Module 5 - GSM Air Interface & Network Planning

GSM Air Interface and Network Planning
GSM Air Interface and

Network Planning
Contents
1 Module Objectives.....................................................................................2
2 Transmission............................................................................................. 3
2.1 Introduction to Radio and Terrestrial Transmission.................................... 3
2.2 Transmission through the Air Interface....................................................... 9
gigijeco.und.slo
2.3 Problems and Solutions of the Air Interface............................................. 21

2.4 Terrestrial Transmission............................................................................25
2.5 Summary of the Learning Points .............................................................. 31
3 Network Planning and Optimisation..................................................... 32
3.1 Introduction................................................................................................32
3.2 Switching Network Planning......................................................................34
3.3 Cellular Transmission Network Planning.................................................. 34
3.4 Radio Network Planning............................................................................35
3.5 Optimising and Developing the Network ...................................................40
3.6 Summary of the Learning Points .............................................................. 42
Nokia Siemens Networks 2008 1-1

1 Module Objectives
At the end of this module, the student should be able to:
Explain the difference between physical and logical channels.
List and describe at least nine different types of logical channels and their
functions with the help of their abbreviations.
Name two problems in the air interface and suggest a way of decreasing
each of these problems.
Describe the main function of the transcoder.
List three Base Station Controller (BSC)/Base Transceiver Station (BTS)
connections.
List five steps in the network planning process.
Explain why and how frequencies are reused in the GSM network.
Name at least one advantage with a sectorised Base Transceiver Station
compared with an omnidirectional BTS.
Name three sources of information that can be used when monitoring the
network's performance.
gigijeco.und.slo
1-2 Nokia Siemens Networks 2008

2 Transmission
2.1 Introduction to Radio and Terrestrial
Transmission
The radio link is the most vulnerable part of the connection and a great deal of
work is needed to ensure its high quality and reliable operation.
When designing a radio interface solution of a modern wireless communication
system, following questions have to be answered:
How can we support duplex transmission, i.e. how can it be possible that a
subscriber has the impression, that he or she can transmit and receive user
data simultaneously?
How can we support multiple access, so that several users within one
geographical area can use the mobile communication system at the same
time?
2.1.1 Duplex Transmission

In GSM, Frequency Division Duplex (FDD) is used for duplex transmission. In
FDD, frequency resources are allocated to the mobile communication system.
gigijeco.und.slo
Duplex transmission is enabled in GSM by using different frequency bands.

Therefore, uplink and downlink communications are separated by frequency.
The frequency ranges of GSM 900 and GSM 1800 are indicated below:
Fig. 1 Frequency allocations for GSM

Note that the uplink refers to a signal flow from the Mobile Station (MS) to the
Base Transceiver Station (BTS) and the downlink refers to a signal flow from
the Base Transceiver Station (BTS) to the Mobile Station (MS) . The
simultaneous use of separate uplink and downlink frequencies enables
communication in both the transmit (TX) and the receive (RX) directions. The
radio carrier frequencies are arranged in pairs and the difference between these
two frequencies (uplink-downlink) is called the duplex frequency. Hence, it can
be seen that for GSM900, duplex frequency is 45 MHz whereas for GSM1800,
duplex frequency is 95MHz.
2.1.2 Multiple Access

Imagine that one GSM 900 subscriber gets 25 MHz for both downlink and uplink
communication. Then there would be no resources left for other GSM 900 users
in the same geographical area. One user would block all available radio interface
resources. Luckily, this is not the case in GSM. It can support multiple users on
same radio interface and this is known as Multiple Access.
Two principles are applied to allow multiple access in GSM:
FDMA: Frequency Division Multiple Access, and
TDMA: Time Division Multiple Access.
Multiple Access is achieved in FDMA by dividing the UL and DL frequency bands
into sub-frequency bands. Each active subscriber thus gets one downlink and one
gigijeco.und.slo
uplink sub-frequency band allocated. The table below shows how this is done in
GSM. The uplink and downlink resources are organised in 200 kHz
sub-frequency bands. The sub-frequency bands are called carriers. In GSM 900,
the 25 MHz bands for uplink and downlink communication are divided into 124
carriers. (Two times 100 kHz are used as guard bands on the left and right hand
side of the UL and DL frequency ranges.) Channel numbers also known as
Absolute Radio Frequency Channel Number (ARFCN), identify the uplink and
downlink carriers. If a subscribers call is conducted on Channel 2, then uplink is
performed on the frequency band 890.3-890.5 MHz and downlink on the
frequency band 935.3-935.5 MHz. In GSM 900, as mentioned before, the duplex
frequency is 45 MHz.

Table 1. GSM 900 frequencies
Channel Uplink signal (MHz) Downlink signal (MHz)

1 890.1 890.3 935.1 935.3
(890.2 -centre freq.) (935.2 -centre freq.)
2 890.4 (centre freq.) 935.4 (centre freq.)

... ... ...
In GSM 1800 the number of carriers is 374 and the duplex frequency is 95 MHz.
The devices in the Base Transceiver Station (BTS) that transmit and receive the
radio signals in each of the GSM channels (uplink and downlink together) are
known as Transceivers (TRX). (Transmitter+Receiver=Transceiver).
FDMA is combined with TDMA to increase the number of simultaneous users of
the radio interface. Time Division Multiple Access (TDMA), as the name suggests,
is a method of sharing a resource (in this case a radio frequency) between
gigijeco.und.slo
multiple users, by allocating a specific time (known as a time slot) for each user.
This is in contrast to the analogue mobile systems where one radio frequency is
used by a single user for the duration of the conversation. In Time Division
Multiple Access (TDMA) systems each user either receives or transmits bursts of
information only in the allocated time slot. These time slots are allocated for
speech only when a user has set up the call. Some timeslots are, however, used
to provide signalling and location updates etc. between calls.

Fig. 2 Time Division Multiple Access Principle
2.1.3 GSM Frequency Bands

In Europe, Australia, most parts of Asia and South America, GSM 900 and GSM
1800 are available. In North America, some parts of Asia and South America,
there is GSM 1900. Numbers, such as 900 and 1900, indicate the frequency
bands applied. All the specified bands are listed below. GSM 900, 1800 and 1900
gigijeco.und.slo
are the most commonly used frequency bands.
GSM 450
UL : 450,4 457,6 MHz and DL : 460,4 467,6 MHz
GSM 480
UL : 478,8 486 MHz and DL : 488,8 496 MHz
GSM 700
UL : 747 762 MHz and DL : 777 792 MHz
GSM 850
UL : 824 849 MHz and DL : 869 894 MHz
GSM 900 (standard GSM, P-GSM, Primary GSM)
UL : 890 915 MHz and DL : 935 960 MHz
Extended GSM 900 (E-GSM; only in combination with GSM 900)
UL : 880 915 MHz and DL : 925 960 MHz
Railway GSM 900 (R-GSM; in combination with GSM)
UL : 876 980 MHz and DL : 921 925 MHz
GSM 1800 (DCS 1800)
UL : 1710 1785 MHz and DL : 1805 1880 MHz
GSM 1900 (PCS 1900)
UL : 1850 1910MHz and DL : 1930 1990 MHz

