Module 6 Planning

1

UMTS Radio Network Planning Fundamentals

(FDD mode)Prerequisites: GSM Radio Network Engineering Fundamentals Introduction to UMTS

2

UMTS Radio Network Planning FundamentalsTable of content

1. Introduction

2. Inputs for Radio Network Planning

3. Link Budget (in Uplink) and Cell Range Calculation

4. Initial Radio Network Design

5. Basic Radio Network Parameter Definition

6. Basic Radio Network Optimization

7. UMTS/GSM co-location and Antenna Systems

AppendixAbbreviations and acronyms

3

1. Introduction

UMTS Radio Network Planning FundamentalsDuration: 2h30

4All rights reserved © Alcatel - 3FL 11194 ABAA WBZZA Ed.01P04

1. IntroductionSession presentation

Objective: to get the necessary background information in

regards of UMTS basics and RNP principles for a good start in UMTS Radio Network Planning.

Program: 1.1 UMTS Basics1.2 UMTS RNP notations1.3 UMTS RNP tool overview1.4 UMTS RNP process overview


1. Introduction

1.1 UMTS Basics

Objective:

to be able to describe the UMTS network architecture and main radio mechanisms


1.1 UMTS BasicsUMTS network architecture(1)

Iu

PLMN, PSTN,ISDN, ...

IP networks

External Networks

USIM

ME

Cu

UE

Uu(air)

User Equipmen

t

Node B

Node B

Iur

UTRAN

RNC

RNC

Node B

Node B

Iub

RNS

RNS

UMTS Radio Access

Network

MSC/VLR

CN

GMSC

GGSN

HLR

SGSN

Iu-CS

Iu-PS

Core Network

Entities and interfaces

Iub


1.1 UMTS BasicsUMTS network architecture(2)

Alcatel OMC-UR architecture

A9100 MBS

UTRAN

A9140RNC

Iub

RNS

RNS

LAN

A1353 OMC-URRNO

NM

ItfB

ItfR

A9155RNP tool

Radio Network OptimizerNetwork Performance AnalyzerNetwork Manager (used to perform supervision and configuration of the UTRAN)

RNO NPA NM

Note: NM is provided from R3 onwards. In R2, the NM function are implemented in two separate servers EM (Element Manager) and SNM (Sub-network Manager)

+NPA

A9140RNC

A9100 MBS

A9100 MBS

A9100 MBS

Note: the Alcatel NodeB is called A9100 MBS (Multi-standard Base Station) from R2 onwards


1.1 UMTS Basics3GPP: the UMTS standardization body

Members:ETSI (Europe) ARIB/TTC (Japan) CWTS (China)T1 (USA) TTA (South Korea)

UMTS system specifications: Access Network

WCDMA (UTRAN FDD) TD-CDMA (UTRAN TDD)

Core Network Evolved GSM All-IP

Note: 3GPP has also taken over the GSM recommendations (previously written by ETSI)

Releases defined for the UMTS system specifications: Release 99 (sometimes called Release 3) Release 4 Release 5

In the following material we will only deal with UMTS FDD R99.

(former Release 2000)


1.1 UMTS Basics3GPP UMTS specifications

3GPP UMTS specifications are classified in 15 series (numbered from 21 to 35), e.g. the serie 25 deals with UTRAN aspects.

Note: See 3GPP 21.101 for more details about the numbering scheme and an overview about all UMTS series and specifications.

Interesting specifications for UMTS Radio Network Planning:3GPP TS 25.101: "UE Radio transmission and Reception (FDD)"3GPP TS 25.104: "UTRA (BS) FDD; Radio transmission and Reception“3GPP TS 25.133: "Requirements for support of radio resource management (FDD)"3GPP TS 25.141: "Base Station (BS) conformance testing (FDD)3GPP TS 25.214: "Physical layer procedures (FDD)".3GPP TS 25.215: "Physical layer - Measurements (FDD)”3GPP TS 25.942: "RF system scenarios".

3GPP specifications can be found under

3GPP specifications can be found under

www.3gpp.org

www.3gpp.org


1.1 UMTS BasicsAlcatel UTRAN releases

Alcatel UTRAN equipment (RNC, NodeB and OMC-UR) is designed by a joint-venture between Alcatel and Fujitsu, called Evolium.

Note: the Alcatel UMTS equipment is called EvoliumTM 9100 MBS, EvoliumTM 9140 RNC and EvoliumTM 1353 OMC-UR

Relationship between Evolium UTRAN releases and 3GPP releases:Evolium UTRAN

releases 3GPP releases

R1 (former 3GR1)

R99 (Technical Status December

2000)R2 R99

(Technical Status June 2001)R3 R99

(Technical Status March 2002)R4 R4R5 R5

PrevisionStand: June 2004


1.1 UMTS BasicsUMTS main radio mechanisms(1)

Sector/Cell/Carrier in UMTSSector and cell are not equivalent anymore in UMTS: A sector consists of one or several cells A cell consists of one frequency (or carrier)Note: a given frequency (carrier) can be reused in each sector of each NodeB in the network (frequency reuse=1)



CDMA (called W-CDMA for UMTS FDD) as access method on the air a given carrier can be reused in each cell (frequency reuse=1)no

FDMA all active users can transmit/receive at the same timeno TDMA As a consequence, there are inside one frequency:

Extra-cell interference: cell separation is achieved by codes (CDMA)

Intra-cell interference: user separation is achieved by codes (CDMA)

Multiple frequencies (carriers) first step of UMTS deployment: a single frequency (e.g. frequency 1) is used for the whole network of an operator second step of UMTS deployment: additional frequencies can be used to enhance the capacity of the network: an additional frequency (e.g frequency 2) works as an overlap on the first frequency.

Frequency 1

Frequency 2



Channelization and scrambling codes (UL side)

2chc

1chc

scramblingc

air interfac

eModulator

3chc

UE

Phys

ical

cha

nnel

s

Channelization codes (spreading codes)short codes (limited number, but they can be reused with another scrambling code)code length chosen according to the bit rate of the physical channel (spreading factor)assigned by the RNC at connection setup

Scrambling codeslong codes (more than 1 million available)fixed length (no spreading)1 unique code per UE assigned by the RNC at connection setup

Bit rateA

Bit rateB

Bit rateC

3.84 Mchips/s

3.84 Mchips/s

3.84 Mchips/s 3.84 Mchips/s

.

.

.



Channelization and scrambling codes (DL side)

2chc

1chc

scramblingc

air interfac

eModulator

3chc

NodeBsector

Phys

ical

cha

nnel

s

Channelization codes (spreading codes)same remarks as for UL sideNote: the restricted number of channelization codes is more problematic in DL, because they must be shared between all UEs in the NodeB sector.

Scrambling codeslong codes (more than 1 million available, but restricted to 512 (primary) codes to limit the time for code research during cell selection by the UE)fixed length (no spreading)1(primary) code per NodeB sector defined by a code planning: 2 adjacent sectors shall have different codes (see §5)Note: it is also possible to define secondary scrambling codes, but it is seldom used.

Bit rateA

Bit rateB

Bit rateC

3.84 Mchips/s

3.84 Mchips/s

3.84 Mchips/s 3.84 Mchips/s

.

.

.



Physical channels Physical channels are defined mainly by:

a specific frequency (carrier) a combination channelization code / scrambling code

used to separate the physical channels (2 physical channels must NOT have the same combination channelization code / scrambling code)

start and stop instants physical channels are sent continuously on the air interface

between start and stop instants

Examples in UL: DPDCH: dedicated to a UE, used to carry traffic and signalling between

UE and RNC such as radio measurement report, handover command DPCCH: dedicated to a UE, used to carry signalling between UE and

NodeB such as fast power control commands

Examples in DL: DPCH: dedicated to a UE , same functions as UL DPDCH and UL DPCCH P-CCPCH: common channel sent permanently in each cell to provide

system- and cell-specific information, e.g. LAI (similar to the time slot 0 used for BCCH in GSM)

CPICH: see next slide



CPICH (or Pilot channel) DL common channel sent permanently in each cell to provide:

srambling code of NodeB sector: the UE can find out the DL scrambling code of the cell through symbol-by-symbol correlation over the CPICH (used during cell selection)

power reference: used to perform measurements for handover and cell selection/reselection (function performed by time slot 0 used for BCCH in GSM)

time and phase reference: used to aid channel estimation in reception at the UE side

Pre-defined symbol sequence

Slot #0 Slot #1 Slot #i Slot #14

Tslot = 2560 chips , 20 bits = 10 symbols

1 radio frame: Tf = 10 ms

The CPICH contains:a pre-defined symbol sequence (the same for each cell of all UMTS networks) scrambled with the NodeB sector scrambling codeat a fixed and low bit rate (Spreading Factor=256): to make easier Pilot detection by UE



Power control Near-Far Problem: on the uplink way an overpowered mobile

phone near the base station (e.g. UE1) can jam any other mobile phones far from the base station (e.g. UE2).

