Cell Planning Overview

GSM CELL PLANNING OVERVIEW
GSM CELL PLANNING

OVERVIEW
For Study Only
FOR INTERNAL USE ONLY 1
TABLE OF CONTENTS
1. Cell Planning Introduction............................................................................1
1.1 INTRODUCTION............................................................................................... 2
1.2 CELL PLANNING PROCESS............................................................................2
. Sy!te" #e!cri$tion.......................................................................................%
2.1 GLOBAL SYSTEM FOR MOBILE COMMUNICATIONS....................................7
2.2 THE DIFFERENT GSM-BASED NETWORKS..................................................7
2.3 NETWORK HARDWARE..................................................................................7
2.4 OPERATION AND SUPPORT SYSTEM (OSS.................................................!
2." SWITCHING SYSTEM (SS..............................................................................!
2.6 BASE STATION SYSTEM (BSS...................................................................1#
&. 'ra((ic............................................................................................................1&
3.1 TRAFFIC AND CHANNEL DIMENSIONING...................................................14
3.2 CHANNEL UTILI$ATION.................................................................................17
). No"inal Cell Plan........................................................................................1*
4.1 INTRODUCTION.............................................................................................2#
4.2 SYSTEM BALANCING....................................................................................2#
4.3 CHANNEL LOADING PLAN............................................................................21
4.4 WAVES........................................................................................................... 2!
4." GENERATION OF RADIO WAVES.................................................................3#
4.% SUPERIMPOSING INFORMATION ON RADIO WAVES................................34
4.7 AIR INTERFACE DATA...................................................................................3"
4.! RADIO WAVE PROPAGATION.......................................................................3!
4.& SIGNAL VARIATIONS..................................................................................... 41
+. Sur,ey!.........................................................................................................))
".1 INTRODUCTION.............................................................................................4"
".2 RADIO NETWORK SURVEY..........................................................................4"
".3 RADIO MEASUREMENTS..............................................................................4!
%. Sy!te" 'uning.............................................................................................+1
%.1 INTRODUCTION............................................................................................."2
%.2 SYSTEM DIAGNOSTICS................................................................................"2
%.3 CELL PARAMETER AD'USTMENT................................................................"%
%.4 SYSTEM GROWTH........................................................................................ "7
1. Cell Planning Introduction
C-a$ter 1
This chapter is designed to provide the student with an
Introduction to cell planning.
O./EC'IVES0
Upon completion of this chapter, the student will be able to:
Explain briefly the major steps in cell planning
escribe what support !uawei offers regarding cell
planning services
1.1 INTRODUCTION
This course, "ell #lanning $verview, is intended to give the
student an understanding of the %adio &re'uency (%&)
engineering processes and what elements they contain.
The course is bro*en down into chapters that explain the
different elements of the process.
1.2 CELL PLANNIN PROCESS
"ell planning can be described briefly as all the activities
involved in determining which sites will be used for the radio
e'uipment, which e'uipment will be used, and how the
e'uipment will be configured.
In order to ensure coverage and to avoid interference, every
cellular networ* needs planning. The major activities involved
in the cell planning process are depicted in &igure +,+.
F()*+, 1-1 T-, .,// 0/122(2) 0+3.,44
S'EP 10 'RAFFIC AN# COVERAGE ANAL1SIS 2S1S'EM
RE34IREMEN'S5
The cell planning process starts with traffic and coverage
analysis. The analysis should produce information about the
geographical area and the expected need of capacity. The types
of data collected are:
- "ost
- "apacity
- "overage
- .rade of /ervice (.o/)
- 0vailable fre'uencies
- 1it Error %ate (1E%)
- /ystem growth capability
The traffic demand (i.e. how many subscribers will join the
system and how much traffic will be generated) provides the
basis for cellular networ* engineering. .eographical distribution
of traffic demand can be calculated by using demographic data
such as:
- #opulation distribution
- "ar usage distribution
- Income level distribution
- 2and usage data
- Telephone usage statistics
- $ther factors such as subscription charges, call charges, and
price of mobile stations
S'EP 0 NOMINAL CELL PLAN
Upon compilation of the data received from the traffic and
coverage analysis, a nominal cell plan is produced. The nominal
cell plan is a graphical representation of the networ* and simply
loo*s li*e a cell pattern on a map. !owever, a lot of wor* lies
behind it (as described previously).
3ominal cell plans are the first cell plans and form the basis for
further planning. 4uite often a nominal cell plan, together with
one or two examples of coverage predictions, is included in
tenders.
0t this stage, coverage and interference predictions are usually
started. /uch planning needs computer,aided analysis tools for
radio propagation studies, e.g. !uawei5s planning tool *nown as
the !uawei #lanning Tool (0//ET).
S'EP &0 S4RVE1S 2AN# RA#IO MEAS4REMEN'S5
The nominal cell plan has been produced and the coverage and
interference predictions have been roughly verified. 3ext, radio
measurements are performed at the sites where the radio
e'uipment will be placed. This is a critical step because it is
necessary to assess the real environment to determine whether it
is a suitable site location when planning a cellular networ*,
since even better predictions can be obtained by using field
measurements of the signal strengths in the actual terrain where
the mobile station will be located.
S'EP )0 2FINAL CELL PLAN5 S1S'EM #ESIGN
$nce we optimi6e and can trust the predictions generated by the
planning tool, the dimensioning of the 1T/ e'uipment, 1/",
and 7/" is performed. The final cell plan is then produced. 0s
the name implies, this plan is later used during system
installation.
In addition, a document called "ell esign ata (") is filled
out containing all cell parameters for each cell.
S'EP +0 IMPLEMEN'A'ION
/ystem installation, commissioning, and testing are performed
following final cell planning and system design.
S'EP %0 S1S'EM '4NING
0fter the system has been installed, it is continually evaluated to
determine how well it meets the demand. This is called system
tuning. It involves:
8 "hec*ing that the final cell plan was implemented
successfully
8 Evaluating customer complaints
8 "hec*ing that the networ* performance is acceptable
8 "hanging parameters and performing other measures (if
needed)
The system needs constant retuning because the traffic and
number of subscribers increases continuously. Eventually, the
system reaches a point where it must be expanded so that it can
manage the increasing load and new traffic. 0t this point, a
coverage analysis is performed and the cell planning process
cycle begins again.
FOR INTERNAL USE ONLY "
2. S!"te# De"cri$tion
C-a$ter
This chapter is designed to provide the student with an overview
of the ./7 system.
O./EC'IVES0
Upon completion of this chapter the student will be able to:
9 Explain the basic functionality of a ./7 system
9 escribe the networ* nodes of a ./7 system
9 escribe general terms used in the ./7 system
9 escribe the geographical networ* structure
FOR INTERNAL USE ONLY %
2.1 LOBAL S%STE& FOR &OBILE CO&&UNICATIONS