2.1.4 Modulation
GSM uses digital techniques where the speech and control information are
represented by 0s and 1s. How is it possible to transmit digital information over
an analogue radio interface?
The digital values 0 and 1 are used to change one of the characteristics of an
analogue radio signal in a predetermined way. By altering the characteristic of a
radio signal for every bit in the digital signal, we can "translate" an analogue
signal into a bit stream in the frequency domain. This technique is called
modulation. Analogue signals have three basic properties: amplitude, frequency
and phase. Therefore, there are basically three types of modulation processes in
common use:
amplitude modulation
frequency modulation
phase modulation
gigijeco.und.slo
Fig. 3 Examples of frequency and amplitude modulation
The transmitter sends a bit value, which is represented by an analogue

electro-magnetic wave. By changing the characteristics of the electro-magnetic
wave, a new bit value is represented. The number of changes during one second
gives the data rate. This also defines the duration of one bit. In frequency
modulation, after the transmission of the first bit, the transmitter changes the
frequency in accordance to the new bit value.
In most modern mobile communication systems, Phase Shift Keying (PSK) is
applied. Binary phase shift keying (BPSK) is sketched with the figure below. As
can be seen, a new bit value is represented either by a zero degree phase shift
or 180 degrees phase shift. The amplitude and the frequency are not changing.

Fig. 4 Binary Phase Shift Keying (BPSK)
GSM uses a phase modulation technique over the air interface known as
Gaussian Minimum Shift Keying (GMSK). In order to understand how it works, let
us take a simple example. In this case, a new bit value is represented by a plus
90 degrees or 90 degrees phase shift. In contrast to BPSK, the phase shift does
not happen abruptly in the MS. The phase shift is done gradually over the bit
period. (By doing so, the MS causes less inter-frequency interference.)
In GMSK, phase transitions are smoothed by filtering data with a Gaussian curve.
This enables smooth phase shifts, keeping the bandwidth comparably narrow.
Thus, a bandwidth of only 200 kHz can be achieved. The duration of one bit is
gigijeco.und.slo
3.69 s in GSM.
Fig. 5 Example of phase modulation
The radio air interface has to cope with many problems, such as variable signal
strength due to the presence of obstacles along the way, radio frequencies
reflecting from buildings, mountains etc. with different relative time delays and
interference from other radio sources.
With such levels of interference, complex equalisation techniques are required
with GMSK.

2.2 Transmission through the Air Interface

To enable us to understand the principles of the air interface, let us imagine that
an army has to be moved from one place to another, and a convoy of vehicles is
set aside to do the job. The army consists of soldiers and officers.
Each vehicle has eight seats and, therefore, only eight people can be carried in
each vehicle.
Fig. 6 A small logistical problem
Obviously, the only solution is to divide the army into groups of eight people. One
officer and seven soldiers are allocated to each vehicle. The officer sits in the
front seat and seven soldiers sit in the others.
gigijeco.und.slo
There are different types of people in the army, soldiers and officers. These could
be referred to as " logical" differences, as they are all human beings, but their
functions are different. In addition, there can be many different ranks of officers,
each one with different responsibilities.
To move them from one place to another, a " physical" connection is employed,
that is, the vehicles and the seats.
2.2.1 Physical and Logical Channels

Time Division Multiple Access (TDMA) divides one radio frequency channel into
consecutive periods of time, each one called a " TDMA frame". The TDMA frame
can be compared to the vehicle in our example.
Each TDMA frame contains eight shorter periods of time known as " timeslots".
These timeslots can be compared to the seats in the vehicle. The TDMA
timeslots are called " physical channels ", as they are used to physically move
information from one place to another.
The radio carrier signal between the mobile station and the BTS is divided into a
continuous stream of timeslots, which in turn are transmitted in a continuous
stream of TDMA frames - like a long line of vehicles with eight seats in each.

If the time slots of the TDMA frame represent the physical channels, what about
the contents? The contents of the physical channels that is, the soldiers and
officers travelling in the eight seats of the vehicle, according to their roles, are
called " logical channels ". In the example of the army, the soldiers are one type
of logical channel and the officers are other types of logical channels, and they
exercise some kind of control depending on their responsibilities.
In GSM, the logical channels can be divided into two types:
Dedicated channels
Common channels
Let us look at things in a more practical way: A subscriber switches on his mobile
phone and receives a call. This simple act of switching on the phone involves the
following steps:
1. The mobile scans all the radio frequencies and measures them.
2. It selects the frequency with the best quality and tunes to it.
3. With the help of a synchronisation signal in a TDMA frame, the mobile
synchronises itself to the network.
gigijeco.und.slo
Fig. 7 Tuning into the network
The synchronisation information required by this process is broadcast by the

network and analysed by the mobile.
Registration and authentication are the next steps and they consist of the
following operations:
1. A point to point connection must be set up. The mobile station makes a
request for a channel to establish the connection.
2. The network acknowledges the request and allocates a channel. The mobile
receives and reads this information.
3. The mobile then moves to the allocated (dedicated) channel for further
transactions with the network and continues authentication/registration
procedures with VLR and HLR on that channel.

Fig. 8 Initiation of a call
Once the subscriber is registered in the network and the authentication is

successful, calls can be set up. In the case of a mobile terminated call, the
subscriber has to be paged. This process is described below:
1. The network sends a paging message to all Base Transceiver Stations (BTS)
within the Location Area (LA) where the subscriber is registered.
2. The mobile station answers the paging message by sending a
service/channel request.
3. The network acknowledges this request and again an authentication is
gigijeco.und.slo
needed. A dedicated signalling channel is assigned in order to transmit the

data related to the call.
4. A traffic channel is assigned for the conversation.
During the conversation, the mobile measures the signal strength of adjacent
carriers and sends measurement reports to the Base Station Controller (BSC).
A channel must be dedicated for this function also.
Fig. 9 Call completion from the called side

This is a simplified description of the process, but it conveys the idea that there
are many functions involved in the air interface to enable a mobile user to have
conversation. Each one of these functions requires a separate "logical channel",
as the data contents are different. Some of them are uplink , others are downlink
and some are bi-directional.
2.2.2 Logical Channels

There are twelve different types of logical channels, which are mapped into
physical channels in the radio path. Logical channels comprise of common
channels and dedicated channels . Common channels are those that are used
for broadcasting different information to mobile stations and for setting up
signalling channels between the MSC/VLR and the mobile station.
gigijeco.und.slo
Fig. 10 Logical channels
Over the radio path, different types of signalling channels are used to facilitate
the discussions between the mobile station and the BTS, BSC and the MSC/VLR.
All these signalling channels are called dedicated control channels .
Traffic channels are also dedicated channels, as each channel is dedicated to
only one user to carry speech or data.