NodeB

UE1

UE2

an efficient and fast power control is necessary in UL to avoid near-far effect

power control is also used in DL to reduce interference and consequently to increase the system capacity

Power control mechanisms (see Appendix for more details): open loop (without feedback information) for common

physical channels closed loop (with feedback information) for dedicated

physical channels (1500 Hz command rate, also called fast power control)



RNC

Node B

Soft/softer Handover (HO) a UE is in soft handover state if there are two (or more) radio links between this UE and the UTRAN it is a fundamental UMTS mechanism (necessary to avoid near-far effect) only possible intra-frequency, ie between cells with the same frequencyNote: hard handover is provided if soft/er handover is not possible A softer handover is a soft handover between different sectors of the same Node B

Soft handover (different sectors of different NodeBs)

Softer handover (different sectors of the same NodeB)

RNC

Node B Node B

UE

UE



Active Set (AS) and Macro Diversity Gain All cells, which are involved in soft/softer handover for a given

UE belong to the UE Active Set (AS): usual situation: about 30% of UE with at least 2 cells in

their AS. up to 6 cells in AS for a given UE

The different propagation paths in DL and UL lead to a diversity gain, called ‘Macro Diversity’ gain: UL

one physical signal sent by one UE and received by two different cells

soft handover: selection on frame basis (each 10ms) in RNC

softer handover: Maximum Ratio Combining(MRC) in NodeB

DL two physical signals (with the same content) sent by

two different cells and received by one UE soft/softer handover: MRC in UE


1. Introduction

1.2 UMTS RNP notations and principles

Objective:

to be able to understand the vocabulary and notations* used in this course in regards of UMTS planning

* unfortunately, UMTS RNP notations are not clearly standardized, so that the meaning of a notation can be quite different from one reference to another one.


Received power and power density

Power [dBm]

Power Density

[dBm/Hz]

Comment (Power Density=Power/B

with B=3.84MHz)

Received (useful) signal

C (or

RSCP)Ec

Ec = Energy per chip=C/B

Thermal Noise -108.1 Nth=-174Nth = k.T0 with k=1.38E-20mW/Hz/K (Bolztmann constant) and T0=293K (20°C)

Thermal Noise at receiver N -

N =-108.1dBm+NFreceiver [dB] (=Thermal noise + Noise generated at receiver)

Interference intra-cellIintra

(Iown)-

interference received from transmitters located in the same cell as the receiverNote: C is included in Iintra

Interference extra-cellIextra

(Iother;Iinte

r)-

interference received from transmitters not located in the same cell as the receiver

Interference I -I=Iintra+ Iextra

(no “Thermal noise at receiver” included)

1.2 UMTS RNP notations and principlesNotations (1)


Received power and power density

Power [dBm]

Power Density

[dBm/Hz]

Comment Power Density=Power/B with

B=3.84MHzTotal received power (“Total noise”)

I+N(RSSI) Io

I+N= Iintra+ Iextra +NNote: C is included in (I+N)

Total received power (“Total noise” without useful signal)

I+N-C No(Nt)

No=( Iintra+ Iextra +N-C)/BNote: C is not included in No


Note: Io can be measured with a good precision, whereas No is not easy to measure (but it is useful for theoretical demonstrations)


Ratio in [dB] Comment

Received energy per chip over “noise”

Ec/IoHere “noise”=IoThis ratio can be accurately measured: it is used for physical channels without real information bits, especially for CPICH (Pilot channel)

Ec/No(“C/I”)*

Here “noise”=NoThis ratio is difficult to measure, but is useful for theoretical demonstrations: it is used for physical channels with real information bits, especially for P-CCPCH and UL/DL dedicated channels.

Received energy per bit over “noise”

Eb/NoEb/No=Ec/No+PG with PG (Processing Gain) = 10 log [(3.84 Mchips/s) / (service bit rate)]e.g. for speech 12.2 kbits/s, Processing Gain = 25dB

Required energy per bit over “noise”

(Eb/No)req

Fixed value which depends on service bit rate...(see §3.5)Eb/No shall be equal or greater than the (Eb/No)req


*This ratio is often written with the classical GSM notation “C/I” (Carrier over Interference ratio): this notation is incorrect, it should be C/(I+N-C)


Two more interesting

ratios!in [dB] Comment

f (or little i)

Iextra / Iintra

In a homogenous network (same traffic and user distribution in each cell), f is a constant in uplink. Typical value for macro-cells with omni-directional antennas: 0.55 (in uplink)

Noise Rise (I+N)/NVery useful UMTS ratio to characterize the moving interference level I compare to the fixed “Thermal Noise at receiver” level N.



1.2 UMTS RNP notations and principlesExercise (1/2)

Assumptions:- n active users in the serving cell with speech service at 12.2kbits/s

and (Eb/No)req =6 dB- Received power at NodeB: C=-120dBm (for each user)- homogenous network (f=0.55)- NFNodeB = 4dB and NFUE =8dB

NodeB

Serving cellSurrounding cells

Uplink considered


1.2 UMTS RNP notations and principlesExercise (2/2)

1. What is the processing gain for speech 12.2kbits/s ?

2. The users in the serving cell are located at different distance from the NodeB: is it desirable and possible to have the same received power C for each user?

3. What is the value of the “Thermal Noise at receiver” N?

4. Complete the following table:n [users

]

I [dBm]

I +N[dBm]

Noise Rise [dB]

Ec/No [dB]

Eb/No [dB] Comment

1

10

25

100


1. Introduction

1.3 UMTS RNP Tool Overview

Objective:

to be able to describe briefly the structure of a RNP tool


1.3 UMTS RNP Tool OverviewRNP tool requirements(1)

Digital maps topographic data (terrain height)

Resolution: typically 20m for city areas and 50 m for rural areas possibly building and road databases for more accuracy

Coordinates system important for interfacing with measurement tools e.g. UTM based on WGS-84 ellipsoid

morphographic data (clutter type) Resolution: same as topographic data

Propagation model dialog e.g. setting Cost-Hata propagation model parameters (see §3.2)

Site/sector/cell/antenna dialog importing sites (e.g GSM sites) setting site/sector/cell/antenna parameters (“Network design

parameters”, see §4.1)Note: in UMTS, sector and cell are not equivalent


1.3 UMTS RNP Tool OverviewRNP tool requirements(2)

Link loss calculation Traffic simulation

Setting traffic parameters (§2.2) Traffic map generation

Resolution: same as topographic data UE list generation (a snapshot of the UMTS network)

Coverage predictions displaying the results on the map showing the results as numerical tables

Automatic neighborhood planning Automatic scrambling code planning Interworking with other tools (dimensioning tools, OMC-UR,

measurements tools, transmission planning tool...)


1. Introduction

1.4 RNP Process Overview

Objective:

to be able to describe briefly the 11 steps of the RNP Process, which starts with Radio Network Requirements definition and ends with Radio Network Acceptance.


(12. Further Optimization)

1.4 RNP Process OverviewThe 11 steps of RNP process

1. Radio Network Requirements (see §2.4)

2. Preliminary Network Design

(see §3)3. Project Setup and

Management


(see §4)5. Site Acquisition

Procedure6. Technical Site

Survey

7. Basic Parameter Definition(see §5)

8. Cell Design CAE Data Exchange over COF

9. Turn On Cycle10. Basic Network

Optimization(see §6)

11. Network Acceptance


1.4 RNP Process Overview Step 1: Definition of Radio Network Requirements

The Request for Quotation (RfQ) from the operator prescribes the requirements which consists mainly in: Coverage TrafficQoS

see §2.4 for more details


1.4 RNP Process Overview Step 2: Preliminary Network Design

The preliminary design lays the foundation to create the Bill of Quantity (BoQ) List of needed network

elements Geo data procurement

Digital Elevation Model DEM/Topographic map

Clutter map Definition of standard equipment

configurations dependent on clutter type traffic density

Definition of roll out phases Areas to be covered Number of sites to be

installed Date, when the roll out

takes place. Network architecture design

Planning of RNC, MSC and SGSN locations and their links

Frequency spectrum from license conditions


1.4 RNP Process Overview Step 3: Project Setup and Management

This phase includes all tasks to be performed before the on site part of the RNP process takes place.

This ramp up phase includes: Geo data procurement if required Setting up ‘general rules’ of the project Define and agree on reporting scheme to be used

Coordination of information exchange between the different teams which are involved in the project

Each department/team has to prepare its part of the project Definition of required manpower and budget Selection of project database (MatrixX)


1.4 RNP Process Overview Step 4: Initial Radio Network Design

Area surveys As well check of correctness of geo data

Frequency spectrum partitioning design RNP tool calibration

For the different morpho classes:Performing of drive measurementsCalibration of correction factor and standard deviation by

comparison of measurements to predicted received power values of the tool

Definition of search areas (SAM – Search Area Map) A team searches for site locations in the defined areas The search team should be able to speak the national language

Selection of number of sectors/cells per site together with project management and operator

Get ‘real’ design acceptance from operator based on coverage prediction and predefined design level thresholds


1.4 RNP Process Overview Step 5: Site Acquisition Procedure

Delivery of site candidates Several site candidates shall be

the result out of the site location search

Find alternative sites If no site candidate or no

satisfactory candidate can be found in the search area

Definition of new SAM (Search Area Map)

Possibly adaptation of radio network design

Check and correct SAR (Site Acquisition Report) Location information Land usage Object (roof top, pylon, grassland)

information Site plan

Site candidate acceptance and ranking If the reported site is accepted as

candidate, then it is ranked according to its quality in terms ofRadio transmission

High visibility on covered areaNo obstacles in the near field of the antennasNo interference from other systems/antennas

Installation costsInstallation possibilitiesPower supplyWind and heat

Maintenance costsAccessibilityRental rates for objectDurability of object


1.4 RNP Process Overview Step 6: Technical Site Survey

Agree on an equipment installation solution satisfying the needs of RNE (Radio Network Engineer) Transmission planner Site engineer Site owner