The next step in the ./7 evolution was the specification of
#ersonal "ommunication 3etwor* (#"3) for the +:;; 7!6
fre'uency range. This was named the igital "ellular /ystem
("/) +:;; (or !uawei5s ./7 +:;;). The #ersonal
"ommunication /ervices (#"/) +<;; (or !uawei5s ./7 +<;;)
for the +<;; 7!6 fre'uency range was also established.
2.2 T'E DIFFERENT S&(BASED NET)OR*S
ifferent fre'uency bands are used for ./7 <;;=+:;; and
./7 +<;; (&igure >,+). In some countries, an operator applies
for the available fre'uencies. In other countries, e.g. United
/tates, an operator purchases available fre'uency bands at
0uctions.
3etwor*
type
&re'uency band
U2=2(7h6)
!uawei
implementations
./7 <;; :<;,<+?=<-?,<9; ./7<;;=+:;;
./7+:;; +@+;,+@:?=+:;?,
+::;
./7<;;=+:;;
F()*+, 2-1 F+,5*,2.6 71284 93+ :-, 8(99,+,2: GSM-714,8
2,:;3+<4
2.+ NET)OR* 'ARD)ARE
Every cellular system has hardware that is specific to it and each
piece of hardware has a specific function. !uawei5s ./7
system comply with the ./7 standard, with the addition of
!uawei specific improvements.
The system solutions integrate existing !uawei hardware and
new technology to provide a AtotalA solution to the mobile
telephony mar*et. The major system in the networ* are:
: $peration and /upport /ystem
: /witching /ystem
: 1ase /tation /ystem
The system is normally configured as depicted in &igure >,>.
F()*+, 2-2 H*1;,( GSM-714,8 464:,= =38,/
2., OPERATION AND SUPPORT S%STE& -OSS.
&or ./7 system administration, the $// supports the networ*
operator by providing:
: "ellular networ* administration
: 3etwor* operation and maintenance
2./ S)ITC'IN S%STE& -SS.
FOR INTERNAL USE ONLY !
F()*+, 2-3 S;(:.-(2) S64:,=
MOBILE SERVICES SWITCHING CENTER MSC
The 7/" is responsible for set,up, routing, and supervision
of calls to and from mobile subscribers. $ther functions are
also implemented in the 7/", such as authentication. The
7/" is built on an "";: platform.
HOME LOCATION REGISTER HLR
In ./7, each operator has a database (the !2%) containing
information about all subscribers belonging to that specific
#ublic 2and 7obile 3etwor* (#273). 2ogically there is
only one !2% per #273 but it can be implemented
physically in one or more databases. Examples of
information stored in the database are the location
(7/"=B2% service area) of the subscribers and the services
attached to the subscription. The !2% is built on an "";:
platform.
VISITOR LOCATION REGISTER VLR
In the !uawei ./7 based solution, the B2% is integrated
with the 7/". This is referred to as the 7/"=B2%. The
B2% contains non,permanent information about the mobile
subscribers visiting the 7/"=B2% service area (e.g. which
location area the 7/ is in currently and which services are
activated).
GATEWAY MSC GMSC
The .ateway 7/" (.7/") supports the function for
routing incoming calls to the 7/" where the mobile
FOR INTERNAL USE ONLY &
subscriber is currently registered. It is normally integrated in
the same node as an 7/"=B2%.
AUTHENTICATION CENTER AUC
&or security reasons, speech, data, and signaling are
ciphered, and the subscription is authenticated at access. The
0U" provides authentication and encryption parameters for
subscriber verification to ensure call confidentiality.
EQUPI IDENTITY REGISTER EIR
In ./7 there is a distinction between subscription and
mobile e'uipment. 0s mentioned above, the 0U" chec*s
the subscription at access. The EI% chec*s the mobile
e'uipment to prevent a stolen or non,type,approved 7/
from being used.
Short Message Service Gatea! MSC "SMS#GMSC10 10
0 /hort 7essage /ervice .ateway 7/" (/7/,.7/") is
capable of receiving a short message from a /ervice "enter
(/"), interrogating an !2% for routing information and
message waiting data, and delivering the short message to
the 7/" of the recipient 7/. In !uawei5s ./7 system, the
/7/,.7/" functionality is normally integrated in an
7/"=B2% node.
Short Message Service I$terWor%i$g MSC "SMSIWMSC10
0 /hort 7essage /ervice InterCor*ing 7/" (/7/IC7/")
is capable of receiving a mobile originated short
message from the 7/" or an 0lert message from the
!2% and submitting the message to the recipient /". The
/7/,IC7/" functionality is normally integrated in the
7/"=B2% node.
2.0 10 BASE STATION S%STE& -BSS.
The 1ase /tation /ystem (1//) is comprised of two major
components. They are:
+; 1ase /tation "ontroller (1/")
FOR INTERNAL USE ONLY 1#
++ 1ase Transceiver /tation (1T/)
B
F()*+, 2-4 B14, S:1:(32 S64:,=
.SC
The 1ase /tation "ontroller (1/") is the central point of the
1//. The 1/" can manage the entire radio networ* and
performs the following functions:
++ !andling of the mobile station connection and handover
++ %adio networ* management
++ Transcoding and rate adaptation
++ Traffic concentration
++ Transmission management of the 1T/s
++ %emote control of the 1T/s
.'S
The 1ase Transceiver /tation (1T/) includes all radio and
transmission interface e'uipment needed in one cell. The
!uawei 1T/ corresponds to the e'uipment needed on one site
rather than one cell. Each 1T/ operates at one or several pairs of
fre'uencies. $ne fre'uency of each pair is used to transmit
signals to the mobile station and the other is used to receive
signals from the mobile station. &or this reason at least one
transmitter and one receiver is needed.
+. Tra11ic
C-a$ter &
of basic concepts of networ* dimensioning.
O./EC'IVES0
+> efine the terms trafficAand A.rade of /erviceA(.o/)
+> Use Erlang5/ 1,table to dimension the number of channels
needed in the system
+.1 TRAFFIC AND C'ANNEL DI&ENSIONIN
"ellular system capacity depends on a number of different
factors. These include:
+- The number of channels available for voice and=or data
+- The grade of service the subscribers are encountering in the
system
Traffic theory attempts to obtain useful estimates of, e.g., the
number of channels needed in a cell. These estimates depend on
the selected system and the assumed or real behavior of the
subscribers.
Chat is trafficD Traffic refers to the usage of channels and is
usually thought of as the holding time per time unit (or the
number of Acall hoursAper hour) for one or several circuits
(trun*s or channels). Traffic is measured in Erlangs (E). &or
example, if one subscriber spends all of his=her time on the
telephone, he=she can generate one call per hour or + E of traffic.
!ow much traffic can one cell carryD That depends on the
number of traffic channels available and the acceptable
probability that the system is congested, the so,called .rade of
/ervice (.o/). ifferent assumptions on subscriber behavior
lead to different answers to this 'uestion. Erlang5s (a anish
traffic theorist) 1,table is based on the most common
assumptions used. These assumptions are:
+- 3o 'ueues
+8 3umber of subscribers much higher than number of traffic
channels
+8 3o dedicated (reserved) traffic channels
+8 #oisson distributed (random) traffic
+8 1loc*ed calls abandon the call attempt immediately
This is referred to as a Aloss systemA. Erlang5s 1,table relates
the number of traffic channels, the .o/, and the traffic offered.
This relationship is tabulated in &igure -,+. 0ssuming that one
cell has two carriers, corresponding typically to >x:,>E+8 traffic
channels and a .o/ of >F is acceptable, the traffic that can be
offered is 0E:.>; E (&igure -,+).
This number is interesting if an estimate on the average traffic
per subscriber can be obtained. /tudies show that the average
traffic per subscriber during the busy hour is typically +?,>; mE
(this can correspond to, e.g., one call lasting ?8,@> seconds per
hour). ividing the traffic that one cell can offer, 0cellE:.>; E,
by the traffic per subscriber, here chosen as 0subE;.;>? E, the
number of subscribers one cell can support is derived as
:.>;=;.;>? E ->: subscribers.
n
.;;@ .;;: .;;< .;+ .;> .;- .;? .+
.> .8
n
+
+.;;@;? .;;:;9 .;;<;: .;+;+; .;>;8+ .;-;<- .;?>9- .+++++ .>?;;; .999
+
>
.+>9;; .+-?-> .+88+9 .+?>?< .>>-8@ .>:+?? .-:+-> .?<?8- +.;;;; >.;;;;
>
-
.-<998 .8+@?@ .8-@++ .8??8< .9;>>+ .@+?+- .:<<8; +.>@;: +.<><< -.8@<:
-
8
.@@@>< .:+;>< .:8;:? .:9<8> +.;<>- +.>?:< +.?>89 >.;8?8 >.<8?> ?.;>+;
8
?
+.>-9> +.>:+; +.->>- +.-9;: +.9?@+ +.:@?> >.>+:? >.::++ 8.;+;8 9.?<??
?
9
+.@?-+ +.:;<- +.:9+; +.<;<; >.>@?< >.?8-+ >.<9;- -.@?:8 ?.+;:9 :.+<;@
9
@
>.-+8< >.-:>; >.88-@ >.?;;< >.<-?8 -.>8<@ -.@-@: 8.999> 9.>-;> <.@<<:
@
:
>.<+>? >.<<;> -.;9+? -.+>@9 -.9>@+ -.<:9? 8.?8-; ?.?<@+ @.-9<> ++.8+<