Fig. 11 TDMA frames with common and dedicated channels
As seen in the figure above, different logical channels are placed in different
timeslots depending on whether they are common channels used by several
mobile stations (blue) or if they are dedicated to a certain mobile station in
connection with a call (yellow).
The only dedicated channel among the blue ones above is the SDCCH. It is used
for system signalling during idle periods and also for call set-up before a traffic
gigijeco.und.slo
channel has been allocated (in the example above located in one of the timeslots
1-7).
Furthermore, the common and dedicated channels are grouped in different
multiframes. The common channels are grouped in a 51 TDMA frame order,
and the dedicated channels are grouped in a 26 TDMA frame order. Why this
organisation? The reason is that we need to be able to receive and decode
common channels along with (but independently of) the dedicated channels used
in busy mode.
By multiplying 51 with 26, we can conclude that any TDMA frame number will
occur simultaneously in both multiframes every 1326 TDMA frames (which
corresponds to 6.12 seconds). Such cycles are called superframes.
Let us now take a closer look at the different logical channels and their functions.
2.2.2.1 Broadcast channels

Base stations can use several TRXs, but there is always only one TRX that can
carry common channels. Broadcast channels are downlink point-to-multipoint
channels. They contain general information about the network and the
broadcasting cell. There are three types of broadcast channels:

1. Frequency Correction Channel (FCCH)

FCCH bursts consist of all "0"s that are transmitted as a pure sine wave. This
acts like a flag for the mobile stations and enables them to find the TRX
among several TRXs, which contains the broadcast transmission. The MS
scans for this signal after it has been switched on, since it has no information
as to which frequency to use.
2. Synchronisation Channel (SCH)
The SCH contains the Base Station Identity Code (BSIC) and a reduced
TDMA frame number. The BSIC is needed to identify that the frequency
strength being measured by the mobile station is coming from a particular
base station. In some cases, a distant base station broadcasting the same
frequency can also be detected by the mobile station. The TDMA frame
number is required for speech encryption.
3. Broadcast Control Channel (BCCH)
The BCCH contains detailed network and cell specific information such as:
Frequencies: Used in the particular cell and neighbouring cells.

Frequency hopping sequence: Is designed to reduce the negative
effects of the air interface, which sometimes results in the loss of
transmitted information; the mobile station may transmit information on
different frequencies within one cell. The order in which the mobile
station should change the frequencies is called the "frequency hopping
gigijeco.und.slo
sequence". ( However, implementing frequency hopping in a cell is
optional.)
Channel combination: As mentioned earlier, there are a total of twelve
logical channels. All logical channels except traffic channels are mapped
into timeslot 0 or timeslot 1 of the broadcasting TRX. The channel
combination informs the mobile station about the mapping method used
in the particular cell.
Paging groups: Normally there is more than one paging channel in one
cell (describer later). To prevent a mobile from listening to all the paging
channels for a paging message, the paging channels are divided in such
a way that only a group of mobile stations listens to a particular paging
channel. These are referred to as paging groups.
Information on surrounding cells : A mobile station has to know what
the cells surrounding the present cell are and what frequencies are being
broadcast on them. This is necessary if, for example, the user initiates a
conversation in the current cell, and then decides to move on. The
mobile station has to measure the signal strength and quality of the
surrounding cells and report this information to the base station
controller.
All the above mentioned information is sent over as SYSTEM INFO messages on
BCCH timeslot.

2.2.2.2 Common control channels

Common control channels comprise the second set of logical channels. They are
used to set up a point to point connection. There are three types of common
control channels:
1. Paging Channel (PCH)
The PCH is a downlink channel that is broadcast by all BTSs of a location
area in the case of a mobile terminated call.
2. Random Access Channel (RACH)
The RACH is the only uplink and the first point to point channel in the
common control channels. It is used by the mobile station in order to initiate
a transaction, or as a response to a PCH.
3. Random Access Channel (RACH)
4. Access Grant Channel
The AGCH is the answer to the RACH. It is used to assign a mobile a
Stand-alone Dedicated Control Channel (SDCCH). It is a downlink, point to
point channel.
2.2.2.3 Dedicated control channels

Dedicated control channels compose the third group of channels. Once again,
there are three dedicated channels. They are used for call set-up, sending
gigijeco.und.slo
measurement reports and handover. They are all bi-directional and point to
point channels. There are three dedicated control channels:
1. Stand Alone Dedicated Control Channel (SDCCH)
The SDCCH is used for system signalling: call set-up, authentication, location
update, assignment of traffic channels and transmission of short messages.
2. Slow Associated Control Channel (SACCH)
An SACCH is associated with each SDCCH and Traffic Channel (TCH). It
transmits measurement reports and is also used for power control, time
alignment and in some cases for transmitting short messages.
3. Fast Associated Control Channel (FACCH)
The FACCH is used primarily during handovers. It is mapped onto a TCH
and it replaces 20 ms of speech and, therefore, it is said to work in "stealing"
mode. However, the user hardly gets to know the occurrence of FACCH
during his call. That makes the mobility hidden from end user, which is a
basic requirement of all wireless networks.
2.2.2.4 Traffic Channels (TCH)

Traffic channels are logical channels that transfer user speech or data, which can
be either in the form of half rate traffic (5.6 Kbits/s) or full rate traffic (13
Kbits/s). Another form of traffic channel is the Enhanced Full Rate (EFR) traffic
channel (12.2 Kbits/s). The speech coding in EFR is still done at 13 Kbits/s, but

the coding mechanism is different than that used for normal full rate traffic. EFR
coding gives better speech quality at the same bit rate than normal full rate.
Traffic channels can transmit both speech and data and are bi-directional
channels.
Adaptive Multi Rate (AMR) is the fourth speech codec defined for the GSM
system. The AMR codec combines the benefits of the EFR and HR codecs in
order to achieve an improved standard of voice quality and greater capacity. AMR
generates improved speech quality in both half-rate and full-rate modes by
varying the balance between speech and channel coding for the same gross
bit-rate. This process, known as codec mode adaptation, results in improved
voice quality throughout the cell and increases overall coverage.
AMR achieves this goal by dynamically adapting its bit-rate allocation between
speech and channel coding, thereby optimising speech quality in various radio
channel conditions. Depending on the conditions, AMR dynamically uses either
the GSM full rate traffic channel with a gross bit rate of 22.8 kbps or the GSM
half rate traffic channel with a gross bit rate of 11.4 kbps. A part of this bit rate is
used for speech coded bits and a part for error control.
2.2.3 Time Slots and Bursts

We have already seen that the technique used in the air interface is Time
Division Multiple Access (TDMA) where one frequency is shared by, at the most,
eight users. Consider the example of a 2 Mbit/s PCM signal that can carry 30
gigijeco.und.slo
speech channels, with each channel occupying 64 Kbits/s. The speech signals
from the mobile stations must be placed into a 2 Mbit/s signal that connects the
BTS and the BSC.
It is very important that all mobile stations in the same cell send the digital
information at the correct time to enable BTS to place this information into the
correct position in the 2 Mbit/s signal.
How do we manage the timing between multiple mobile stations in one cell? The
aim is that each mobile sends its information at a precise time, so that when the
information arrives at the Base Transceiver Station, it fits into the allocated time
slot in the 2 Mbit/s signal. Each mobile station must send a burst (a burst
occupies one TDMA timeslot) of data at a different time to all the other mobile
stations in the same cell. The mobile then falls silent for the next seven timeslots
and then sends the next burst and so on.
It can be seen that the mobile station is sending information periodically. All the
mobile stations send their information like this. If we go back to the analogy of the
army, the road is the radio carrier frequency, the vehicle is the TDMA frame and
the seats in each vehicle are the TDMA timeslots.