The Technical Site Survey Report (TSSR) defines Antenna type, position,

orientation and tilt Mast/pole or wall mounting

position of antennas EMC rules are taken into

accountRadio network engineer

and transmission planner check electro magnetic compatibility (EMC) with other installed devices

BTS/Node B location Power and feeder cable mount Transmission equipment

installation Final Line Of Site (LOS)

confirmation for microwave link planningE.g. red balloon of around

half a meter diameter marks target location

If the site is not acceptable or the owner disagrees with all suggested solutions The site will be rejected Site acquisition team has to

organize a new date with the next site from the ranking list


1.4 RNP Process Overview Step 7: Basic Parameter Definition

After installation of equipment the basic parameter settings are used for Commissioning

Functional test of BTS/NodeB and VSWR check

Call tests RNEs define cell design data Operations field service generates

the basic software using the cell design CAE data

Cell parameters definition LAC/RAC... Frequencies Neighborhood/cell

handover relationship Transmit power Cell type (macro, micro,

umbrella, …) Scrambling code planning

39




2. Inputs for Radio Network Planning Session presentation

Objective: to be able to describe the UMTS RNP inputs in

regards of frequency spectrum, traffic parameters, equipment parameters and radio network requirements

Program: 2.1 UMTS FDD frequency spectrum2.2 UMTS traffic parameters2.3 UMTS Terminal, NodeB and

Antenna overview2.4 UMTS Radio Network Requirements



2.1 UMTS FDD frequency spectrum

Objective:

to be able to describe the UMTS FDD frequency parameters defined by the 3GPP


2.1 UMTS FDD frequency spectrum Frequency spectrum

1920-1980 2110-2170

Frequency spectrum (UMTS FDD mode) UL: 1920 MHz – 1980 MHz DL: 2110 MHz – 2170 MHz Duplex spacing: 190 MHz


2.1 UMTS FDD frequency spectrum Carrier spacing

Carrier spacing: 5MHz 2110 MHz – 2170 MHz = 60 MHz; 60 MHz / 5 MHz =12

frequencies One operator gets typically 2–3 frequencies (carriers) So typically 4–6 licenses per country as a maximum

Required bandwidth: 4.7MHz The chip rate is 3.84Mchip/s, therefore at least 3.84MHz bandwidth are needed

to avoid inter-symbol interference (Nyquist-Criterion) The roll-of factor of the pulse-shaping filter is 0.22 (root-raised cosine) The needed minimum bandwidth is 3.84MHz x 1.22 4.7MHz

Examples: 60MHz

5MHz

6 operators

4 operators


2.1 UMTS FDD frequency spectrum Frequency channel numbering

UTRA Absolute Radio Frequency Channel Number (UARFCN) UARFCN formula (3GPP 25.101 and 25.104):

MHz.fMHzwith

[MHz]fUARFCN

nlinkUplink/DowCenter

nlinkUplink/DowCenternlinkUplink/Dow

632760.0

5

UARFCN is integer: 0 <= UARFCN <= 16383


2.1 UMTS FDD frequency spectrum Center Frequency

Center Frequency fcenter

Consequence of UARFCN formula (see previous slide): fcenter must be set in steps of 0.2MHz (Channel Raster=200

kHz) fcenter must terminate with an even number (e.g 1927.4 not

1927.5)

fcenter values Uplink (1920Mhz-1980MHz)

1922.4MHz <= fcenter <= 1977.6MHz 9612 <= UARFCN Uplink <= 9888

Downlink (2110Mhz-2170MHz) 2112.4MHz <= fcenter <= 2167.6MHz 10562 <= UARFCN Downlink <= 10838


2.1 UMTS FDD frequency spectrum Further comments

Frequency adjustment If an overlap between frequency bands belonging to same

operator is set, guard band between different operators will increase.

This feature can be used to enlarge the guard band between frequency blocks belonging different operators and prevent dead zones.

Example:it shows an overlap of 0.3 MHz between two carriers of one operator0.6 MHz additional channel separation between the operators is created.

0.6 MHz additionalguard band

5 MHz5 MHz

4.7 MHz 4.7 MHz0.3 MHz overlap

1920 1940Operator 1 Operator

2

Frequency coordination at country borders (see Appendix)

0.3 MHz overlap



2.2 UMTS traffic parameters (UMTS traffic map)

Objective:

to be able to describe the method to create a traffic map


2.2 UMTS traffic parameters Step 1: Terminal parameters

Tx power (dBm)

Terminal parameters

(typical values) Min Max

AntennaGain (dB)

Internal Losses+ Indoor Margin (dB)

Noise Factor (dB)

Active set size

Deep Indoor 20 Indoor 18

Indoor First Wall 15 Incar 8

Mobile phone

Outdoor

21

0 Deep Indoor 20

Indoor 18 Indoor First Wall 15

Incar 8

Personal Digital Assitent (PDA)

Outdoor

-50

24

0

0

8

3

The indoor margin (also called penetration loss) is part of UE parameters.


2.2 UMTS traffic parameters Step 2: Service parameters(1)

(Eb/No)req (dB)

DL traffic Power (dBm)

3 Km/h 50 km/h 120 km/h Service

parameters (typical values) UL DL UL DL UL DL Ty

pe

SHO

allo

wed

Prio

rity

UL n

omin

al ra

te

(Kb/

sec)

DL n

omin

al ra

te

(Kb/

sec)

Codi

ng F

acto

r UL

/DL

Activ

ity F

acto

r (U

L/DL

)

Min Max

Body

loss

(d

B)

Speech 12.2 3 12.2 12.2 0.6 3

CS 64 CS

2 64 64 PS 64 1 64 64 PS 128 0 64 128 PS 384

see next page PS

Y

0 64 384

1 1

-50 +40 0

Activity factor and Body loss are part of service parameters


2.2 UMTS traffic parameters Step 2: Service parameters(2)

(Eb/No)req typical values• fixed values which depends on link

direction (UL or DL )service bit rate, BLER (or BER), UE speed, UE multipath environment, TX/RX diversity and processing/hardware imperfection margin (2dB)

Uplink Downlink2 rx ants 1 tx ant

Vehicular A - 3 km/h 5,8 7,6Vehicular A - 50 km/h 6,2 8,1Vehicular A - 120 km/h 7,1 8,7

SPEECH 12.2



CIRCUIT 64



PACKET 64



PACKET 128



PACKET 384

PS services for a target BLER of 0.05

CS services for a target BLER of 0.0001 (10-4)

Speech services for a target BLER of 0.01(10-2)

Source: Alcatel simulations


2.2 UMTS traffic parameters Step 3: User Profile parameters

Traffic Density

Volume (Kb/sec)

User Profile

(Examples) Service

(see Step2) Terminal

(see Step1) Calls/ hour

Duration (sec)

UL DL Surfing user PS 384 PDA Deep Indoor 1 - 8 60 Videocall user PS 64 PDA Deep Indoor 1 - 5 20

Phonecall user Speech 12.2 Mobile phone Deep Indoor 1 115.2 - -

Speech 12.2 1 72 - - CS64 1 72 - - PS64 PS128

City user

PS384

Mobile Phone Outdoor 0.2 - 40 200

Standard user same as City User without PS384 service All of this data has to be provided by the operator: as the user profiles

will be different for different operators in different countries, no typical values can be given.


2.2 UMTS traffic parameters Step 4: Environment Class parameters

User profiles have been used to describe single user types. Environment classes are used to distribute and quantify these user

profiles on the planning area.

Environment class*

(Examples)

User profiles (see Step

3)

Geographical density (users/km2)

low traffic

medium traffic high traffic

Dense Urban city user 1000 3000 6000Urban city user 750 1500 3000Suburban city user 50 250 500Rural standard user 10 20 40

*BE CAREFUL: environment classes and clutter classes have often the same names, although they refer to quite different concepts: an environment class refers to a traffic property whereas a clutter class refers to an electromagnetic wave propagation property. The reason is that environment classes are very often mapped on clutter classes to generate a traffic map (see Step 5)


2.2 UMTS traffic parameters Step 5: Traffic Map definition

Mapping of Environment Classes (see Step 4) on a map: Example with 4 environment classes: Dense Urban, Urban, Suburban, Rural

Dense Urban

Urban

Rural

Suburban

Resolution:20m…100m

Planning Area(also called Focus Area)

Map Traffic map

Note: an easy way to generate a traffic map is to use the clutter map and to associate each clutter class to an environment class (e.g. Dense Urban environment class is mapped on Dense Urban clutter class…)



2.3 UMTS Terminal, Antenna overview

Objective:

to be able to describe briefly the main characteristics of the UMTS radio equipment (UE, Alcatel NodeB and antenna)


2.3 UMTS Terminal, NodeB and Antenna overviewUE characteristics

According to 3GPP 25.101 (Release 1999): UE power classes at antenna connector*:

Power class 1: (+33 +1/-3)dBm Power class 2: (+27 +1/-3)dBm Power class 3: (+24 +1/-3)dBm Power class 4: (+21 ±2)dBm

UE minimum output power: <-50dBm

According to UE manufacturers: UE Noise Figure: 8dB (typically) UE internal losses + UE antenna gain = 0dB

What is EIRP for a UE of power class 4?* the notation means e.g. for class 1:- Maximum output power: +33dBm- Tolerance: +1dBm/-3dBm