:
<
-.?-<? -.9>@8 -.@;:; -.@:>? 8.-88@ 8.@8@< ?.-@;> 9.?898 :.?>+@ +-.;8?
<
+;
8.+<++ 8.>::< 8.-@:8 8.89+> ?.;:8; ?.?><8 9.>+?@ @.?+;9 <.9:?; +8.9@@
+;
++
8.:9-@ 8.<@;< ?.;9<+ ?.+?<< ?.:8+? 9.->:; @.;@98 :.8:@+ +;.:?@ +9.-+8
++
+>
?.??8- ?.9@;: ?.@@@8 ?.:@9; 9.9+8@ @.+8+; @.<?;+ <.8@8; +>.;-9 +@.<?8
+>
+-
9.>9;@ 9.-:9- 9.?;++ 9.9;@> @.8;+? @.<99@ :.:-8< +;.8@; +-.>>> +<.?<:
+-
+8
9.<:++ @.++?8 @.>-:> @.-?+@ :.>;;- :.:;-? <.@><? ++.8@- +8.8+- >+.>8-
+8
+?
@.@+-< @.:?9: @.<:@8 :.+;:; <.;;<9 <.9?;; +;.9-- +>.8:8 +?.9;: >>.:<+
+?
+9
:.8?@< :.9;<> :.@8@8 :.:@?; <.:>:8 +;.?;? ++.?88 +-.?;; +9.:;@ >8.?8+
+9
+@
<.>++< <.-@+8 <.9+@+ <.9?+9 +;.9?9 ++.-9: +>.89+ +8.?>> +:.;+; >9.+<>
+@
+:
<.<@?+ +;.+8- +;.><9 +;.8-@ ++.8<+ +>.>-: +-.-:? +?.?8: +<.>+9 >@.:88
+:
+<
+;.@8@ +;.<>> ++.;:> ++.>-; +>.--- +-.++? +8.-+? +9.?@< >;.8>8 ><.8<:
+<
>;
++.?>9 ++.@;< ++.:@9 +>.;-+ +-.+:> +-.<<@ +?.>8< +@.9+- >+.9-? -+.+?>
>;
>+
+>.-+> +>.?;- +>.9@@ +>.:-: +8.;-9 +8.::? +9.+:< +:.9?+ >>.:8: ->.:;:
>+
>>
+-.+;? +-.-;- +-.8:8 +-.9?+ +8.:<9 +?.@@: +@.+-> +<.9<> >8.;98 -8.898
>>
>-
+-.<;8 +8.++; +8.><@ +8.8@; +?.@9+ +9.9@? +:.;:; >;.@-@ >?.>:+ -9.+>+
>-
>8
+8.@;< +8.<>> +?.++9 +?.><? +9.9-+ +@.?@@ +<.;-+ >+.@:8 >9.8<< -@.@@<
>8
>?
+?.?+< +?.@-< +?.<-< +9.+>? +@.?;? +:.8:- +<.<:? >>.:-- >@.@>; -<.8-@
>?
>9
+9.--8 +9.?9+ +9.@9: +9.<?< +:.-:- +<.-<> >;.<8- >-.::? >:.<8+ 8+.;<9
>9
>@
+@.+?- +@.-:@ +@.9;+ +@.@<@ +<.>9? >;.-;? >+.<;8 >8.<-< -;.+98 8>.@??
>@
>:
+@.<@@ +:.>+: +:.8-: +:.98; >;.+?; >+.>>+ >>.:9@ >?.<<? -+.-:: 88.8+8
>:
><
+:.:;? +<.;?- +<.>@< +<.8:@ >+.;-< >>.+8; >-.:-- >@.;?- ->.9+8 89.;@8
><
-;
+<.9-@ +<.:<+ >;.+>- >;.--@ >+.<-> >-.;9> >8.:;> >:.++- --.:8; 8@.@-?
-;
-+
>;.8@- >;.@-8 >;.<@> >+.+<+ >>.:>@ >-.<:@ >?.@@- ><.+@8 -?.;9@ 8<.-<?
-+
->
>+.-+> >+.?:; >+.:>- >>.;8: >-.@>? >8.<+8 >9.@89 -;.>-@ -9.><? ?+.;?9
->
F()*+, 3-1 P1+: 39 E+/12)>4 B-:17/,? 6(,/8(2) :-, :+199(. ((2
E+/12)4 14 1 9*2.:(32 39 :-, G3S (.3/*=24 128 2*=7,+ 39
:+199(. .-122,/4 (+3;4
imensioning the networ* now implies using demographic data
to determine the si6es of the cells. The preceding example is
simplified, however, it provides an understanding of what is
meant by traffic and traffic dimensioning.
The problem may be that given a number of subscribers in one
particular area, e.g. an airport, how many carriers do we need to
support the traffic if only one cell is to be usedD imensioning a
whole networ* while maintaining a fixed cell si6e means
estimating the number of carriers needed in each cell. In
addition, traffic is not constant. It varies between day and night,
different days, and with a number of other factors. 7obile
telephony implies mobility and hence subscribers may move
from one area to another during the course of a day.
FOR INTERNAL USE ONLY 1"
It is important that the number of signaling channels (/""!s)
is dimensioned as well, ta*ing into account the estimated system
behavior in various parts of the networ*. &or example, cells
bordering a different location area may have lots of location
updating, and cells on a highway probably have many
handovers. In order to calculate the need for /""!s the
number of attempts for every procedure that uses the /""! as
well as the time that each procedure holds the /""! must be
ta*en into account. The procedures areG location updating,
periodic registration, I7/I attach=detach, call setup, /7/,
facsimile and supplementary services. The number of false
accesses must also be estimated. This is typically 'uite a high
number, but still small compared to the traffic.
Chen the .o/ that should be used to consult the traffic tables is
chosen, the fact that calls go through two different devices must
be *ept in mind.
+.2 C'ANNEL UTILI2ATION
0ssume the tas* is to find the necessary number of traffic
channels for one cell to serve subscribers with a traffic of -- E.
The .o/ during the busy hour is not to exceed >F. 1y
considering the above re'uirements and consulting Erlang 1table,
8- channels are found to be needed (&igure -,>).
n
.;;@ .;;: .;;< .;+ .;> .;- .;?
.+ .> .8
n
8-
-;.@-8 -+.;9< -+.-@8 -+.9?9 --.@?: -?.>?- -@.?9? 8>.;++ 8<.:?+ 9<.-8>
8-
F()*+, 3-2 P1+: 39 E+/12)>4 B-:17/, 93+ 43 .-122,/4 )(@(2) :-,
399,+,8 :+199(. (E 14 1 9*2.:(32 39 :-, G3S (A
0ssume five cells are designed to cover the same area as the
single cell. These five cells must handle the same amount of
traffic as the cell above, -- E. 0cceptable .o/ is still >F. &irst,
the total traffic is divided among the cells (&igure -,-). Traffic
distribution over several cells results in a need for more
channels than if all traffic had been concentrated in one cell.
FOR INTERNAL USE ONLY 1%
This illustrates that it is more efficient to use many channels in a
larger cell than vice versa. To calculate the channel utili6ation,
the traffic offered is reduced by the .o/ of >F (yielding the
traffic served) and dividing that value by the number of channels
(yielding the channel utili6ation).
Cith 8- channels (as in the previous single cell example) the
channel utili6ation is --.;:- = 8- E @@F, i.e. each channel is
used approximately @@F of the time. !owever, by splitting this
cell into smaller cells, more traffic channels are re'uired and
hence the channel utli6ation decreases
cell traffic(F) traffic(E) 3o.of channel "hannel
utili6ation(F)
0 8; +-.> >+ 9>
1 >? :.>? +? ?8
" +? 8.<? +; 8<
+; -.- : 8;
E +; -.- : 8;
T
+;; -- 9>
F()*+, 3-3 W-1: -100,24 ;-,2 1 .,+:1(2 1=3*2: 39 :+199(. (4
8(4:+(7*:,8 3@,+ 4,@,+1/ .,//4B
0s we will see in the following chapter, capacity and
interference problems prevent us from always using the most
effective channel utili6ation scheme and so solutions in real
networ*s must compromise between efficiency (i.e., cost) and
'uality.
,. No#inal Cell Plan
C-a$ter )
of system balancing, channel planning, and basic radio concepts.
O./EC'IVES0
+: escribe the *ey terms when relating to cell structure
+: Explain the T70 concept
+: Explain how to balance a cellular system, e.g. to be able to
set the output power
+: escribe the most common re,use patterns and their channel
plans
+: Explain briefly why interference occurs
+: iscuss general properties of electromagnetical waves
+: escribe how radio waves are generated
+: escribe how information is superimposed on radio waves
+: escribe radio wave propagation and attenuation
+: escribe the pathloss concept without using mathematical
formulas
+: escribe the origin of some fast signal variations
FOR INTERNAL USE ONLY 1!
,.1 INTRODUCTION
"ellular networ* engineering encompasses all wor* re'uired to
design a cellular (radio base station) networ*. uring the initial
phase of a system design, the system re'uirements are collected
and analy6ed. These include cost, capacity, coverage, .o/,
speech 'uality, and system growth capability. (The traffic related
factors were discussed in the previous chapter.) Chen the
system re'uirements phase is complete, it is time to prepare a
nominal cell plan. This plan covers the distribution (location)
and configuration of radio base stations and is based on the
system re'uirements. The nominal cell plan must
later be verified so that it is as accurate as possible. $nce the
system design has been implemented, cell planning wor*
continues using data from the existing networ*.
,.2 S%STE& BALANCIN
0n area is referred to as being covered if the signal strength
received by an 7/ in that area is higher than some minimum
value. 0 typical value in this case is around ,<; d1m (+ pC).
!owever, coverage in a two,way radio communication system is
determined by the wea*est transmission direction. 1oth uplin*
and downlin* are ta*en into consideration here. That is, the
signal received by the 1T/ from an 7/ in an area must be
higher than some minimum value. It ma*es no sense to have
different coverage on uplin* and downlin* because this causes
an excess amount of energy to be dissipated into the system
adding extra interferences and costs. 0 system balance must be
obtained before coverage calculation can start.
To achieve this balance it is necessary to ma*e sure that the
sensitivity limit, 7/(sens), of the 7/ (for downlin* transmission)
is reached at the same point as the sensitivity limit, 1T/(sens), of
the 1T/ (for uplin* transmission).
FOR INTERNAL USE ONLY 1&
F()*+, 4-2 S.-,=1:(. )+10- 39 :-, .3=032,2:4 (2./*8,8 (2 1
464:,= 71/12.,. A77+,@(1:(324 -1@, :-, 93//3;(2) :+124/1:(324C
GDG1(2? LDL344? ADA2:,221? FDF,,8,+? CDC3=7(2,+?