Fig. 12 TDMA bursts and timeslots

gigijeco.und.slo

In the air interface, a TDMA timeslot is a time interval of approximately 576.9 s,

which corresponds to the duration of 156.25 bit times. All bursts occupy this
period of time, but the actual arrangement of bits in the burst depends on the
burst type. Two examples are:
Normal burst, used for traffic channels, stand alone dedicated channels,
broadcast control channel, paging channel, access grant channel, and slow
and fast associated control channels.
Access burst, used to send information on the Random Access Channel
(RACH). This burst contains lowest number of bits. The purpose of this extra
free space is to measure the distance between the Mobile Station and the
BTS at the beginning of a connection. This process determines a parameter
called " timing advance" which ensures that the bursts from different mobile
stations arrive at the correct time, even if the distances between the various
MSs and BTS are different. This process is carried out in connection with the
first access request and after a handover. A maximum theoretical distance of
about 35 kilometres is allowed between the BTS and the MS.
gigijeco.und.slo
Fig. 13 Normal bursts and access bursts

As you can see, the normal burst contains a training sequence and an SB
(Stealing Bit). The training sequence is known, which makes it possible to correct
bit errors by applying Viterbi Equalisation. There are eight different sequence
patterns, and the one to be used is determined at the time of call set-up. The
stealing bits indicate if the burst contains FACCH information.
The Frequency Correction Burst is used by the MS to synchronise on a
cell. In each cell, the Frequency Correction Burst is transmitted on the same
carrier as the BCCH. It is transmitted every 10th or every 11th TDMA frame
(see 51 multiframe) with a higher output power than every other burst. This
helps the MS to find to determine the 51 multiframe structure as well as the
position of the SCH and BCCH.
The Frequency Correction Bursts transmits a sinusoid wave, which is used to
synchronise the local oscillator of the MS to the BCCH carrier frequency
band.
The Synchronisation Burst is equipped with a very long, well known
training sequence. This long sequence helps the MS to get time
synchronised downlink. If the MS is time synchronised downlink, it knows
when BCCH and PCH information is transmitted and it can read it. The
Synchronisation Burst also transmits the cells BSIC.
On the BCCH carrier of a cell, there must be permanently a high output
power. In timeslot 0, there is 51 multiframe with FCCH, SCH, and BCCH. If
timeslot 1 to 7 are not in use, or resources within the 51 multiframe of TS 0
are not used (e.g. SDCCH resources), then the dummy burst is transmitted
gigijeco.und.slo
to guarantee, that there is permanently a load on this carrier.
Fig. 14 Frequency correction, synchronisation, and dummy bursts2.3.4

Multiframe: Channel Combinations

There are seven different schemes to co-ordinate the logical channels in
multiframes. Three schemes are used for co-ordination of Full rate and Half rate
Traffic Channels. Four schemes are used to co-ordinate signaling, depending on
the requirements of the individual cell. The network operator has to decide, which
channel combinations are used for a cell.
Combination I III are used for TCH Multiframe co-ordination (Full rate / Half
rate).
Combination IV VII are used for Signaling Multiframe co-ordination.
Combination I : TCH/F + FACCH/F + SACCH/F
Combination I is used to transmit Full rate user data and speech. The frames
011 and 13-24 are used for user data, frame 12 is used for SACCH (user
related control data) and frame 25 is not used (I: Idle).
Combination II & III : TCH(0,1) + FACCH/H(0,1) + SACCH/H(0,1) respectively
TCH/H(0) + FACCH/H(0) + SACCH/H(0) + TCH/H(1) + FACCH/H(1) +
SACCH/H(1)
Combination II & III are used to transmit Half rate user data and speech. 2 TCH/H
users have to share the 26 multiframes. Data from user 1 or user 2 are filled
alternately into the frames. The SACCH of user 1 is on frame 12, the SACCH of
user 2 is on frame 25.
Combination IV : FCCH +SCH + CCCH (PCH & AGCH) + BCCH
gigijeco.und.slo
Combination IV offers much space for the Common Control Channels CCCH.
Therefore, this combination is used often for cells with many carrier. BCCH
carrier is the cell beacon and so it must be used exactly only on one carrier of the
cell.It is allocated on TS 0 of this carrier and has to be transmitted continuously. If
no useful information is to be transmitted, Dummy Bursts have to be used. There
is no Power Control used on the cells beacon. Combination IV lacks dedicated
signaling channels (SDCCH and SACCH). Therefore, it has to be used together
with combination VII.

2.3 Problems and Solutions of the Air

Interface
It has already been pointed out that the radio air interface link is the most
vulnerable part of the GSM connection. In this section we will briefly discuss
some of the problems that occur in air interface and some solutions. There are
three major sources of problems in the air interface, which can lead to loss of
gigijeco.und.slo
data. These are:

multipath propagation
shadowing
propagation delay
2.3.1 Multipath Propagation

Whenever a mobile station is in contact with the GSM network, it is quite rare that
there is a direct "line of sight" transmission between the mobile station and the
base transceiver station. In the majority of cases, the signals arriving at the
mobile station have been reflected from various surfaces. Thus a mobile station
(and the base transceiver station) receives the same signal more than once.
Depending on the distance that the reflected signals have travelled, they may
affect the same information bit or corrupt successive bits. In the worst case, an
entire burst might get lost.
Depending on whether the reflected signal comes from near or far, the effect is
slightly different. A reflected signal that has travelled some distance causes "inter
symbol interference" whereas near reflections cause "frequency dips". There are
a number of solutions that have been designed to overcome these problems:
Viterbi equalisation
Channel coding
Interleaving

Frequency hopping
Antenna diversity
Fig. 15 The effects of multipath propagation
Viterbi equalisation
Viterbi equalisation is generally applicable for signals that have been reflected
from far away objects. When either the base transceiver station or mobile station
transmits user information, the information contained in the burst is not all user
data. There are 26 bits that are designated for a " training sequence" included in
each transmitted TDMA burst. Both the mobile station and base transceiver
gigijeco.und.slo
station know these bits and by analysing how the radio propagation affects these
training bits, the air interface is mathematically modelled as a filter. Using this
mathematical model, the transmitted bits are estimated based on the received
bits. The mathematical algorithm used for this purpose is called "Viterbi
equalisation".
Channel coding
Channel coding is normally used for overcoming the problem caused by fading
dips. In channel coding, the user data is coded using standard algorithms. This
coding is not for encryption, but for error detection and correction purposes and
requires extra information to be added to the user data. In the case of speech,
the amount of bits is increased from 260 per 20 ms to 456 bits per 20 ms. This
gives the possibility to regenerate up to 12.5% of data loss.
Interleaving
Interleaving is the spreading of coded speech into many bursts. By spreading the
information onto many bursts, we will be able to recover the data even if one
burst is lost. (Ciphering is also carried out for security reasons.)

Fig. 16 Speech processing in the mobile station
Frequency hopping
With frequency hopping, the frequency on which the information is transmitted is
changed for every burst. Frequency hopping generally does not significantly
improve the performance if there are less than four frequencies in the cell.
Frequency hopping randomizes errors over multiple user connections. The error
correction can thus combat errors more efficiently since each user will suffer from
error during a limited time.
gigijeco.und.slo
Fig. 17 Example of frequency hopping
Antenna receiver diversity

In this case, two physically separated antennas receive and process the same
signal. This helps to eliminate fading dips. If a fading dip occurs at the position of
one antenna, the other antenna will still be able to receive the signal. Since the
distance between two antennas is a few metres, it can only be implemented at
the Base Transceiver Station.