Answer: UE EIRP=UE TX Power+ UE Antenna Gain - UE Internal Loss=21dBm + 0 dB = 21 dBm


2.3 UMTS Terminal, NodeB and Antenna overview UMTS antennas (1)

Constraints for antenna system installation: visual impact space or building constraints co-siting with existing GSM BTS (see §7)

Note: the antenna system includes not only the antennas themselves, but also the feeders, jumpers and connectors as well as diplexers (in case of antenna system sharing) and TMAs (tower mounted amplifiers)

Whenever possible, a solution with a standard antenna has to be chosen: Model: 65° horizontal beam width Azimuth: 0°, 120° and 240° (3 sectored site) Gain: 17-18dBi Height (above ground): 20-25 m for urban and 30-35 m for

suburban Downtilt: electrical downtilt adjustable between 0° and 10°


2.3 UMTS Terminal, NodeB and Antenna overview UMTS antennas (2)

Antenna parameters are key parameters which can be tuned to decrease interference in critical zones, especially: Antenna downtilt

by increasing the antenna downtilt of the interfering cell downtilt changes with a difference less than 2° compared

to the previous value do not make sense, since the modification effort (requiring on-site tuning) does not stand in relation to the effect.

rule of thumb: the downtilt in UMTS should be at least 1°-2° higher than the value a planner would chose for GSM

Antenna azimuth by re-directing the beam direction of the interfering cell azimuth modifications of 10°-20° compared to the

previous value do not make senseNote: Azimuth/downtilt modifications can be restricted or even forbidden due to antenna system installation constraints (especially the constraints for UMTS/GSM co-location, see §7 for more details)



2.4 Radio Network Requirements

Objective:

to be able to understand the parameters, which define the UMTS radio network requirements in terms of coverage, traffic and quality of service


2.4 Radio Network Requirements Definition of radio network requirements (1)

Traffic mix and distribution for traffic simulation with the aim to predict power load in DL and UL noise rise (see §2.2)

Covered area Polygon surrounding the area to be covered (focus zone for

RNP tool)

Definition of what coverage is CPICH Ec/Io coverage

(CPICH Ec/Io)required=-15dB (Alcatel value coming from simulations and field measurements)

Required coverage probability for CPICH Ec/Io: e.g. Average probability {CPICH Ec/Io > (CPICH Ec/Io)req} > 95%(with this definition a minimum average quality in the covered area is guaranteed*)

*other definitions of required coverage probability are possible, e.g. 95% of area with CPICH Ec/Io > (CPICH Ec/Io)required

(with this definition, a minimum percentage of covered area is guaranteed)



UL and DL service coverage (Eb/No)reqspecific value for each service and for each direction (UL/DL), see §2.2 Required coverage probability for DL and UL services:

e.g. Average probability {Eb/No > (Eb/No)req} > 95% (for each direction UL/DL and for each service)Note: It is possible to define different required coverage probabilities for different services.

Eb/No values can not easily be measured, but nevertheless service coverage predictions are a good source of information to improve the radio network design (to find the limiting resources).



CPICH RSCP coverage (optional) (CPICH RSCP)required: it can be defined, if the maximum

allowed path loss is determined by calculating a link budget and taking into account the CPICH output power (if no traffic mix is available, the link budget would base on the limiting service)

Required coverage probability for CPICH RSCPe.g. Average probability {CPICH RSCP > (CPICH RSCP)req}>95%(To guarantee an average reliability, that the minimum level is fulfilled in the covered area)

CPICH RSCP prediction is not mandatory, but: it can be a help to guarantee a certain level of indoor

coverage from outdoor cells, taking into account different indoor losses for different areas.

CPICH RSCP can easily be measured using a 3G scanner.

62




3. Link Budget (in Uplink) and Cell Range Calculation Session presentation

Objective: to be able to calculate the cell range for a given

service by doing a manual link budget in UL.

to be able to describe the typical UMTS radio effects in UL and in DL.

Program: 3.1 Inputs for a manual UL link budget3.2 UMTS propagation model 3.3 UMTS shadowing and fast fading modeling3.4 Calculation of Node B reference sensitivity3.5 UMTS interference modeling3.6 Calculation of cell range



3.1 Inputs for a manual UL link budget

Objective:

to be able to define the necessary inputs for an UL link budget (in order to prepare cell range calculation).


3.1 Inputs for a manual UL link budgetPrinciple for Cell Range calculation

We consider a link budget in UL (assuming that the coverage is UL limited). It is known that:

the pathloss Lpath depends on the distance UE-NodeB d (see §3.2). Lpath = MAPL for d=Cell Range.

We calculate MAPLk for the limiting service k in UL:

NodeB

UE

dBGainsdBLossesdBMargins

dBmysensitivitReference_dBmEIRPdBMAPL kNodeB,UEk

EIRPUE

(see §2.3)

Reference_sensitivityNodeB,k

(see §3.4)Margins

Losses

Gains

d=Cell Range


3.1 Inputs for a manual UL link budget Inputs for the UL link budget

MarginsShadowing margin* see §3.3Fast fading margin see §3.3Interference margin see §3.5Losses

Feeders and connectorsNodeB typically 3dB (it depends on the feeder length..)

Body loss see §2.2Penetration loss (indoor margin) see §2.2

Gains*Antenna gainNodeB typically 18dBi

*Soft/softer handover gain is included in the shadowing margin (see §3.3)



3.2 UMTS propagation model

Objective:

to be able to describe the parameters involved in UL/DL wave propagation.

to find out the relationship between the pathloss and the distance UE-NodeB


3.2 UMTS propagation modelHow to calculate the Pathloss Lpath?

For UMTS link budget calculations, we have to find out the value of the Pathloss Lpath between the NodeB and the UE using: The free-space formula:

It cannot be used in mobile networks such as UMTS, because the Fresnel ellipsoid is obstructed in the environment of the UE over a big distance (due to low height above the ground of the UE).

Empirical formulas: The most effective approach is based on the classical COST 231-Hata formula, extended for the usage on higher frequencies or additional propagation effects.e.g. Alcatel selected as UMTS propagation model a slightly modified COST 231-Hata model, called the Standard Propagation Model*.

In UMTS radio environment, the propagation waves are subject to complex mechanisms:

Free Space Propagation Reflections/Refractions/Scattering Diffraction

Slow fading (Shadowing) Fast Fading (Multipath fading)

*see Appendix for the relationship between COST231- Hata and the Alcatel Standard Propagation Model


3.2 UMTS propagation model Alcatel Standard Propagation Model

Lpath formula:

Important: this formula takes into account free space propagation, reflections /refractions/scattering and

diffraction not slow and fast fading effects (never considered in

propagation model, but as margins see §3.3)

(m) UEof height antenna effective :H(m) NodeBof height antenna effective:H

(m) UE-NodeB distance:d*with

eff

eff

UE

NodeB

path

clutterfKHfKHdK

ndiffractiofKHKdKKL

clutterUENodeB

NodeB

effeff

eff

)(loglog)(loglog

65

4321

*see next slides for the values of the 7 multiplying factors K1, ..., K6, Kclutter and the calculations of the 3 functions f(diffraction), f(HUEeff), f(clutter)



Can we consider for the antenna height in the Lpath formula the height above the sea? the height above the ground?

What is the effective antenna height of NodeB and UE? Typical values for the antenna height of NodeB and UE above

the ground level are:HNodeB above ground = 20-25 m for urban and 30-35 m for suburbanHUE above ground = 1.5 m

These values and the topographic information between NodeB and UE are used to calculate an effective antenna height HNodeB

eff and HUE eff , in order to model the real effect of antenna height on the pathloss.

The effective height and the height above the ground : are equal on a flat terrain (of course) can be very different on a hilly terrain Answer:

Height above the sea: no (Mexico isn’t better than Shanghai due to its higher altitude!)Height above ground: it is can be a strong approximation on a hilly terrain. Indeed assume a 20 m antenna is located on the top of a 500 m hill. The height above ground is 20 m, but the antenna height shoud be 520 m.



Multiplying factors (directly derived from COST-Hata model)

Name Value Factor related to

Comment

K1 23.6(for f=

2140MHz)

constant offset

used to take into account free space propagation and reflections/refractions/scattering mechanisms for a standard clutter class.

K2 44.9 d same comment as K1.

K3 5.83 HNodeB eff same comment as K1.

K5 -6.55 d , HNodeB eff same comment as K1.

K6 0 HUEeff same comment as K1. As the contribution of f(HUEeff) is close to zero, K6 is set to zero.

Propagation model parameters (1)



Multiplying factors (not included in COST-Hata model)

Name Value Factor related to

Comment

K4 1 f(diffraction)

used to take into account diffraction mechanisms see further comments on f(diffraction).

Kclutter 1 f (clutter) used to take into account the necessary correction compared to the standard clutter class see further comments on f(clutter).

Propagation model parameters (2)



Clutter Class* Clutter Loss1 buildings -1.0

2 dense urban -3.03 mean urban -6.04 suburban -8.05 residential -11.06 village -14.07 rural -20.08 industrial -14.09 open in urban -12.010 forest -9.011 parks -15.012 open area -24.013 water -27.0

Propagation model parameters (3) clutter losses based on experienced values

*BE CAREFUL: do not confuse clutter classes and environment classes (see §2.2)



Calculation of the diffraction loss f(diffraction)Approximation: an obstacle of height H between NodeB and UE is modeled as an infinite conductive plane of height H. Case 1: one obstacle

NodeB

UE

What is the diffraction loss in case 1 (use the curve on the next page)?