MSDM37(/, S:1:(32? BTSDB14, T+124.,(@,+ S:1:(32?
DDD(@,+4(:6? P(2D(20*: 03;,+? P3*:D3*:0*: 03;,+? 128 L0D01:-
/344
The input power, Pin(ms), at the 7/ receiver e'uals the output
power, Pout(bts), of the 1T/ plus gains and losses. If input power
is set e'ual to the sensitivity level, a system balance can be
found.
+The 1T/ output power should never be changed once the system
is balanced for a particular configuration and mobile class.
3ote: Asmaller cellsADare desired, the power can be decreased
because it can be matched by a corresponding, forced, decrease
in the output power of the 7/.
,.+ C'ANNEL LOADIN PLAN
The simplest cell planning problem solution is to have one cell
and use all available carriers in that cell (&igure 8,-). !owever,
such a solution has severe limitations. It is seldom that coverage
can be maintained in the entire area desired. In addition, even
though the channel utili6ation may be very high, limited capacity
soon becomes a problem due to the limited number of carriers
available to any operator.
0 cellular system is based upon re,use of the same set of
fre'uencies which is obtained by dividing the area needing
coverage into smaller areas (cells) which together form clusters
(&igure 8,8). 0 cluster is a group of cells in which all available
carriers have been used once (and only once). /ince the same
carriers are used in cells in neighboring clusters, interference
may become a problem. Indeed, the fre'uency re,use distance,
i.e. the distance between two sites using the same carrier, must
be *ept as large as possible from a interference point,of,view.
0t the same time they must be *ept as small as possible from a
capacity point of view.
F()*+, 4-3 EE1=0/, 39 12 1+,1 4,+@,8 9+3= 32, .,// 76 24 .1++(,+4
F()*+, 4-4 T-, 41=, 1+,1 14 (2 F()*+, 4-3
7*: 23; 4.-,=1:(.1//6 8(@(8,8 (2:3 93*+ ./*4:,+4? ,1.- ./*4:,+
*4(2) 1//
(-,+, 24 .1++(,+4. T-, 4=1// .(+./,4 (28(.1:, (28(@(8*1/ .,//4
;-,+, :-, 9+,5*,2.6 91 (4 *4,8 128 1 8(4:12., 7,:;,,2 :-,
.3++,40328(2) 4(:,4? <23;2 14 9+,5*,2.6 +,-*4, 8(4:12.,? (4
(28(.1:,8 76 :-, 83*7/, 1++3;.
IN'ERFERENCE
"ellular system are often interference limited rather than signal
strength limited. Therefore some elementary information about
different problems associated with the re,use of carriers is
provided in this section.
F()*+, 4-" C3-.-122,/ (2:,+9,+,2.,
"o,channel interference is the term used for interference in a
cell by carriers with the same fre'uency present in other cells.
&igure 8,? illustrates the situation. /ince the same carrier
fre'uency is used for the wanted carrier as for the unwanted
carrier, 'uality problems can arise if the signal from the
unwanted carrier is too strong.
The ./7 specification states that the signal strength ratio, "=I,
between the carrier, ", and the interferer, I, must be larger than <
d1. !owever, !uawei recommends using "=I H+> d1 as a
planning criterion. If fre'uency hopping is implemented, it adds
extra diversity to the system corresponding to a margin of
approximately - d1, i.e.:
"=I H +> d1 (without fre'uency hopping)
"=I H < d1 (with fre'uency hopping)
0djacent carrier fre'uencies (i.e., fre'uencies shifted I>;; *!6)
with respect to the carrier cannot be allowed to have too strong a
signal strength either. Even though they are at different
fre'uencies, part of the signal can interfere with the wanted
carrier5s signal and cause 'uality problems (&igure 8,9). The
./7 specification states that the signal strength ratio, "=0,
between the carrier and the adjacent fre'uency interferer, 0,
must be larger than ,< d1. !owever, adjacent channel
interference also degrades the sensitivity as well as the "=I
performance. uring cell planning the aim should be to have
"=0 higher than - d1, according to !uawei, i.e.:
"=0 H ,9 d1
0djacent fre'uencies must be avoided in the same cell and
preferably in neighboring cells as well.
F()*+, 4-% A8F1.,2: .-122,/ (2:,+9,+,2.,
1y re,using the carrier fre'uencies according to well,proven reuse
patterns (&igure 8,@ and &igure 8,:), neither co,channel
interference nor adjacent channel interference will cause
problems, provided the cells have isotropic propagation
properties for the radio waves. Unfortunately this is hardly ever
the case. "ells vary in si6e depending on the amount of traffic
they are expected to carry, and nominal cell plans must be
verified by means of predictions or radio measurements to
ensure that interference does not become a problem.
The re,use patterns recommended for ./7 are 8=+>, and -=<,
patterns. 8=+> means that each cluster has four three,sector sites
supporting twelve cells (&igure 8,@).
F()*+, 4-7 4G12 +,-*4, 01::,+2
The re,use pattern in &igure 8,@ is compatible with the condition
"=IH+> d1. 0 shorter re,use distance, given a smaller "=I,ratio,
is used in the -=<,pattern (&igure 8,:).
F()*+, 4-! 3G& +,-*4, 01::,+2
This re,use pattern (&igure 8,:) is recommended only if
fre'uency hopping is implemented. It has a higher channel
utili6ation because the carriers are distributed among nine cells
rather than +>. $ther re,use patterns with much higher re,use
distances (such as the @=>+) must be used for system which are
more sensitive to interferenceG e.g. analog mobile telephone
system.
IN'ERS1M.OL IN'ERFERENCE 2ISI5
Inter/ymbol Interference (I/I) is caused by excessive time
dispersion. It may be present in all cell re,use patterns. I/I can
be thought of as co,channel interference. !owever in this case
the interferer, %, is a time delayed reflection of the wanted
carrier. 0ccording to ./7 specifications, the signal strength
ratio "=% must be larger than < d1 (compared to the "=I criterion).
!owever, if the time delay is smaller than +? s (i.e.,
8 bits or approximately 8.8 *m), the e'uali6er can solve the
problem. I/I is not affected by the re,use pattern chosen, but is
still an issue for the cell planner.
!ow can the cell planner avoid I/I in the cellular networ*D
3ormally, the reflected waves are much wea*er than the direct
wave. !owever, if the direct wave is obstructed (shadowed), or
+if the reflected wave has a very advantageous path of
propagation, the "=% ratio may creep down to dangerous values
if the time delay is outside the e'uali6er window. !ence, time
dispersion may cause problems in environments with, e.g.,
mountains, la*es with steep or densely built shores, hilly cities,
+and high metal,covered buildings. The location of the 1T/ can
thus be crucial. &igure 8,< and &igure 8,+; suggest some
possible solutions.
F()*+, 4-& L3.1:(2) :-, BTS ./34, :3 :-, +,9/,.:(2) 37F,.: :3
.3=71: ISI
F()*+, 4-1# P3(2:(2) :-, 12:,221 1;16 9+3= :-, +,9/,.:(2)
37F,.: :3 .3=71: ISI
,., )A3ES
There are many seemingly different types of electromagnetic
waves. They include radio waves, infrared rays, light, x,rays,
and gamma rays among others. %adio waves are one type of
electromagnetic radiation. They are typically generated as
disturbances sent out by oscillating charges on a transmitting
antenna. $ther types of electromagnetic radiation are caused by
intense heat, atomic reactions, and stimulated emission (lasers).
%egardless of its origin, an electromagnetic wave is comprised
of oscillating electric and magnetic fields. &or a simple,
traveling, plane wave, the electric and magnetic fields are
perpendicular to each other and also to the direction of
propagation. Caves can be described by simple sinusoidal
functions (&igure 8,++) and are conveniently characteri6ed by
their wavelength, , the length of one cycle of oscillation. This
can be calculated as:
F()*+, 4-11 A2 ,/,.:+3=1)2,:(. 0/12, ;1@, H9+3I,2HB(2 :(=,
#ropagation properties are different across the fre'uency
spectrum. %adio waves fall in the fre'uency spectrum between -
!6 and -;;; .!6. This part of the spectrum is divided into
twelve bands (&igure 8,+>). $nly the Ultra !igh &re'uency
(U!&) band is considered from now on, since properties of U!&
waves and fre'uency allocations have made this the mobile
telephony fre'uency band.
F()*+, 4-12 F+,5*,2.6 40,.:+*= 71284
,./ ENERATION OF RADIO )A3ES
!igh fre'uency radio waves are typically generated by
oscillating charges on a transmitting antenna. In the case of a
radio station, the antenna is often simply a long wire (a dipole)
fed by an alternating voltage=current sourceG i.e., charges are
placed on the antenna by the alternating voltage source. Ce can
thin* of the electric field as being disturbances sent out by the