Fig. 18 Antenna receiver diversity
2.3.2 Shadowing
Hills, buildings and other obstacles between antennas cause shadowing (also
called log normal fading). Instead of reflecting the signal, these obstacles
attenuate the signal.
gigijeco.und.slo
Fig. 19 Shadowing effect
Shadowing is generally a problem in the uplink direction, because a Base

Transceiver Station transmits information at a much higher power compared to
that from the mobile station. The solution adopted to overcome this problem is
known as adaptive power control. Based on quality and strength of the
received signal, the base station informs the mobile station to increase or
decrease the power as required. This information is sent in the Slow Associated
Control Channel (SACCH).

2.3.3 Propagation delay

As you remember, information is sent in bursts from the mobile station to the
Base Transceiver Station (BTS). These bursts have to arrive at the Base
Transceiver Station so that they can map exactly into their allocated time slots.
However, the further away the mobile station is from the BTS, the longer it will
take for the radio signal to travel over the air interface. This means that if the
mobile station or base station transmits a burst only when the time slot appears,
then when the burst arrives at the other end, it will cross onto the time domain of
the next timeslot, thereby corrupting data from both sources.
The solution used to overcome this problem is called " adaptive frame alignment
". The Base Transceiver Station measures the time delay from the received signal
compared to the delay that would come from a mobile station that was
transmitting at zero distance from the Base Transceiver Station. Based on this
delay value, the Base Transceiver Station informs the mobile station to either
advance or retard the time alignment by sending the burst slightly before the
actual time slot. The base station also adopts this time alignment in the downlink
direction. This is also referred to as Timing Advance.
gigijeco.und.slo
Fig. 20 Propagation delay problem and solution
2.4 Terrestrial Transmission

So far, we have concentrated solely on the radio link between the Base
Transceiver Station (BTS) and the mobile station. Now we are going to follow the
signal to the next phase and take a look at the transmission between the other
network elements, in particular from the Base Transceiver Station to the Base
Station Controller (BSC) and up to the Mobile services Switching Centre (MSC).

2.4.1 Base Transceiver Station

A base transceiver station is a physical site from where the radio transmission in
both the downlink and uplink direction takes place. The radio resources are the
frequencies allocated to the Base Station. The particular hardware element inside
the Base Transceiver Station (BTS) responsible for transmitting and receiving
these radio frequencies is appropriately named " transceiver (TRX)". A Base
Station site might have any number of TRXs from one to twelve. These TRXs are
then configured into one, two or three cells. If a BTS is configured as one cell it is
called an " omnidirectional BTS " and if it is configured as either two or three
cells it is called a " sectorised BTS ".
gigijeco.und.slo
Fig. 21 Examples of BTS configurations
2.4.2 Transmission between BSC and BTS

There are three alternative methods to provide connections between a BSC and
several BTSs. The method used will depend on a number of factors, such as the
distance between the Base Station Controller (BSC) and the Base Transceiver
Station (BTS), the number of TRXs used at a particular BTS site, the signalling
channel rate between the Base Station Controller (BSC) and the Base
Transceiver Station (BTS). The methods to provide connection between a BSC
and several BTSEs include star, point to point connection, multidrop, multidrop
loop, loop and cross connect.

Fig. 22 BTS - BSC connections
Point-to-point connection indicates that the Base Station Controller (BSC) is

connected directly to every BTS with a PCM line. This is a simple and effective
method, particularly in cases when the distance between the BSC and the BTS is
short. However, if the BSC - BTS distance is a few kilometres, whereas the
distance between a group of BTSs is much shorter, it does not make sense to
draw a point-to-point connection to every BTS. One PCM line has ample capacity
to transfer data to several BTSs simultaneously. Therefore, it is possible to draw
gigijeco.und.slo
just one BSC - BTS connection and link the BTSs as a chain. This technique is
called " multidrop chain". The BSC sends all the data in one PCM line and each
BTS in turn analyses the signal, collects the data from the correct timeslots
assigned for itself and passes on the signal to the next BTS.
There is, however, one problem with a multidrop chain. Consider what would
happen if there is a malfunction somewhere along the line and the chain breaks.
More BTSs are isolated and, if the BSC is not informed, it will continue to send
data. The solution to this problem is called " multidrop loop " and instead of a
chain we connect the BTSs in the form of a loop. Previously a dynamic node was
needed to split the signal into the two directions around the loop, but later
versions of BTS are capable of carrying out this function. The flow of the signal is
similar to the signal flow in a multidrop chain, except that a BTS will change the
listening direction if the signal from one side fails. This ensures that the BTSs
always receive information from the BSC even if the connection is cut off at some
point in the loop.

2.4.3 Excursion: Voice Transmission in Fixed Networks

Acoustic signals, such as speech, must be converted into electrical signals to
transport them over longer distances - they have to be modulated on a carrier:
Nowadays, acoustical signals are even digitalised prior to transport. At the
senders side:
1. Acoustical signals are transformed into electrical signals, then
2. An analogue-digital transformation takes place, and finally,
3. The digital information is modulated on a carrier frequency band.
Then, the user data can be transmitted over long distances. At the receivers side
these processes are carried out in the reverse order to regenerate the original
acoustic signal:
1. De-modulation,
2. Digital/analogue conversation, and
3. Conversion of the analogue electrical signal into an acoustical signal.
gigijeco.und.slo
Fig. 23 Pulse Code Modulation
Pulse Code Modulation

Many different digitalisation methods exist. The most common one is Pulse Code
Modulation (PCM). Telephony networks were originally developed to transmit
speech. For the human ear, the most relevant speech information is transmitted
in the frequency bands between 300 and 3400 Hertz. Therefore, the acoustical
signals arriving at the microphone are filtered and only frequencies between 0.3
and 3.4 kHz are considered in the digitalisation process.
The highest frequency to be represented is thus 3.4 kHz. In order to represent it,
at least 2 sample values are required. In PCM 8000 sample values are taken,
e.g. one sample value every 125 s.

In the next step, the sampled values must be digitalised. In PCM, each sample
value is represented by a 8 bit value. The resulting data stream for speech is thus
8000 times 8 bits = 64 kbps.
PCM 30
If each call requires its one wire to be transmitted, then tremendous amounts of
cables are required. Luckily, several calls can be transmitted via the same
physical medium. The process of combining several calls on one physical
medium is called multiplexing. Time Division Multiplexing (TDM) is nowadays
often used.
One example of TDM is PCM30 (or PCM 24 in North America), which is quite
common in mobile communication networks. The transmission between network
elements is organised in 125 s time frames. 125 s time frames is subdivided in
32 time slots. Time slot number 0 is required for synchronisation and alarms.
Time slot number 16 is used in many countries to transmit common channel
signalling information. (Also other time slots can be used to carry this signalling
information.) If this is the case, then 30 time slots are left for user data transport.
Each time slot transmits 8 bits. If a speech call is set-up, one of the time slots
can be reserved for the subscriber. Let us take time slot 5. As long as the speech
call takes place, 8 bits of speech information are cyclically transmitted in time slot
5 between the two network elements. The remaining time slots can be allocated
to other calls.
gigijeco.und.slo
Fig. 24 Pulse Code Modulation
2.4.4 The Concept of Multiplexing

According to GSM 900 and GSM 1800 specifications, the bit rate for speech in
the radio access interfaces is 13 Kbits/s (full rate, enhanced full rate) and the bit
rate in the Mobile services Switching Centre (MSC) and PSTN interfaces is 64
Kbits/s. This means that the bit rate has to be converted at some point after the
signal has been received by the BTS and before it is sent to other networks. But
the specifications do not put a constraint as to where exactly the conversion
should take place. This brings up some interesting scenarios.