LOS r

h0

Fresnel Ellipsoid (first order)

Infinite conductive plane

H

Answer: h0=r v=-1 f(diffraction)=14dB



Knife-edge diffraction function

-5

0

5

10

15

20

25

30

35

-9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3

Clearance of Fresnel ellipsoid (v)

F(v)

[dB

]

Calculation of the diffraction loss f(diffraction) Case 1: one obstacle (continuing)

Diffraction loss for one obstacle:

v: clearance parameter, v=-h0/rr: Fresnel ellipsoid radius, h0: height of obstacle above line of sight (LOS)

Note:h0 = 0 v =0 F(v) = 6 dB



Calculation of the diffraction loss f(diffraction) Case 2: several obstacles

NodeB

UE

LOS

The diffraction loss in case 2 is not easy to calculate: it is not equal to the sum of the contributions of each obstacle alone (it is usually smaller).

Different calculations methods can be applied based on the General method for one or more obstacles described in ITU 526-5 recommendations, e.g Deygout, Epstein-Peterson or Millington



Calculation of f(clutter): In the Lpath formula, the multiplying factors K1,..,K6 are

calculated for a standard clutter class: f(clutter) is a correction factor compared to the standard clutter class.

f(clutter) is calculated taking into account a clutter loss* average of all pixels located in the line of sight and in a circle around the UE (the circle radius, called Max distance, is typically 200m).

Pixel

NodeBMax distance

UE

Water clutter class pixel clutter loss = -27 dB (typically)

Forest clutter class pixel clutter loss = -9 dB (typically)

*(also called clutter or morpho correction factor)

in this example, 3 pixels are considered to calculate f(clutter)



Calculation of f(clutter): How are provided the clutter loss values?

based on experienced values: simple, accuracy of +/-3 dB (see previously)

based on calibration measurements: complex and expensive way, but accuracy of +/-1 dB.

Is it possible to reuse GSM1800 calibration measurements(in order to save costs of expensive measurement campaigns)?The difference between 1850 MHz (middle of GSM1800 band) and 2140 MHz (middle of DL UMTS FDD band) involves: fixed offset of 0.9dB for all clutters taken into account in

K1:K1=24.5 (COST-Hata value for f=2140MHz) – 0.9dB

= 23.6 no significant correction offset per clutter except if large

vegetation is penetratedConclusion: GSM 1800 calibrations can be reused. Only for clutter type mainly covered by vegetation, additional calibration is recommended.



Calculation of f(clutter) (simplified*): all the values are negative and are given compared to the

“standard clutter class” for which f(clutter) =0 dB (the worst case)

Example:

Clutter Class f(clutter) (simplified*)

Dense urban -3

Urban -6

Sub-urban -8

Rural -20

*Assumption:homogeneous clutter class around the UE


3.2 UMTS propagation model Other Propagation Models

Other propagation models can be applied, especially for micro-cell planning: e.g. Walfish-Ikegami or Ray-Tracing necessary to have building and road databases (expensive)


3.2 UMTS propagation model Alcatel Standard Propagation Model (simplified formula)

Clutter class

dUE-

NodeB [km]

C1 [dB]

C2 x log(dUE-

NodeB)[dB]

Lpath [dB]

Dense Urban

0.512

Suburban

0.512

*Assumptions:-HNodeBeff=30m-no diffraction-homogeneous clutter class around the UE

Exercise: Let’s consider the simplified* formula of the Alcatel Standard

Propagation Model:Lpath[dB] = C1 + C2 x log(dUE-NodeB[km])

Can you complete the table?



3.3 UMTS shadowing and fast fading modeling

Objective:

to be able to find out the UL margins due to fading effects (fast fading and shadowing)

to be able to describe the fading effects in UL and in DL


3.3 UMTS shadowing and fast fading model Definition of fading(1)

Let’s consider a the received power level C of a UE at the cell edge, taking into account the pathloss, all gains, all losses and all margins, except shadowing and fast fading margins.

NodeBUE

EIRPUE

Reference_SensitivityNode

B,k= Cthreshold

(fixed value for a given service k)

– Lpath – Losses + Gains

– Margins (except fading)

UE received power C

Time

Cmean

=Cthreshold

(fixed value)

UE received power C oscillates around a mean value Cmean equal to Cthreshold

Cell Range


3.3 UMTS shadowing and fast fading model Definition of fading(2)

Shadowing (or Slow Fading or long-term fading )

Fast Fading (or Multipath fading or small-scale fading or Rayleigh fading)Cmean

Cthreshold

(fixed value)

Time

UE received power C

Shadowing and fast fading margins are necessary to maintain the UE received power C above the fixed Cthreshold during the most part of

the time


3.3 UMTS shadowing and fast fading model Shadowing (1)

Cause: Shadowing holes appear in the received power C when the UE is in the “shadow” of large objects (size>10m)

Modeling:The received power C can be modeled as a Log-normal distribution with: a mean value Cmean

a standard deviation , typically =7-8 dB (clutter dependent)

Note: GSM1800 calibrations can be reused for the values.

Signal distribution

Prob

abili

ty

std dev=8 dB

std dev = 4dB

std dev= 2dB

std dev= 6dB

Cmean



Definition of reliability level and reliability margin: Reliability level* =% of time for the received power C to be

above Cthreshold (for a sufficient observation time period) at a given pixel

Reliability marginx% =Cmean offset compared to the fixed Cthreshold to get a reliability level of x%

Wanted reliability level=50% Reliability margin50%=0dB Cmean = Cthreshold

UE received power C

Time

Cmean

=Cthreshol

d

(fixed value)

UE received power C

Time

Cthreshold

(fixed value)

Cmean

reliability margin

50%

95%

Wanted reliability level=95% Reliability margin95%=10dB (for =6)Cmean = Cthreshold +10dB (see next slide for calculation of Reliability marginx%)

*also called local coverage probability or coverage probability per pixel



Reliability level (also called local coverage probability or coverage probability per pixel)

0%

20%

40%

60%

80%

100%

-20 -10 0 10 20F = (Fmed - Fthr) /dB

Reliability margin95.2%=10dB

95,2%

50% probabilityfor Fmed=Fthr

Curve for a standard deviation =6dB

k - -0.5 0 1 1.3 1.65 2 2.33 +

Reliability level

0% 30% 50% 84% 90% 95% 97.7%

99% 100%

Reliability margin*=k

* be careful! the reliability margin (defined above) corresponds to the GSM shadowing margin, but not to the UMTS shadowing margin (see further)

Calculation of reliability margin*: It depends on the reliability level and on the standard deviation



Values for the standard deviation : Power level [dBm] (e.g CPICH RSCP):

it can be modeled as a log-normal variable with a standard variation (clutter dependent value, typically 7dB or 8dB)

Ratio [dB] (e.g CPICH Ec/Io or UL/DL Eb/No) it can normally NOT be modeled as a log-normal variable, because

the numerator and the denominator are modeled as separate log-normal variables with separate standard deviations.

Approximation: a ratio is modeled as a log-normal variable with a standard deviation which is estimated according to the correlation between the numerator and the denominator: CPICH Ec/Io : strong correlation between shadowing effect on Ec

and shadowing effect on Io. CPICH Ec/Io is constant (Field value:3dB)

DL Eb/No: same as CPICH Ec/No UL Eb/No: no specific correlation between Eb and No. UL Eb/No is

a clutter dependent value as for CPICH RSCP



Reliability level=87%



Cell coverage probability=95%

Definition of area (cell) coverage probability: If the reliability levels are provided at each pixel of a area (or a

cell), it is easy to calculate the Area(or cell) coverage probability as the average of all reliability levels.

Area (cell) coverage probability=% of time for the received power C to be above Cthreshold (for a sufficient observation time period) in average over the area(cell).

Average



Definition of shadowing margin: If the area (cell) coverage probability is provided (from the

radio network requirement, see §2.4), it is possible to find out the reliability levels in the area (cell).

Reliability level=?Reliability Margincell edge=?

Reliability level=?

Reliability level=?Cell coverage probability=95%

For a UE at cell edge: Shadowing margin* = Reliability Margincell edge – Soft/Softer HO Gain

*the UMTS shadowing margin (defined above) is NOT the same as the GSM shadowing margin(=Reliability Margin)



How to calculate the shadowing margin for a received power C? It depends on:

Wanted cell coverage probability Clutter class of the UE UE soft/softer handover state and correlation factor

between UE radio links (0=no correlation, typically 0.5) Examples in uplink (Source: Alcatel simulations)

Note:in case of soft/er handover (it is typically the case for a UE at cell edge), the soft/er handover gain partially compensates for the additional path loss caused by shadowing.