dipole source and the fre'uency of the oscillating electric field
(the electromagnetic wave) is the same as the fre'uency of the
source.
Each antenna has a uni'ue radiation pattern. This pattern can be
represented graphically by plotting the received, time,averaged
power, as a function of angle with respect to the direction of
maximum power in a log,polar diagram. The pattern is
representative of the antenna5s performance in a test
environment. !owever, it only applies to the free,space
environment in which the test measurement ta*es place. Upon
installation, the pattern becomes more complex due to factors
affecting propagation in the reality. Thus, the real effectiveness
of any antenna is measured in the field.
0n isotropic antenna is a completely non,directional antenna
that radiates e'ually in all directions. /ince all practical antennas
exhibit some degree of directivity, the isotropic antenna exists
only as a mathematical concept. The isotropic antenna can be
used as a reference to specify the gain of a practical antenna (see
the appendix for a general discussion on gain=loss and
logarithmic units). The gain of an antenna referenced
isotropically is the ratio between the power re'uired in the
practical antenna and the power re'uired in an isotropic antenna
to achieve the same field strength in the desired direction of the
measured practical antenna. irective gain in relation to an
isotropic antenna is expressed in units of Ad1iA
0 half,wave dipole antenna may also be used as a gain reference
for practical antennas. The half,wave dipole is a straight
conductor cut to one,half of the electrical wavelength with the
radio fre'uency signal fed to the middle of the conductor.
&igure 8,+- illustrates the radiation pattern of the half,wave
dipole which normally is referred to as a dipole. Chereas the
isotropic antenna5s three dimensional radiation pattern is
spherical, the dipole antenna5s three dimensional pattern is
shaped li*e a donut.
irective gain in relation to a dipole is expressed in units of
Ad1dA &or a dipole and an isotropic antenna with the same
input power, the energy is more concentrated in certain
directions by the dipole. The difference in directive gain
between the dipole and the isotropic antenna is >.+? d1.
&igure 8,+8 illustrates the differences in gain between the
isotropic, dipole and practical antenna. The vertical pattern
(&igure 8,+8) for the practical antenna is that of a directional
F()*+, 4-13 D(03/, +18(1:(32 01::,+2
F()*+, 4-14 G1(2 .3=01+(432
F()*+, 4-1" V,+:(.1/ 128 -3+(I32:1/ 12:,221 01::,+24 93+ 1 H+,1/H
12:,221
Chen choosing an antenna for a specific application, the
manufacturer5s data sheet must be consulted. The data sheet
contains information including antenna gain, beamwidth
(vertical and hori6ontal), and graphs showing the vertical and
hori6ontal patterns. Examples of the graphs normally found in a
data sheet are shown in &igure 8,+?. The patterns displayed are
those of a directional antenna. The antenna5s gain is
approximately +? d1d.
The beamwidth, 1, is defined as the opening angle between the
points where the radiated power is - d1 lower than in the main
direction (&igure 8,+9). 1oth the hori6ontal and vertical
beamwidths are found using the - d1 down points, alternatively
referred to as half,power points.
F()*+, 4-1% D,9(2(:(32 39 7,1=;(8:-
,.0 SUPERI&POSIN INFOR&ATION ON RADIO )A3ES
Information is seldomly transmitted in the same fre'uency range
as it was generated. The reason is that if, as an example, we
want to broadcast a > *!6 signal, the antenna would have to be
@? *m long (half a wavelength). !owever, by translating the
signal to a much higher fre'uency band (e.g., the U!& band of
cellular telephony) antenna si6es drop to a few decimeters. In
addition, in order to have numerous AchannelsA simultaneously,
a higher fre'uency is re'uired. &re'uency translation is
implemented by modulating the amplitude, fre'uency or phase
of a so,called carrier wave in accordance with the wave form of
the wanted signal. /everal modulation schemes exist (e.g.,
amplitude modulation) common for analog radio signals and
phase modulation. 0ny modulation scheme increases the carrier
bandwidth and hence limits the capacity of the fre'uency band
available. /ince the bandwidth of the carrier increases if the bit
rate increases, a high carrier fre'uency is necessary to obtain
many different AchannelsA The cell planner cannot choose
modulation techni'ues, but the conse'uences of the system
choice are very important, since carrier bandwidth and carrier
separation affects, e.g., interference properties. Cave
propagation also behaves differently in different fre'uency
bands.
The modulation techni'ue used in ./7 <;; is called .aussian
7inimum /hift Jeying (.7/J). This narrow,band digital
modulation techni'ue is based on phase shifting. That is, bits are
represented by continuous positive or negative phase shifts. 1y
changing the phase continuously, sharp discontinuities are
avoided, thus narrowing the bandwidth of the modulated carrier.
.7/J modulation also involves filtering the incoming bit
stream with a .aussian filter to obtain a more narrow bandwidth
of the modulated carrier. In fact the full width at half maximum
of the carrier becomes +9> *!6, corresponding nicely to the >;;
*!6 carrier separation.
Transmitting the information on the air interface in digiti6ed
form has an advantage over analog techni'ues, since channel
coding protects bits, the signal is less sensitive to perturbations.
In addition, it enables Time ivision 7ultiple 0ccess (T70)
which means that one carrier fre'uency can be used for several
connections. Each connection uses only one particular time slot
(out of the eight available in ./7). This has the advantage that
the mobile is released from transmitting=receiving continuously
and can perform, e.g., measurements on neighboring cells. $ne
main advantage with T70 is that it enables 7obile 0ssisted
!and $ver (70!$) which is essential for effective connection
control.
,.4 AIR INTERFACE DATA
1elow is a summary of some important air interface data for
./7 <;;, ./7 +:;;, and ./7 +<;;.
FRE34ENC1 SPEC'R4M
ifferent fre'uency bands are used for ./7 <;;, ./7 +:;;,
and ./7 +<;; (refer to &igure 8,>). In some countries,
operators apply for the available fre'uencies. In other countries
e.g., the United /tates), operators purchase fre'uency bands at
auctions.
In ecember of +<<8, the &ederal "ommunications "ommission
(&"") auctioned AbroadbandA licenses to prospective operators
offering personal communications services. Each operator owns
the rights to the licenses for a period of ten years. The United
/tates is divided into ?+ regions or 7ajor Trading 0reas (7T0)
and 8<- 1asic Trading 0reas (1T0). The &"" issued two
./7 +<;; licenses for each 7T0 and four for each 1T0. $ne
7T0 can be geographically as large as a state, while one 1T0
can be compared in si6e to a large city. 1T0s are designed for
use in major metropolitan areas.
The &"" has specified the fre'uency range and output power.
The fre'uency band is divided into six fre'uency bloc*s
(&igure 8,+@): three duplex bloc*s 0, 1, and " (<; 7!6 total
spectrum bandwidth) and three other duplex bloc*s , E, and &
(-; 7!6 total spectrum bandwidth).
F()*+, 4-17 S0,.:+*= 1//3.1:(32 93+ GSM 1&## (2 U2(:,8
S:1:,4. 14# MHI 93+ GSM 1&## (12# MHI /(.,24,8 128 2# MHI
*2/(.,24,8
#4PLE6 #IS'ANCE
The distance between the uplin* and downlin* fre'uencies is
*nown as duplex distance. The duplex distance is different for
the different fre'uency bands (&igure 8,+:).
F()*+, 4-1! D*0/,E 8(99,+,2.,4 93+ 8(99,+,2: 9+,5*,2.6 71284
C7ANNEL SEPARA'ION
The distance between adjacent fre'uencies on the uplin* or the
downlin* is called channel separation. The channel separation is
>;; *!6, regardless of the standard chosen from the ones
mentioned above. This separation is needed to reduce
interference from one carrier to another neighboring fre'uency.
ACCESS ME'7O# AN# 'RANSMISSION RA'E
./7 has chosen the T70 concept for access. In ./7, there
are eight T70 time slots per frame (&igure 8,+<). Each time
slot is ;.?@@ ms long and has room for +?9.>? bits (+8: bits of
+information and a :.>? bits long guard period) yielding a bit rate
on the air interface of >@;.: *bits.
F()*+, 4-1& B14(. TDMA 9+1=,? :(=,4/3:? 128 7*+4: 4:+*.:*+,4
,.5 RADIO )A3E PROPAATION