The actual hardware that does the conversion from 13 Kbits/s to 64 Kbits/s and
vice versa is called a Transcoder. In theory, this piece of equipment belongs to
the Base Transceiver Station. However, by putting the transcoder at a different
place we can take some advantages in reducing the transmission costs.
If the transcoder is placed at the BTS site (in the BSC interface), the user data
rate from the BTS to the Base Station Controller (BSC) would be 64 Kbits/s. The
transmission for this would be similar to standard PCM line transmission with 30
channels per PCM cable. The same would also apply between the BSC and the
MSC.
If we put the transcoder somewhere else, say just after the MSC, we cannot get a
significant advantage either. This is because, although after transcoding the bit
rate reduces to 13 Kbit/s, we still have to use the PCM structure to send the
traffic channels, with 8 bits per time slot. However, since after transcoding we
have a bit rate of 13 Kbits/s and an additional 3 Kbits/s (making 16 Kbits/s) only
two bits per time slot will be used. The other 6 bits are effectively wasted.
The additional 3 Kbits/s are used between the transcoder and the BTS for
carrying the 260-bit vocoder block (equivalent with 20 ms of speech) and for
inband signalling (information about speech coding algorithm, type of call, etc)
when a call is connected.
Independent from its actual position, the transcoder belongs to the BSS even if it
is placed next to the MSC. (A TC that is placed away from the BTS is called a
remote TC according to the GSM recommendations).
gigijeco.und.slo
Fig. 25 Implementations of the transcoder at different sites
The real advantage, however, comes if we use the second configuration shown in
the figure. We saw that from the MSC, data comes out at a 64 Kbits/s rate and
from the Transcoder it comes out at 16 Kbits/s. Each PCM channel (timeslot) has
2 bits of information. It appears that we are able to put in data from three more
PCM lines by multiplexing. Consequently, we are able to multiplex up to four
PCM lines and send up to 120 traffic channels in one PCM line from the MSC
(transcoder) towards the BSC. Common Channel Signalling information, OMC
data and some other network information are, however, not transcoded.

2.5 Summary of the Learning Points

GSM networks use Time Division Multiple Access (TDMA) technology in the
air interface. By this method, one frequency resource can be shared by a
maximum of eight mobile users.
There are eight physical channels per frequency in the air interface.
Logical channels are classified according to the type of information contained
within each channel.
There are eleven logical channels. Two of them are half and full rate traffic
channels. The remaining nine are various control channels used to transfer
information related to call set-up.
Information is sent from the mobile station to the Base Transceiver Station
(BTS) in intermittent bursts.
There are four primary types of bursts: Normal burst, access burst,
synchronisation burst and frequency correction burst.
There are primarily three sources of problems in the air interface. There is
multipath propagation, shadowing and propagation delay. The methods
adopted to overcome these problems are viterbi equalisation, channel coding,
frequency hopping, interleaving, antenna receiver diversity, adaptive power
control and adaptive frame alignment.
A Base Transceiver Station (BTS) site can be configured as an
omni-directional BTS or a sectorised BTS.
gigijeco.und.slo
There are three different methods of connecting a Base Transceiver Station

(BTS) to the Base Station Controller (BSC): point to point, multidrop chain
and multidrop loop.
The transcoder is placed at the Mobile services Switching Centre (MSC). A
transcoder makes it possible to multiplex traffic channels from three PCM
lines, thereby reducing transmission costs.

3 Network Planning and

Optimisation
3.1 Introduction
The geographical distribution of the subscribers poses a difficult problem for GSM
networks. Without wire-connected telephones the subscribers can be virtually
everywhere, but still the network must be able to provide a connection in spite of
their movements. A good geographical coverage is the basis for providing
network services. Careful network planning is thus a primary aspect of
implementing GSM networks. One common misperception is that GSM network
planning is the task of selecting and placing out base stations, but this is only one
part. Network planning is an ongoing process, which requires inputs from a lot
competence areas, such as transmission, access systems, data communications,
mobile switching, Intelligent Network, site acquisition, etc.
Several requirements must be taken into consideration already in the early stages
of the planning process:
gigijeco.und.slo
Fig. 26 Network planning map
Costs of building the network

Capacity of the network
Coverage and location of network elements
Maximum congestion allowed (grade of service)
Quality of calls
Further development of the network.
Various factors that affect the demand for network services must also be
considered. These are mostly related to the inhabitants of the area, such as
distribution of the population and vehicles, income of the population and statistics
on telephone usage.

The main steps of a network planning process are as follows:

1. Collection of all relevant information, for instance:
- Regulations and laws.
- Key information concerning, for instance, demography, income level,
penetration forecast, geographical extension forecast, services to be
supported, market segmentation, etc.
- Availability of leased lines, microwave frequency availability,
required connections with other systems, etc.
- Numbering, addressing and routing principles.
- Topographical maps.
- Existing infrastructure, for example transmission network and transmission
media.
Already in this first planning step, site acquisition management
must be involved, as it is necessary to identify the potential sites at an early
stage.
2. Network dimensioning based on coverage and capacity requirements. The
main objective here is to optimise the network in a cost-efficient way. In order
to succeed, detailed information about the network is needed, such as growth
estimates, protection, available and needed infrastructure, plus goals for
quality and performance. The outcome is an integrated network architecture
design that shows how the different services will be implemented and what
equipment will be needed. Also, a preliminary rollout plan should be included.
gigijeco.und.slo
3. Selection of MSC, BSC and Base Station sites (in this order).
4. Survey of intended MSC, BSC and Base Station sites, in other words to
evaluate the intended location of each MSC, BSC and BTS. It is important to
analyse if the locations fulfil the requirements. Factors to consider when, for
instance, making a BTS site survey are the surroundings, possible structural
and geographical obstacles and existing radio equipment.
5. Detailed network planning. Computer aided design systems and tools are
used for coverage prediction, interference analysis, frequency planning,
microwave link planning, documentation etc.
By co-ordinating different information and different competencies, it is possible to
prepare a network plan that includes the coverage predictions and dimensioning
of the network.
In the following, we shall take a closer look at three different areas of network
planning, namely Switching Network planning, Cellular Transmission Network
planning and Radio Network planning. The main purpose is to illustrate the
complexity of the planning process. Note also that areas like Fixed Transmission
planning, Network Access planning, DCN (Data Communications Network)
planning, Intelligent Network planning, 3G and IP Network planning etc., are not
covered in this book, but that they also have to be included in the planning
process.