S h ad ow in g m arg in (d B ) (n o S H O )

U L S h ad ow in g m arg in (d B ) (S H O , 2 leg s )

C ell co verag e

pro b ab ility = 6 = 8 = 12 = 6 = 8 = 12 95 % 5 .9 8 .7 14 .6 3 .1 4.8 8 .5 90 % 3 .3 5 .4 10 .0 0 .6 2.1 6 .4

Soft Handover Zone

“Shadowing Hole”


3.3 UMTS shadowing and fast fading model Fast Fading (1)

Cause: summation and cancellation of different signal components of the same signal which travel on multiple paths

Modeling Rayleigh distributed fading with correlation distance /2

Note: =15 cm for f=2GHz positive fades are less strong than negative fades (unequal

power variance)

RayleighSmall-ScaleFading

RayleighPDF


3.3 UMTS shadowing and fast fading model UL Fast Fading (2)

How to compensate for fast fading losses in UPLINK? Case 1: slow moving UE (0-50km/h)

Power control (inner loop at 1500Hz) compensates fairly well with a TX power increase for the fast fading losses in the serving cell, but: It works only if the UE has enough TX power Power

Control Headroom (called Fast Fading Margin) necessary, especially for the UEs at the cell edge (see further)

Side effect: increase of f value (little i value) for the surrounding cells (see further)

Case 2: fast moving UE (>50km/h) Power Control loop is too slow to compensate for fast fading A margin is necessary to compensate for the fast fading

losses: this margin is not explicit, but implicitly included in the (Eb/No)req values (see §2.2)



How to calculate Power Control Headroom (Fast Fading Margin) for slow moving UEs (Case 1)? Fast fading depends on:

required BER (or BLER) UE speed Multipath environment (Vehicular A, Pedestrian A…) UE soft/softer handover state and power difference

between UE radio links Example for uplink (Source: Alcatel simulations)

Fast fading margin (dB) for several target BLER Multipath

environment 10-1 10-2 10-3 10-4

Dense urban, urban, suburban (Veh. 3km/h) 0.6 1.7 2.5 3.3

Rural (Veh. 50 km/h) -0.3 -0.3 -0.3 -0.2

Assumption:Soft handover considered with 2 links and 3dB power difference between the 2 links



- 5

- 1 0

- 1 5

0

5

1 0

1 5

0 0 . 2 0 . 4 0 . 6 0 . 8 1 1 . 2 1 . 4 1 . 6 1 . 8 2S e c o n d s , 3 k m / h

dB

C h a n n e l

T r a n s m i t t e dp o w e r

N o d e - B r e c e i v e d

p o w e r

A v e r a g e t r a n s m i t

p o w e rP o w e r

r i s e

What about the side-effect for slow moving UE (Case 1)?Fast fading in serving cell and in neighboring cells are not correlated: impact on neighboring cells due to UE TX power increase which

causes additional UL extra-cell interference (called average power rise)

increase of f value (little i value)


3.3 UMTS shadowing and fast fading model DL Fast Fading (5)

How to compensate for fast fading losses in DOWNLINK?Case 1: slow moving UE (0-50km/h)As in uplink, power control compensates fairly well with a TX power increase the loss due to fast fading in the serving cell, but: Power Control Headroom (called Fast Fading Margin) necessary

for NodeB, but much smaller than in uplink, because: NodeB TX power is a shared power resource: the NodeB has to

compensate channel variations due to fast fading for all UEs in the cell

There is a very low probability that all UEs be in a fading dip at the same time

Typical value: 2 dB on the overall available power

Case 2: fast moving UE (>50km/h)same as in UL (see previous slides)



3.4 Calculation of Node B reference sensitivity

Objective:

to be able to calculate the reference sensitivity for a given service bit rate, BER, UE speed and UE multipath environment


3.4 Calculation of Node B reference sensitivity Definition of Reference_Sensitivity

The received Eb/No for a given UE at the NodeB reference point must apply:

Eb/No[dB] > (Eb/No)req[dB]Note: Eb/No=C/(I+N – C) + PG (definition, see §1.3) NodeB reference point=NodeB antenna

connector (see 3GPP 25.104)

[dB]N

N-CIN[dBm][dB] [dB]– PG (Eb/No)

)[dBm]N-C(I[dB] [dB]– PG (Eb/No)[dBm]C

req

req

min

minmin

Reference_Sensitivity [dBm]defined with reference to N it is service dependent

Interference Margin [dB]= Noise Rise [dB] –10log{1+ (Ec/No)req}see §3.5 for more details

NodeB

UE

As a consequence, the minimum received power Cmin shall apply:

NodeB antenna connector

Feeder

Antenna


3.4 Calculation of Node B reference sensitivity Calculation of Reference_Sensitivity

with: N=-108.1dBm+ NFNodeB =-104.1dBm (assuming NFNodeB=4dB)

PG is the Processing Gain (service dependent): PG=25dB for speech 12.2k PG=17.8dB for CS 64k PG=10dB for PS 384k

(Eb/No)req is a fixed value (see §2.2)Note: (Eb/No)req depends in UE speed and UE multipath environment (Vehicular A 50km/h...) in order to take into account the multipath diversity effect:

gain due to multipath combining in the rake receiver loss due to multipath fading holes (see §3.4)

N[dBm][dB] [dB]– PG (Eb/No)[dBm]nsitivity ference_Se req Re



3.5 UMTS interference modeling

Objective:

to be able to calculate the UL interference margin for a given traffic load

to be able to describe the interference effects in UL and in DL


3.5 UMTS interference modelingCalculation of interference margin

The NodeB reference_sensitivity is defined with reference to the fixed received „thermal noise at receiver“ N: it is necessary to apply a correction factor, called Interference Margin in order to take into account the effect of the movable received interference I:

} linear (Ec/No){e [dB] – Noise Risin [dB] ce MInterferen req ][1log10arg with: Noise Rise [dB] depends on the interference level I (ie on the

traffic load): I=Cmin Noise Rise ~ 0,2dB I=N Noise Rise=3dB I=3N Noise Rise=6dB

{10 log {1+ (Ec/No)req[linear]} typically between 0.1dB (for speech 12.2k) and 0.8dB (for

PS 384k) small value because (Ec/No)req (linear value) <<1 (the

useful signal level is always far below the noise floor in W-CDMA )

it can be neglected except for very high bit rates


3.5 UMTS interference modeling Noise Rise and Traffic load(1)

Definition:Cj[dBm]: received power of the transmitter j (UEj in UL, NodeBj in DL)Xj[%]: load factor for j defined as the contribution of j to the total noise (I+N)

Cj=Xj x (I+N)X[%]: load factor defined as the sum of the contributions for all transmitters

XUL=sumall UEs in the network(Xj) ; XDL=sumall NodeBs in the network(Xj)

We can demonstrate that: X

[dB]Noise Rise

1

1log10

Example in Uplink

0 5

10 15 20 25 30 35

0 11 21 31 41 51 61 71 81 91 100

XUL (%)

50% of cell load (3dB of interference)

max loading : 75%

Noise Risel (dB)


3.5 UMTS interference modeling Noise Rise and Traffic load(2)

UplinkNoise Rise and XUL are cell specific parameters (useful to characterize UL cell load)XUL can tend toward 100% (just by adding new UEs in the network) Noise Rise can tend towards infinity the system can be unstable.

DownlinkNoise Rise and XDL are UE specific parameters (not convenient)XDL can not tend toward 100% (because the TX power of NodeBs has a fix limit Noise Rise can not tend towards infinity the system can not be unstable.


3.5 UMTS interference modelingTraffic load and UL load factor (1)

Relationship between XUL and traffic load for one cell: Does XUL depend on:

the traffic mix? the user distribution in the serving cell? the user distribution in the surrounding cells?

XUL can be calculated analytically with the assumption that Iextra=f x Iintra with f constant value:

Answer:Does XUL depend on:- the traffic mix? yes (due to different (Eb/No)req values and PG values)- the user distribution in the serving cell? no (due to power control)- the user distribution in the surrounding cells? yes, but the most polluting users in the surrounding cells should stop to pollut by taking the serving cell in their active set (soft/softer handover) and being therefore power controlled by the serving cell

cell serving the in usersof number N with

FactorActivity rate Chip

Rate Bit ServiceNoEb1

FactorActivity rate ChipRate Bit Service

NoEb f)(1[%] X N

1kk

kkreq,

kk

kreq,UL


3.5 UMTS interference modelingTraffic load and UL load factor (2)

XUL typical values (commonly used): Very low loadXUL=5%Noise Rise=0.2dB Medium loadXUL=50%Noise Rise=3dB(typical default value) High loadXUL=75% Noise Rise=6dB (at the limit of system

instability)


3.5 UMTS interference modeling What about DL load factor?

As Noise Rise and XDL are not convenient to characterize the DL cell load, another parameter is commonly used:

Orthogonality effect In downlink, the orthogonality of channelization codes reduces

the intra-cell interference Iintra:Iintra [W]=(1-) x sumDL users in the cell (Ci) with Orthogonality

Factor =0no orthogonality Iintra= sumDL users in the cell (Ci) =1perfect orthogonality Iintra= 0 W

3GPP values for Orthogonality Factor : =0.6 for Vehicular A =0.94 for Pedestrian A

Note: there is no orthogonality effect in UL because the codes of UL physical channels come from different UEs and are therefore not synchronized each over.

cell[W] the for NodeBpower TX Maximumcell[W] the for NodeBpower TX[%] factor load powerDL



3.6 Calculation of cell range

Objective:

to be able to calculate the MAPL with a manual UL link budget and to deduce the cell range


3.6 Calculation of cell range Exercise: MAPLUL calculation (1)

Fixed assumptions: Antenna gainUE + Internal lossesUE = 0dB Antenna gainNodeB=18dBi Feeder and Connector losses=3dB Thermal noise=-108.1 dBm and NFNodeB=4dB

EXAMPLE 1: Service/UE mobility assumptions are given (see table EXAMPLE 1) Can you complete the table EXAMPLE 1?