In this course we are primarily interested in the transmission loss
between two antennas: the transmitter=emitter and the receiver.
7any factors including absorption, refraction, reflection,
diffraction, and scattering affect the wave propagation.
!owever, in free space an electromagnetic wave travels
indefinitely if unimpeded. This does not mean that there are no
transmission losses, as we will see in this first simple model
where isotropic emission from the transmitter and line of sight
between the two antennas separated by a distance, G, in free
space are assumed (&igure 8,>;).
F()*+, 4-2# R18(3 ;1@, 0+301)1:(32 (2 9+,, 401.,
3ote that the wavelength dependency of the pathloss does not
correspond to losses in free space as such. It is a conse'uence of
the finite effective receiver area.
This expression is fairly general. The only thing which changes
when we improve our models is the expression for the pathloss.
The antenna gain is normally given in d1(i), i.e., as +;log(G),
where gain means a reduction of the total transmission loss, L,
between a transmitting and receiving antenna.
This model helps us to understand the most important features
of radio wave propagation. That is, the received power decreases
when the distance between the antennas increases and the
transmission loss increases when the wavelength decreases (or
alternatively when the fre'uency increases).
&or cell planning, it is very important to be able to estimate the
signal strengths in all parts of the area to be covered, i.e. to
predict the pathloss. The model described in this section can be
used as a first approximation. !owever, more complicated
models exist. Improvements can be made by accounting for:
the fact that radio waves are reflected towards the earth5s
surface (the conductivity of the earth is thus an important
parameter) transmission losses due to obstructions in the line of sight
the finite radius of the curvature of the earth the topographical
variations in a real case as well as the different attenuation properties
of different terrain types such as forests, urban areas, etc.
The best models used are semi,empirical, i.e., based on
measurements of pathloss=attenuation in various terrains. The
use of such models are motivated by the fact that radio
propagation can not be measured everywhere. !owever, if
measurements are ta*en in typical environments, the parameters
of the model can be fine,tuned so that the model is as good as
possible for that particular type of terrain.
,.6 SINAL 3ARIATIONS
The models described in the previous section can be used to
estimate the average signal level (called the 5global meanA) at
the receiving antenna. !owever, a radio signal envelope is
composed of a fast fading signal super,imposed on a slow fading
signal (&igure 8,+:). These fading signals are the result of
obstructions and reflections. They yield a signal which is the
sum of a possibly wea*, direct, line,of,sight signal and several
indirect or reflected signals.
The fast fading signal (pea*,to,pea* distance ==>) is usually
present during radio communication due to the fact that the
mobile antenna is lower than the surrounding structures such as
trees and buildings. These act as reflectors. The resulting signal
consists of several waves with various amplitudes and phases.
/ometimes these almost completely cancel out each other. This
can lead to a signal level below the receiver sensitivity. In open
fields where a direct wave is dominating, this type of fading is
less noticeable.
/hort,term fading is %ayleigh distributed with respect to the
signal voltage. Therefore, it is often called %ayleigh fading. This
type of fading affects the signal 'uality, and as a result some
measures must be ta*en to counter it.
The first and most simple solution is to use more power at the
transmitter(s), thus providing a fading margin. 0nother way to
reduce the harm done by %ayleigh fading is to use space
diversity, which reduces the number of deep fading dips.
iversity means that two signals are received which have
slightly different 5historiesA and, therefore, the AbestA can be
used. Cith the efficient space diversity system used by !uawei,
voice 'uality is much improved.
F()*+, 4-21 S-3+:-:,+= (914: 128 /32)-:,+= (4/3; 918(2)
The signal variation received if we smooth out the short,term
fading is called the Alocal meanA Its power is often called the
local average power, is expressed in a logarithmic scale, and is
normally distributed. Therefore, this slow fading is called Alognormal
fadingA If we drive through a flat desert without any
obstructions, the signal varies slowly with distance. !owever, in
normal cases the signal path is obstructed.
$bstructions near the mobile (e.g., buildings, bridges, trees, etc.)
cause a rapid change of the local mean (in the range of five to
fifty meters), while topographical obstructions cause a slower
signal variation. 1ecause log,normal fading reduces the average
strength received, the total coverage from the transmitter is
reduced. To combat this, a fading margin must be used.
#roblems generated by multi,path reflections are made more
severe by log,normal fading since the direct beam is wea*ened
by the obstructing object.
#hases between various reflected waves are different. This is
due to the fact that they propagate over different distances or
e'uivalently use different times to reach the receiver. This time
dispersion can cause particular problems if the phase difference
between the reflected waves is very large. &or ./7 <;;, a large
phase difference is on the order of several thousands of
wavelengths (i.e. one *ilometer or more). In this case, different
waves added together in the receiver carry information about
different symbols (bits). If the direct wave is wea*, and
conse'uently the reflected waves are relatively strong, it can be
difficult to determine which symbol (bit) was transmitted.
/. Sur7e!"
C-a$ter +
of radio networ* survey of a cellular networ* as well as some
radio measurements.
O./EC'IVES0
Explain briefly what a site survey is and what to consider
during a survey
escribe three different types of radio measurements
/.1 INTRODUCTION
The cell planning process results in a cell plan with nominal site
positions. If the operator has access to existing locations, it is
necessary to adapt the cell plan according to these locations. &or
this reason, it is important that the cell planner has a basic
*nowledge of the locations that can be used.
The on,site cell planning wor* that ta*es place is called the
Aradio 3etwor* /urveyA.This is described in the following
section. 0 more detailed survey is performed on the base station
sites. This is called the Asite investigationA and is not discussed
in this course.
/.2 RADIO NET)OR* SUR3E%
.ASIC CONSI#ERA'IONS
It is li*ely that the system operator has a number of alternative
buildings which may be used in the cellular networ* planning
phase. $ne reason for this is to reduce the initial cost.
The following aspects of site selection must be studied:
8; #osition relative to nominal grid
8; /pace for antennas
8; 0ntenna separations
8; 3earby obstacles
8+ /pace for radio e'uipment
8+ #ower supply=battery bac*up
8+ Transmission lin*
8+ /ervice area study
8+ "ontract with the owner
POSI'ION RELA'IVE 'O NOMINAL GRI#
The initial study for a cell system often results in a theoretical
cell pattern with nominal positions for the site locations. The
existing buildings must then be adapted in such a way that the
real positions are established and replace the nominal positions.
The visit to the site is to ensure the exact location
(address=coordinates and ground level). It is also possible for
more than one existing site to be used for a specific nominal
position.
SPACE FOR AN'ENNAS
The radio propagation predictions provide an indication on what
type of antennas can be used on the base station and in what
direction the antennas should be oriented.
The predicted antenna height should be used as a guideline when
the on,site study starts. If space can be found within a maximum
deviation of +?F from the predicted height the original
predictions can be used with sufficient accuracy.
If it is possible to install the antennas at a higher position than
the predicted position, the operator must ensure that there is no
ris* for co,channel interference. If the antennas are to be
installed at a lower position than predicted, new predictions