3.2 Switching Network Planning

Switching network planning plays an important role in the process. During the
network dimensioning process, there are several important tasks to consider. The
switches need to be dimensioned in accordance with the estimated needed
capacity, for instance, average conversation time, the need for signalling, such as
number of handovers, location updates, short message distribution, etc.
It is necessary to define the network performance level, for instance the intended
capacity of the switching network. The implementation of switching and signalling
network must also be considered. Moreover, there is a need to create rules for
routing, protection, synchronisation and switch management. Voice and signalling
traffic matrixes must be defined, as well as the equipment to be used must be
planned.
After the network has been dimensioned, a detailed plan is made with a number
of inputs, for instance Data Communications Network (DCN) settings towards the
NMS, network diagrams, the synchronisation plan, a detailed routing plan, digit
analysis, detailed signalling, numbering and charging plans, etc.
In addition to this, the planners must also consider the future expansion plans in
order to have a roadmap to the future.
3.3 Cellular Transmission Network Planning
gigijeco.und.slo
Cellular transmission network refers to usage of microwave links in the GSM
network, for instance between the Base Transceiver Stations and the Base
Station Controllers. Usually, the main alternative is to use leased lines, that is, to
utilise the already existing fixed infrastructure.
In the network dimensioning phase, a lot of different information has to be
collected and co-ordinated. The general transmission network diagram for base
stations access and core networks must be drawn, in order to get a clear general
picture of the network connections. It is also necessary to define the capacity
requirements. Furthermore, rules for the general management of network are to
be defined. Gateway and switching network connections need to be identified, as
well as synchronisation principles for the network. Last, but not least, the future
expansion plans must be taken into consideration.
In the detailed planning, a network layout diagram and a management network
diagram must be drawn in order to identify how the network is to be implemented.
Connections between sites as well as capacities, timeslot allocation of each link,
routing diagrams, etc. must be defined. A more exact synchronisation plan with
sources and hierarchy is needed.
An essential part of the planning of microwave links is to make sure that it is
possible to get free line of sight between different elements (for instance between
BTS and BSC sites).

3.4 Radio Network Planning

The type and location of the BTS depends on the characteristics of the
surroundings. In city areas, cells are usually smaller than in the countryside. A
larger traffic volume also affects the number of channel frequencies in a certain
cell (TRXs).
The maximum theoretical distance from a BTS to the edge of the cell is 35
kilometres. The timing advance is adapted for these distances (that is the ability
of the mobile station to send the bursts in advance so that they arrive in the BTS
in the right timeslot). One factor that limits the cell size is the wavelength of the
frequency. The rule is that the higher the frequency, the smaller the size of the
cell. It means that the potential cell coverage in GSM 900 is larger than for 1800
and 1900 networks. Another factor that affects the cell size is the geographical
condition. As an example, open water attenuates the signal less than for instance
forest or city environment. It means that the signal can travel longer distances on
open water.
As you can see from the picture below, there are three different types of BTS
configurations, omnidirectional, 2 sector and 3 sector cells. The omnidirectional
BTS has an antenna that transmits and receives 360 degrees. The sectorised (2
sector and 3 sector) BTSs have sectorised antennas, which receive and transmit
in certain directions only. The 2 sector BTS is suitable for, for instance, providing
coverage for a highway.
gigijeco.und.slo
Fig. 27 Simulated cellular radio network planning

The omnidirectional BTS is not as popular as the sectorised alternatives. There

are mainly two reasons for this. Firstly, the degree of interference is higher than
that for sectorised BTSs, which in turn restrict the frequency reuse possibilities.
Secondly, the gain value is lower than for omnidirectional BTSs. In other words,
by using a sectorised antenna the covered distance increases.
Fig. 28 BTS configurations
gigijeco.und.slo
After all the installation sites have been surveyed, a detailed network plan can be
made. This includes the design of a transmission network. One important task
is to choose between leased lines and microwave links.
After the installation work has been completed, the radio environment has to be
measured and tested to ensure its proper operation and coverage before putting
it into use. This is carried out in the surroundings of each individual site using
portable test transmitters.
3.4.1 Dimensioning Cells

A cell is the basic construction block of a GSM network. One cell is the
geographical area covered by one BTS. The actual size of a cell depends on
several factors: the environment, number of users, etc. Cells are grouped under
Base Station Controllers (BSC).

Dimensioning a cell means finding answers to two fundamental questions: How

many traffic channels (TCH) does the cell need to handle and how many traffic
channels are necessary? To solve these problems, that is, to determine the traffic
capacity, we have to calculate the number of Erlangs. Erlang is the measuring
unit of network traffic. One Erlang equals the continuous use of a mobile device
for one hour. The traffic is calculated using a simple formula:
The amount of traffic is independent of the observation duration. For example, it

is possible to make observation for only 15 minutes and then, in the formula
above, calls per 15 minutes is taken and it is divided by 900 seconds.
With heavy traffic loads on available resources, the bigger the likelihood that
there will be congestion on these resources. Network planners carefully analyse
the traffic volume on installed traffic channel capacity and, according to quality
limits in the network, decide if there is a need to install more capacity.
Let us take an example: if there are 540 calls per hour and the average
conversation time is 100 seconds, the traffic capacity is 15 Erlangs. After
obtaining this value, we must take a look at the Erlang table .
Table 2 Erlang B Table

gigijeco.und.slo
Chs 1% 2% 3% 5% Chs 1% 2% 3% 5%
1 0.01 0.02 0.03 0.05 21 12.80 14.00 14.90 16.20
2 0.15 0.22 0.28 0.38 22 13.70 14.90 15.80 17.10
3 0.46 0.60 0.72 0.90 23 14.50 15.80 16.70 18.10
4 0.87 1.09 1.26 1.52 24 15.30 16.60 17.60 19.00
5 1.36 1.66 1.88 2.22 25 16.10 17.50 18.50 20.00
6 1.91 2.28 2.54 2.96 26 17.00 18.40 19.40 20.90
7 2.50 2.94 3.25 3.75 27 17.80 19.30 20.30 21.90
8 3.13 3.63 3.99 4.54 28 18.60 20.20 21.20 22.90
9 3.78 4.34 4.75 5.37 29 19.50 21.00 22.10 23.80
10 4.46 5.08 5.53 6.22 30 20.30 21.90 23.10 24.80
11 5.16 5.84 6.33 7.08 31 21.20 22.80 24.00 25.80
12 5.88 6.61 7.14 7.95 32 22.00 23.70 24.90 26.70
13 6.61 7.40 7.97 8.83 33 22.90 24.60 25.80 27.70

14 7.35 8.20 8.80 9.73 34 23.80 25.50 26.80 28.70

15 8.11 9.01 9.65 10.60 35 24.60 26.40 27.70 29.70
16 8.88 9.83 10.50 11.50 36 25.50 27.30 28.60 30.70
17 9.65 10.70 11.40 12.50 37 26.40 28.30 29.60 31.60
18 10.40 11.50 12.20 13.40 38 27.30 29.20 30.50 32.60
19 11.20 12.30 13.10 14.30 39 28.10 30.10 31.50 33.60
20 12.00 13.20 14.00 15.20 40 29.00 31.00 32.40 34.60
As you can see, the table also contains the grade of service (GOS) figure, which
is the maximum congestion allowed. Supposing that GOS is 5 % - which means
that during a certain observation period (usually 1 hour) 5 out of 100 calls fail due
to lack of resources - the required number of channels is 20. Since each carrier
supports eight channels, we can make a rough estimate that this cell must be
equipped with three carriers, that is, three transceivers or TRXs.
3.4.2 Frequency Reuse

Now we have to resolve another problem. There is a limited number of
frequencies available to each Base Station Subsystem. These frequencies must
be distributed between the cells, so that the capacity requirements are met in
different parts of the BSS. Let us make a simplified exercise to illustrate the
gigijeco.und.slo
situation.
You are the network planner and the number of frequencies assigned to this
project is 9. Your task is to distribute the frequencies in the network that is shown
in the following figure with one frequency per cell.