EXAMPLE 2: EIRP, Reference_sensitivity, margins, losses and MAPL are given (see

table EXAMPLE 2) Can you find the service/UE mobility assumptions?



EXAMPLE 1— UL link budget for: UE power class 4 Speech12.2kbits/s Vehicular A 3km/h UE in soft(or softer) handover state with 2 radio links Deep Indoor Cell coverage probability=95%, =8 UL load factor=50%

Value in

Commentf.a.=fixed assumptio

n (see previously

)

A. On the transmitter sideA1 UE TX power dBm see §2.3A2 Antenna gainUE + Internal lossesUE dB f.a.A3 EIRPUE dBm A1+A2B. On the receiver sideB1 (Eb/No)req dB see §2.2B2 Processing Gain dB see §1.3B3 NFNodeB dB f.a.B4 Thermal noise dBm f.a.B5 Reference_SensitivityNodeB dBm B1-

B2+B3+B4(continuing on next slide)



EXAMPLE 1— continuing Value in Commentf.a.=fixed assumptio

n(see

previously)

C. MarginsC1 Shadowing margin dB see §3.3C2 Fast fading margin dB see §3.3C3 Noise Rise dB see §3.5C4 10 log {1+ (Ec/No)req} dB see §3.5C5 Interference margin dB C3-C4D. LossesD1 Feeders and connectors dB f.a.D2 Body loss dB see §2.2D3 Penetration loss (indoor

margin)dB see §2.2

E. GainsE1 Antenna gainNodeB dBi f.a.MAPL dB =?



EXAMPLE 2— UL link budget for: UE power class ? Service: ? Multipath Environment: ? UE in soft(or softer) handover state? Indoor margin:? Cell coverage probability=?, =? UL load factor=?

Value in

Commentf.a.=fixed assumptio

n(see

previously)

A. On the transmitter sideA1 UE TX power 24 dBm see §2.3A2 Antenna gainUE + Internal lossesUE 0 dB f.a.A3 EIRPUE 24 dBm A1+A2B. On the receiver sideB1 (Eb/No)req 3.2 dB see §2.2B2 Processing Gain 17.8 dB see §1.3B3 NFNodeB 4 dB f.a.B4 Thermal noise -108.1 dBm f.a.B5 Reference_SensitivityNodeB -

118.7dBm B1-

B2+B3+B4

(continuing on next slide)



EXAMPLE 2— continuing Value in Commentf.a.=fixed assumptio

n (see previously

)C. MarginsC1 Shadowing margin 4.8 dB see §3.3C2 Fast fading margin -0.3 dB see §3.3C3 Noise Rise 3 dB see §3.5C4 10 log {1+ (Ec/No)req} 0.1 dB see §3.5C5 Interference margin 2.9 dB C3+C4D. LossesD1 Feeders and connectors 3 dB f.a.D2 Body loss 3 dB see §2.2D3 Penetration loss (indoor

margin)8 dB see §2.2

E. GainsE1 Antenna gainNodeB 18 dBi f.a.MAPL 139.3 dB


3.6 Calculation of cell range Exercise: cell range calculation (6)

Can you complete the following table by using the simplified formula of the Alcatel Standard propagation model (see exercise in §3.2)?

Limiting Service Clutter class Cell Range [km]

Speech 12.2k

Dense urbanUrban

Suburban Rural

PS64

Dense urbanUrban

SuburbanRural

114




4. Initial Radio Network Design Session presentation

Objective: to be able to have the theoretical background to

create an initial network design using a RNP tool*: the aim is to fulfill the radio network requirements with lowest possible costs.

Program: 4.1 Positioning the sites on the map4.2 Coverage Prediction for CPICH RSCP 4.3 UMTS Traffic Simulations4.4 Coverage Predictions for CPICH Ec/Io and DL/UL services4.5 “Traffic emulation approach” or “fixed load approach”?

* the aim of this training is not to learn how to use A9155 RNP tool. There is another training course for that purpose (3FL 11195 ABAA Alcatel 9155 RNP Operation)


4. Initial Radio Network DesignOverview

Cell range calculation (see §3)

Positioning the sites on the map

(§4.1)

CPICH RSCP

coverage prediction

(§4.2)

Traffic simulation

(§4.3)

Coverage predictions(§4.4)- CPICH Ec/Io

-UL Eb/No-DL Eb/No

Basic radio network parameter definition (§5)

RNP requirements

fulfilled?

Fixed load default values

Traffic parametersPropagation model parameters

Network design parameters

Basic radio network

optimization (§6)

Traffic map

Traffic emulationapproach

Fixed loadapproach

Change network design parameters

Initial Radio Network Design

YES

NO

RNP requirements

fulfilled?

NO



4.1 Positioning the sites on the map

Objective:

to be able to get a coarse positioning of NodeB sites on the planning area and to apply a UMTS parameter set for network design parameters.


4.1 Positioning the sites on the map Calculation of inter-site distance

Manual Method: Description:

1. calculate MAPLUL for the limiting service by performing a manual UL link budget (see §3)

2. deduce the cell range and the inter-site distance:Inter-site distance = 1.5 x Cell Range for a 3-sectored site

Advantage: quick, because it can be performed by hand even if RNP tool and digital maps are not available yet.

Inconvenient: imprecise, because topographic data and detailed clutter data are not taken into account.

Typical inter-site distance: Dense urban: 350-450 m, Urban: 500-650 m, Sub-urban:900 -1200 m, Rural: 2000 - 3000 m


4.1 Positioning the sites on the map Site map

The sites are positioned in the planning area roughly respecting the inter-site distance for each clutter class: Existing GSM sites can be reused The sites should be positioned close to the dense traffic zones

(see traffic map in §2.2)

Inter-site distance

Planning area The initial site map is regularly updated based on site acquisition and site survey results.

Note: At this stage, search radii may already be issued, in order to start the long process of site acquisition

Site map


4.1 Positioning the sites on the map Network Design Parameters (1)

.Network design parameters – site wise Typical value Comment

Number of UL/DL hardware resources

R2: 2BB boardsR3: 4 BB boards see §2.3

Number of sectors 3



.Network design parameters – sector wise Typical value Comment

Number of carriers 1TMA usage no

Antennaparameters

model 65° horizontal beam width

azimuth 0°, 120° and 240° 3 sectored site

height 20-25m for urban30-35 m for suburban

gain 18dBidowntilt 6° mechanical +electrical

downtiltRXdiv yesTXdiv no

DL feeder and connector losses 3dB see §3.1

UL feeder and connector losses 3dB see §3.1

Noise Figure 4dB see §2.3



.Network design parameters – cell wisealso called Cell Parameters

Typical value Comment

see Appendix for a complete description of Cell Parameters. Here are only described the cell parameters which have an impact on traffic simulations and coverage predictions (§4)

Max. total power (for the cell) 43dBm see §2.3

CPICH (Pilot) power 33dBm 10% of Total powerOther common physical channels power 35dBm CPICH power + 2dB

AS threshold 3dB

maximum threshold between the CPICH Ec/Io of

the best transmitter and the CPICH Ec/Io of another

transmitter so that this transmitter becomes part

of the UE active set



4.2 Coverage Prediction for CPICH RSCP (=CCPICH=Pilot level= Pilot field strength)

Objective:

to be able to check that the CPICH RSCP coverage probability is in line with the network requirements

perform, interpret and improve a CPICH RSCP coverage prediction


4.2 Coverage Prediction for CPICH RSCP (=CCPICH =Pilot level)How to perform the prediction?(1)

Calculation Radius of NodeBj

Calculation Area of NodeBj

NodeBj

Virtual UE scanning the Calculation Areas of all

NodeBs

Step1: enter the prediction inputse.g. definition of Calculation Areas Planning Area


NodeB

Virtual UE

CPICH TX powerCPICH RSCP(=CPICH RX power)Pathloss Lpath

No shadowing (Shadowing margin=0dB in this

step)at each pixel*:CPICH RSCP[dBm] = CPICH TX power[dBm] +GainNodeB antenna [dB]

– LossNodeB feeder cables [dB] – Lpath [dB]

Step2: the tool calculates the CPICH RSCP values for the virtual UE (without considering shadowing effect)

*The calculation is performed for a given resolution, typically at each pixel of the Calculation Areas (see Step1)

4.2 Coverage Prediction for CPICH RSCP (=CCPICH =Pilot level) How to perform the prediction?(2)


4.2 Coverage Prediction for CPICH RSCP (=CCPICH =Pilot level) How to perform the prediction?(3)

Step3: the tool calculates the reliability level for each CPICH RSCP value (calculated in Step2) in order to consider the shadowing effect

(at each pixel) CPICH RSCP- (CPICH RSCP)minimum=Reliability Margin

with (CPICH RSCP)minimum =fixed value

Reliability Margin = f(Reliability Level, Standard deviation ) is given by the clutter map we can deduce a CPICH RSCP reliability level (per pixel)

Example: assume CPICH RSCP=-94 dBm, (CPICH RSCP)minimum =-104dBm, =6dB What is the reliability level for this CPICH RSCP value (use the curve in§3.3)?