must be carried out based on this position.
It is not necessary that all antennas in one particular cell have
the same height or direction. That is, it is possible to have cells
on the same base station with different antenna heights. This can
be the case if space is limited in some directions. There are also
cell planning reasons for placing antennas at different heights.
These include coverage, isolation, diversity, and=or interference.
3ote: /ome of these considerations are discussed in the next
section.
AN'ENNA SEPARA'IONS
There are two reasons for antennas to be separated from each
other and from other antenna system:
8> To achieve space diversity
8> To achieve isolation
The hori6ontal separation distance to obtain sufficient space
diversity between antennas is +>,+: or 8,9 meter for ./7 <;;
and >,- m for ./7 +:;;=+<;;. Typical values of separation
distances between antennas to obtain sufficient isolation
(normally -; d1) are ;.8 m (hori6ontal) and ;.> m (vertical) for
./7 <;;.
NEAR.1 O.S'ACLES
$ne very important part in the %adio 3etwor* /urvey is to
classify the close surroundings with respect to influence on radio
propagation. In traditional point,to,point communication
networ*s, a line,of,sight path is re'uired. 0 planning criterion is
to have the first fresnel 6one free from obstacles. (3$TE: The
fresnel 6one is the area in open space that must be practically
free of obstructions for a microwave radio path to function
properlyG some degree of fresnel consideration is re'uired in the
immediate vicinity of the microwave radio %& envelope=field.)
It is not possible to follow this guideline because the path
between the base and the mobile subscriber is normally not line,of,
sight. In city areas, one cell planning criterion is to provide
margins for these types of obstacles.
If optimal coverage is re'uired, it is necessary to have the
antennas free for the nearest ?;,+;; m. The first fresnel 6one is
approximately five meters at this distance (for <;; 7!6). This
means the lower part of the antenna system has to be five meters
above the surroundings.
SPACE FOR RA#IO E34IPMEN'
%adio e'uipment should be placed as close as possible to the
antennas in order to reduce the feeder loss and the cost for
feeders. !owever, if these disadvantages can be accepted, other
locations for the e'uipment can be considered. In addition,
sufficient space should be allotted for future expansions.
The radio networ* survey includes a brief study with respect to
this matter. 0 more detailed analysis ta*es place when the
location is chosen to be included in the cellular networ*.
POWER S4PPL18.A''ER1 .AC94P
The e'uipment power supply must be estimated and the
possibility of obtaining this power must be chec*ed. /pace for
battery bac*,up may be re'uired.
'RANSMISSION LIN9
The base station must be physically connected to the 1/". This
can be carried out via radio lin*, fiber cable, or copper cable.
etailed transmission planning is not included in this course.
SERVICE AREA S'4#1
uring the networ* survey it is important to study the intended
service areas from the actual and alternate base station locations.
"overage predictions must be chec*ed with respect to critical
areas.
CON'RAC' WI'7 '7E OWNER
The necessary legal documentation must exist between the land
owner and the proposed site user, e.g., a contract for site leasing.
Even though cost is a major consideration in the site ac'uisition
process, cost is not discussed as a factor in this course.
/.+ RADIO &EASURE&ENTS
PA'7 LOSS PARAME'ERS
0 radio survey involves installation of a transportable test
transmitter somewhere in the area where the base station is to be
installed. Using a specially e'uipped vehicle, signal strength can
be measured. &or this purpose, !uawei has designed a
computeri6ed measurement system. 0 locating unit, a measuring
receiver with antenna, a control and processing unit, and a tape
recorder are among the e'uipment contained in the unit. /ignal
level can be measured on a number of channels and, for each
channel, samples are ta*en at an adjustable speed. 3ormally,
samples are ta*en several times per wavelength traveled. The
data is pre,processed before it is stored on either the hard drive
or a dis*ette and presented off,line after the survey. %esults can
be presented with respect to median value, standard deviation,
and number of Ameasuring s'uaresA along the test routes. The
recorded files can be imported into 0//ET and displayed on the
map. The residual values (i.e., the difference between the
prediction and the measurement) can also be displayed. If there
is a difference, the path loss parameters in the prediction model
can be adjusted according to the measurements.
'IME #ISPERSION
7easurements must be performed to verify the time dispersion
predictions. In addition, if there are 'uality problems, time
dispersion measurements must be ta*en to verify that time
dispersion is actually causing the poor 'uality.
The e'uipment used for time dispersion measurements consists
of a transmitter and a receiver (&igure ?,+). The transmitter
sends a short pulse, the signal is received, and the pulse response
is evaluated in a controller (&igure ?,>). In this way, the time
delay and the carrier to reflection ratio can be found.
F()*+, "-1 T(=, 8(40,+4(32 =,14*+,=,2: ,5*(0=,2:
F()*+, "-2 I=0*/4, +,40324,
IN'ERFERING 'RANSMI''ERS
&or sites where a number of other radio transmitters are colocated,
!uawei recommends that radio spectrum
measurements and a subse'uent interference analysis be
performed. !uawei has developed special e'uipment and
methods for this purpose. These include a computer controlled
spectrum analy6er and computer programs for calculating
interference levels at different fre'uencies. The end result of a
radio spectrum measurement is to accept the site from an
interference point of view, to accept it with reservations, or to
reject the site and find another one.
0. S!"te# Tuning
C-a$ter %
This chapter is designed to provide the student with a general
understanding of networ* optimi6ation.
O./EC'IVES0
89 Explain the reasons for optimi6ation of the radio networ*
89 Explain briefly some !uawei tools used for diagnosing the
networ*
89 Explain briefly how parameter adjustment affects the
networ*
89 escribe why system growth affects cell planning
0.1 INTRODUCTION
/ystem tuning means analy6ing the traffic data collected by the
system to better adjust the system to the actual traffic demand
distribution. 0djustments that can be made include:
8@ "hanging handover parameters to move traffic from a
congested cell to a neighboring cell with a low traffic load
8@ "hanging switch parameters to optimi6e the traffic handling
capacity of the system
8@ 0dding cells or adding radio channels to congested cells
and=or reducing the number of radio channels in cells with
lower traffic than expected
/ystem tuning (i.e., diagnosing the networ* and tuning it) often
ta*es place in:
8@ Initial tuning
8@ %adio 3etwor* Investigation (%3I)
Initial tuning is the tuning that ta*es place either as part of the
acceptance test with a customer or just prior to the acceptance
test. This means that there is usually no (or very little) traffic in
the system. 0 %3I can ta*e place when a system has been
commissioned and used for commercial traffic for some time. It
is then possible to collect statistics in the different networ*
elements from the /tatistics and Traffic measurements
/ubsystem (/T/). These statistics are used in the %3I. !uawei
recommends that %3Is be performed on a regular basis.
/ome of the tools that can be used in system diagnostics are
discussed briefly in the next section. This is followed by a short
overview of the use of parameter adjustment as a means of
system tuning.
0.2 S%STE& DIANOSTICS
OSS
$peration and /upport /ystem ($//) can be used, e.g. to
present system diagnostic information as statistics in graphs.
The /T/ data is transferred to $// where it is stored in a
database. In $//, the data can be displayed in different reports
that illustrate networ* performance regarding, e.g., .o/ in the
cells.
$// can also be used to present measurements collected by
7obile Traffic %ecording , "ell Traffic %ecording, and "hannel
Event %ecording . These are bloc*s located in the 1/" exchange but
accessible from $//. In the
graphical reports we can view, e.g. signal strength, 'uality, T0,
and 7/ power