Fig. 29 Frequency planning exercise
As you can see, the frequencies have to be reused. If you do not distribute the
frequencies properly throughout the network, the result will be a high level of
interference caused by overlapping frequencies. To avoid this, the GSM network
gigijeco.und.slo
includes a specification of the Frequency reuse patterns. In the figure below,

one of these is presented. Note, however, that cell planning is normally made
with the help of a computer based design program.
Fig. 30 Frequency reuse pattern example

The next step involves the dimensioning of location areas. This is carried out
according to the traffic characteristics of each area. The final phase is the
dimensioning of the fixed network on the basis of the traffic requirements and
dimensioning of the entire radio network.
3.5 Optimising and Developing the Network

The network planning described so far is only the first part of a continuous
process of improving the network performance. In the following section , we will
discuss the importance of network optimisation and development. Furthermore,
different ways of obtaining information about the network are presented.
3.5.1 Why is it Important?

The mobile network needs to be monitored continuously, and there are several
reasons for this.
Firstly, the increase in the number of subscribers requires network expansion at
the right times and places. A favourable cost structure is of major importance for
any operator in a competitive marketplace. Therefore, excess network capacity
should be avoided. At the same time, it is necessary to offer sufficient grade of
service to the subscribers. It is certainly a paradox that the network should be
large enough (sufficient coverage and quality) and small enough (cost efficient) at
gigijeco.und.slo
the same time.
Secondly, the quality of service experienced by the subscribers has to be high.
Therefore, it is important to, for instance, reduce the number of dropped calls and
the degree of interference in the air interface (in a cost efficient way).
Thirdly, present and future demand for basic and supplementary services must
be satisfied. Today, large investments are directed towards facilitating higher bit
rates for data transmission (High Speed Circuit Switched Data (HSCSD), General
Packet Radio Service (GPRS) and Enhanced Data Rates for Global Evolution
(EDGE). The reason is that huge growth is anticipated in the demand for wireless
data transmission services for the next years to come.
3.5.2 Sources of information

It is obvious that there are several factors to take into consideration when
performing network optimisation and development. When we monitor the
performance of the network, we can utilise different sources of information. We
can obtain a lot of useful information from the Network Management System, and
also other performance management tools can be used, for instance Nokia's
Traffica monitoring tool. Furthermore, we can perform field-testing in order to
identify the performance from the user's point of view. Last, but not least, the
subscribers can give us valuable information concerning the experienced quality
of service.

Fig. 31 Toolkit Drive survey
There are of course advantages and disadvantages associated with each of

these information sources.
NMS and performance management tools give us valuable information about
gigijeco.und.slo
different parts of the network, and at least for larger operators, they are relatively
inexpensive alternatives. It is also possible to tailor what kind of information to
gather, and what kind of reports to produce (for operation and maintenance
purposes, for executive reports, etc.). One of the problems with these tools is that
they cannot fully reflect the quality the way the subscribers experience it.
Field-testing is an excellent way of receiving information in different geographical
areas. The findings also reflect the actual quality as perceived by the subscribers.
Furthermore, it is possible to monitor the competitors' networks and compare
them with the own network. One of the major drawbacks with field tests is that it
they are costly processes.
Customers can give valuable information about the network, as their feedback
reflects the experienced quality. The subscribers may also discover problems that
would not otherwise be identified. It can seem to be an inexpensive way of
gathering information, but do consider that large resources are engaged when
customer calls are followed up (problem identification and location). In addition to
this, low customer satisfaction decreases the success of the operator. Therefore,
customer feedback should be regarded as a complementary information source
besides NMS/performance management tools and field-testing.

3.6 Summary of the Learning Points

Two of the most important factors in planning a GSM network are the
coverage and capacity of the network.
The demographic data of the intended coverage area plays a major role in
network planning.
Frequencies have to be reused in GSM network according to certain
predefined methods.
Omnidirectional cells are less popular than sectorised due to smaller
coverage (lower gain value) and higher degree of interference.
The required number of traffic channels in a cell is estimated with the help of
Erlang's formula of loss and waiting times.
The network needs continuous monitoring and optimisation, as there are
constant changes in both the number of users and technologies. Reliable
information sources are NMS/performance management tools and field
testing. Additional information can be obtained with the help of subscribers'
feedback.
gigijeco.und.slo

Module 5 - GSM Air Interface & Network Planning

Uploaded by

Copyright:

Available Formats

Module 5 - GSM Air Interface & Network Planning

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Module 5 - GSM Air Interface & Network Planning

Uploaded by

Copyright:

Available Formats

GSM Air Interface and Network Planning

GSM Air Interface and

2.3 Problems and Solutions of the Air Interface............................................. 21

Nokia Siemens Networks 2008 1-1

1-2 Nokia Siemens Networks 2008

2.1.1 Duplex Transmission

Duplex transmission is enabled in GSM by using different frequency bands.

Fig. 1 Frequency allocations for GSM

Nokia Siemens Networks 2008 1-3

2.1.2 Multiple Access

1-4 Nokia Siemens Networks 2008

Table 1. GSM 900 frequencies

Channel Uplink signal (MHz) Downlink signal (MHz)

2 890.4 (centre freq.) 935.4 (centre freq.)

Nokia Siemens Networks 2008 1-5

Fig. 2 Time Division Multiple Access Principle

2.1.3 GSM Frequency Bands

1-6 Nokia Siemens Networks 2008

Fig. 3 Examples of frequency and amplitude modulation

The transmitter sends a bit value, which is represented by an analogue

Nokia Siemens Networks 2008 1-7

Fig. 4 Binary Phase Shift Keying (BPSK)

Fig. 5 Example of phase modulation

1-8 Nokia Siemens Networks 2008

2.2 Transmission through the Air Interface

Fig. 6 A small logistical problem

2.2.1 Physical and Logical Channels

Nokia Siemens Networks 2008 1-9

The synchronisation information required by this process is broadcast by the

1-10 Nokia Siemens Networks 2008

Fig. 8 Initiation of a call

Once the subscriber is registered in the network and the authentication is

needed. A dedicated signalling channel is assigned in order to transmit the

Fig. 9 Call completion from the called side

Nokia Siemens Networks 2008 1-11

2.2.2 Logical Channels

Fig. 10 Logical channels

1-12 Nokia Siemens Networks 2008

Fig. 11 TDMA frames with common and dedicated channels

2.2.2.1 Broadcast channels

Nokia Siemens Networks 2008 1-13

1. Frequency Correction Channel (FCCH)

Frequencies: Used in the particular cell and neighbouring cells.

1-14 Nokia Siemens Networks 2008

2.2.2.2 Common control channels

2.2.2.3 Dedicated control channels

2.2.2.4 Traffic Channels (TCH)

Nokia Siemens Networks 2008 1-15

2.2.3 Time Slots and Bursts

1-16 Nokia Siemens Networks 2008

Fig. 12 TDMA bursts and timeslots

Nokia Siemens Networks 2008 1-17

In the air interface, a TDMA timeslot is a time interval of approximately 576.9 s,

1-18 Nokia Siemens Networks 2008

to guarantee, that there is permanently a load on this carrier.

Fig. 14 Frequency correction, synchronisation, and dummy bursts2.3.4

Nokia Siemens Networks 2008 1-19

Multiframe: Channel Combinations

1-20 Nokia Siemens Networks 2008

2.3 Problems and Solutions of the Air