteristic to notice is that the SCH and FCCH are mapped onto the 51 multi-frame, and that the idle frame of the mobile during communication is occurs in on the 26 multiframe. Since 26 and 51 are mutually prime numbers, this means a search frame will be available every 26 modulo 51 frame on the beacon channel.

For instance, let us imagine that an idle frame occurs in the frame 0 of the 51 multiframe. The next idle frames will be programmed on frames 26, 1, 27, 2, and so on. Therefore, after a certain number of search frames, the MS will necessarily decode an FB and an SB.

Another measured parameter during a TCH or SDCCH is the RXQUAL, which represents an indication of the quality of the received link, in terms of BER. For each channel, the measured received signal quality is averaged on that channel over the reporting period of length one SACCH multiframe defined above.