AN'
03T is !uawei5s Test 7obile /ystem for measuring the
radio environment. 03T consists of a mobile station with
special software, a portable #", a transmitter, and a receiver.
F()*+, %-1. ANT -1+8;1+,
03T can be used, e.g., in a vehicle which drives around the
networ* to analy6e the air interface.

ASSE'
0//ET , as you may remember from "hapter + , is a tool that
!uawei is mar*eting and using for predicting radio propagation.
It runs from a windows platform and interfaces with a graphical
windows environment (&igure 9,>).
0lthough it is not a networ* diagnosing tool itself, it is an
important tool used for optimi6ing networ* performance.
F()*+, %-2. G+10-(.1/ U4,+ I2:,+91.,
1y adding the diagnostic information
generated by tools li*e 03T, 0//ETKs theoretical predictions can
be modified and improved with practical measurements from the
field.
It is important when a system is new, i.e. in an initial tuning
phase, to ma*e sure that the mobiles behave as planned and that
area coverage is what is expected.
The operator can use 03T to monitor where handovers occur.
If the predicted cell borders do not correspond to the data
obtained by 03T, the operator may need to ma*e adjustments,
e.g., changing cell parameters or reconstructing the entire
networ* (or parts of it).
CELL4LAR NE'WOR9 ANAL1:ER 2CeNA5
The "ellular 3etwor* 0naly6er is a cellular 'uality
information system that enables optimi6ation of a digital
networ*5s performance. The system consists of 7obile units
that are mounted in vehicles and assigned subscriber
numbers in the cellular networ*. 0 subscriber number in the
public networ* is assigned to a fixed unit .In accordance
with a measurement order, 7TUs regularly call an &TU, execute
measurements, and transfer the results to a database for storage
and transition .0ll measurement,order setups, result
presentations, and report generations are executed from an
operator terminal as seen in &igure 9,-.
F()*+, %-3 C,//*/1+ N,:;3+< A21/6I,+ (C,NA
FOR INTERNAL USE ONLY "#
0.+ CELL PARA&ETER AD8UST&ENT
If measurement analysis shows an inconsistency in the
parameter setting, hysteresis and offset parameters can be tuned
to improve networ* 'uality.
$// provides graphical user interface for changing parameters.
Using $// also reduces the possibility of human errors by
providing validation and consistency chec*s of the parameter
settings. This means that chec*s can be run on parameter
settings before updating the networ*. $// also allows storage of
a bac*up area which can be loaded if errors occur during the
networ* update.
CELL PARAME'ERS
#arameters are necessary so that the operator can adjust and tune
the networ* to fit their specific re'uirements. 0ll parameters
have a specific permitted range and, in most cases, a default
value.
efault values are a good starting point in a new system. 2ater,
when the system is operational and measurements have been
collected, the parameters can be fine,tuned. #arameters should
be changed one at a time because, if more than one parameter is
changed, it is difficult to determine how each parameter affects
the system.
O((!et
0n offset is used to ma*e a cell appear better (worse) than it
really is by increasing (decreasing) measured signal strengths.
7y!tere!i!
0 hysteresis is used to prevent the ping,pong effect i.e., several
consecutive handovers between two cells. The ping,pong effect
can be caused by fading, the 7/ moving in a 6ig,6ag pattern
between the cells, or by non,linearities in the receiver.
Control o( Radio Net;or< Feature!
FOR INTERNAL USE ONLY "1
$ther parameters are used to control radio networ* features such
as iscontinuous Transmission (TL), fre'uency hopping, and
#ower control
'i"er! and Filter!
There are some timers and filters which can be set using
parameters. epending on the time settings or length of filters
the system responds faster or slower to changes. 0 fast system is
less stable than a slow system. 0 fast system is necessary if
micro cells are used because, in this case, handovers are
fre'uent.
Identi(ication
#arameters used to identify, e.g. a cell or a location area in the
networ*.
Penaltie!
#enalties are used in the locating algorithm to punish a cell. The
cell then appears worse then it really is. This is to avoid
handbac* in case of an urgency handover and to avoid several
repeated handover attempts in case of signaling failure.
'-re!-old!
Thresholds for cell ran*ing, call release, and access can be set.
0., S%STE& RO)T'
IN'RO#4C'ION
If the number of subscribers in a system continues to increase, at
some point it becomes necessary to increase the capacity of the
system. There are several ways to do this:
?> increase the fre'uency band (e.g. a ./7 <;; operator might
buy ./7 +:;; licenses)
?> implement half,rate
?> fre'uency re,use tighter (e.g. going from a 8=+> re,use
pattern to a -=< re,use pattern by implementing fre'uency
hopping)
?- ma*e the cells smaller and smaller
0lthough the last solution often implies introducing micro,cells
under a hierarchical cell structure, only the regular procedure for
adding new sites (cell split) is discussed here.
CELL SPLI'
It is clear that a smaller cell si6e increases the traffic capacity.
!owever, a smaller cell si6e means more sites and a higher cost
for the infrastructure. $bviously, it is preferable not to wor*
with an unnecessarily small cell si6e.
Chat is needed is in fact a method that matches cell si6es to the
capacity re'uirements. The system is started using a large cell
si6e and when the system capacity needs to be expanded, the cell
si6e is decreased in order to meet the new re'uirements. This
normally also calls for using different cell si6es in different
areas. The method is called cell split. This is illustrated in
&igure 9,8 though &igure 9,@.
example:
Initially, the largest possible cell si6e is used, considering
coverage range (&igure 9,8). The next step is to introduce three
cells per site, using the original sites and feeding the cells from
the corners (&igure 9,?). This represents a cell split of one to
three (+:-). 3ow the number of sites is still the same, but the
number of cells are three times as many as before. The following
step is to do a cell split of, e.g. one to four (&igure 9,@). 0s seen
from the figures, the old sites are still used in the new cell plan,
but additional sites are now re'uired.
F()*+, %-4 C,// 40/(: 0-14, #
F()*+, %-" C,// 40/(: 0-14, 1
F()*+, %-% C,// 40/(: 1C3 (0-14, 2
"ell split +:- (&igure 9,9) re'uires three times as many cells.
0fter the split, the capacity is three times higher per area unit,
and the cell area is three times smaller. The antenna directions
on the site that existed before the split must be changed -;
egrees.
F()*+, %-7 C,// 40/(: 1C4 (0-14, 2
"ell split +:8 (&igure 9,@) re'uires four times as many sites.
0fter the split the capacity is four times higher per area unit, and
the cell area is four times smaller. There is no need to change the
antenna directions in a +:8 cell split.

Cell Planning Overview

Uploaded by

Copyright:

Available Formats

Cell Planning Overview

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

Cell Planning Overview

Uploaded by

Copyright:

Available Formats

GSM CELL PLANNING OVERVIEW

GSM CELL PLANNING

You might also like