Answer:Reliability Margin=10dB Reliability level=95% (=6dB)


From the radio network requirements (see §2.4), it is known: (CPICH RSCP)minimum

required Area Coverage Probability (typically 95%)

Area Coverage Probability: it is the average of all Reliability Levels per pixel (calculated in

Step3) over the Planning Area it can be calculated by a tool and has to be compared with the

required Area Coverage Probability

4.2 Coverage Prediction for CPICH RSCP (=CCPICH =Pilot level) How to interpret the prediction?

Reliability level=80%Reliability level=98%


Area coverage probability>required value?if yes, network design is OKelse network design has to be improvedReliability level=50%





PlanningArea


1. What happens if you have a bad CPICH RSCP coverage in an area?

2. Does the CPICH RSCP coverage depend on traffic load?

3. Which are the input parameters for the CPICH RSCP coverage prediction?

4. Shall the calculation radius be greater or smaller than the inter-site distance?

5. Make some suggestions to improve the prediction results

4.2 Coverage Prediction for CPICH RSCP (=CCPICH =Pilot level) Exercise



4.3 UMTS Traffic Simulations

Objective:

to be able to check that the network capacity is in line with the traffic demand by performing traffic simulations with a RNP tool


4.3 UMTS traffic simulationsWhy do we need traffic simulations?(1)

Traffic Map (see§2)Traffic demand modeling

Can the capacity cope with the demand in UL and in DL?

Site map (see §4.1)Network capacity modeling

it is necessary to calculate the UL/DL network capacity to check that it is in line with the traffic demand.


4.3 UMTS traffic simulationsWhy do we need traffic simulations?(2)

How to calculate the UL/DL network capacity? Problem: the capacity depends on the user distribution (at least

in DL)

Solution: a traffic simulation can be performed (= a snapshot of UMTS network at a given time, one possible scenario among infinite number of scenarii).

User distribution 1 User distribution 2

384k

12.2k

Cell

NodeB

12.2k

384k (in outage)

Cell

NodeB

Suburban environment class

Network capacity 1 > Network capacity 2 (for the same traffic map)


4.3 UMTS traffic simulationsHow to perform a traffic simulation?(1)

Traffic simulation inputs

typicalvalue Comment

Traffic simulation parameters (only used for traffic simulations)

Maximum UL load factor 75% limit of system instability. If this threshold is overcome, some UEs are put in outage.

Number of iterations 100 RNP tool dependent values. Trade off between precision and calculation time

Convergence criteria 3%

Orthogonality factor (per clutter) 0.6 0.6 for Vehicular A ; 0.94 for Pedestrian A

Traffic mapsee §2.2Propagation model parameterssee §3.2Network design parameterssee §4.1

Step 1: enter the traffic simulation inputs



Step 2: the RNP tool provides a realistic user distribution Used input: traffic map The RNP tool provides a snapshot of the network at a given time

(based on the traffic map and Monte-Carlo random algorithm): a distribution of users (with terminal used, speed and multipath

environment) in the planning area a distribution of services among the users a distribution of activity factors among the speech users in order

to simulate the DTX (Discontinuous Transmission) featureExample:

Mobile phoneVehicular 50km/h

Speech 12.2k (active)

PDAVehicular 3km/h

PS384

24 users



Step 3: the RNP tool checks the UL/DL service availability for each user Used inputs: user distribution (see Step1) +Propagation model

parameters+Network design parameters+ traffic simulations parameters

UL/DL link loss calculations are performed iteratively due to (fast) power control mechanisms in order to get: needed UE TX power for each UE needed NodeB TX power for each cell

Each of the following conditions is checked: if one of them is not fulfilled, the concerned user will be ejected (service blocked): Conditions in UL:

1) needed UE TX power < Maximum UE TX power2) UL load factor < Maximum UL load factor (typical value: 75%)3) enough UL NodeB processing capacity

Conditions in DL:1) CPICH Ec/Io < ( CPICH Ec/Io)required2) needed NodeB TX power < Maximum

NodeB TX power (ie DL Power load<100%)

3) (for each traffic channel) needed TX power < Max TX power per channel

4) enough DL NodeB processing capacity 5) needed number of codes < max number

of codes


4.3 UMTS traffic simulationsTraffic simulation outputs

DL (power) load factor per cell UL load factor per cell Percentage of soft handover Percentage of blocked service requests and reasons for blocking

(ejection causes)Example of ejection causes with A9155 RNP tool: the signal quality is not sufficient:

on downlink: not enough CPICH quality: Ec/Io<(Ec/Io)min

not enough TX power for one traffic channel(tch): Ptch > Ptch maxon uplink: not enough TX power for one UE (mob): Pmob > Pmob max

the network is saturated: the maximum UL load factor is exceeded (at admission or

congestion). not enough DL power for one cell (cell power saturation) not enough UL/DL NodeB processing capacity for one site (channel

element saturation) not enough DL channelization codes (code saturation)


4.3 UMTS traffic simulationsLimitation of traffic simulation

Limitation: a simulation is only based on one user distribution another simulation based on the same traffic map but on a

different user distribution can give different results for DL/UL service availabilities

Solution: to average the results of several simulations (statistical effect)

to be closer to the reality

Other interest of traffic simulation Some traffic simulation ouputs (that are DL (power) and UL load

factors per cell) can be used as inputs for CPICH Ec/Io and DL/UL service coverage predictions (see §4.4).



4.4 Coverage Predictions for CPICH Ec/Io and DL/UL services

Objective:

to be able to check that the coverage probabilities for UL/DL services are in line with the networks requirements by performing coverage predictions with an RNP tool


4.4. Coverage Predictions for CPICH Ec/Io and DL/UL services (based on traffic simulations)

Why do we need coverage predictions?

What is the coverage probability at this pixel for:-CPICH Ec/Io?-UL service coverage?-DL service coverage?

What is the probability for a user to get UL/DL services at a given point of the planning area?

Problem: traffic simulations can be used, but it is necessary to average an enormous number of traffic simulations (see§4.3) to get the answer for each service at each pixelunrealistic calculation time Solution: Coverage Predictions can be performed



Different types of coverage predictions CPICH RSCP prediction plot (see §4.2) CPICH Ec/Io prediction plot

Only the pilot quality from best server is considered (no soft handover)

Standard deviation: 3dB no UL/DL service coverage if CPICH Ec/Io < (CPICH Ec/Io)minimum

UL Coverage area prediction plots for each service soft/softer handover possible Standard deviation: same as clutter map values Uplink service area is limited by maximum terminal power.

DL Coverage area prediction plots for each service soft/softer handover possible Standard deviation: 3dB Downlink service area is limited by maximum allowable traffic

channel power



How to perform a coverage prediction?(1) Step 1: enter the Coverage Prediction inputsTraffic simulation inputs typical

value Comment

Coverage Predictions parameters (only used for predictions)Calculation Radius (per cell) 4 km same as for CPICH RSCP prediction (see §4.2)

Probe UE

Service parameters

see §2.2

The probe UE characterizes the service/terminal/multipath environment for which the Coverage Prediction is performed, e.g. PS64/PDA/Vehicular 3km/hNote: in case of CPICH/Io prediction, no service parameters are entered.

Multipath environment

Terminal parameters and indoor margin

UL load factor(per cell) 50% used to simulate UL/DL interference levelFixed load approach: same values for all cellsTraffic emulation approach: specific values for each cell (see §4.5)DL(power) load factor(per cell) 50%

(ratio value)minimum-15dB (typically) for CPICH Ec/Io ratio (see §2.4)(Eb/No)req values for UL/DL (Eb/No) ratios (see §2.2)

Stand. deviation (per clutter) 3dB for CPICH Ec/Io and DL (Eb/No) ratios, clutter map values for UL (Eb/No) ratio (typically 7-8dB)

Orthogonality factor (per clutter) 0.6 0.6 for Vehicular A ; 0.94 for Pedestrian A

Propagation model parameters(see §3.2) + Network design parameters(see §4.1)



How to perform a coverage prediction?(2) Step 2: calculation of the ratio values (e.g. CPICH Ec/Io values) at

each pixel A probe UE (causing no interference) is scanning each pixel of

the planning area. Pathloss calculations are performed for this probe UE to get the

ratio values:e.g. CPICH Ec/Io values per pixel or UL PS64 (Eb/No) values per pixel

Probe UE scanning each pixel of the calculation areas



How to perform a coverage prediction?(3) Step 3: calculation of the reliability level for each ratio value

(calculated in Step2) in order to consider the shadowing effect.(at each pixel) Ratio value - (ratio value)minimum=Reliability Margin

with (ratio value)minimum =fixed value Reliability Margin = f(Reliability Level, Standard deviation )

is given by the prediction inputs (see Step 1) we can deduce a reliability level (per pixel) for the ratio

value

Example:what is the reliability level for the following pixels(use the curve in

§3.3): CPICH Ec/Io value = -12 dB? UL (Eb/No) value= 4dB (for PS64, Vehicular 50km/h)? Answer:

CPICH Ec/Io (CPICH Ec/Io)minimum =-15dBReliability Margin=3dBk=1 (=3dB) Reliability level=84%UL (Eb/No)(Eb/(No)req=3.2dBReliability Margin=0.8dBk=0.1 (=8dB) Reliability level~50%



How to interpret a coverage prediction? From the radio network requirements (see §2.4), it is known:

(ratio value)minimum required Area Coverage Probability (for a given ratio)

Area Coverage Probability (for a given ratio): it is the average of all Reliability Levels per pixel (calculated in

Step3) over the Planning Area it can be calculated by a tool and has to be compared with the

required Area Coverage Probability



Area coverage probability>required value?if yes, network design is OKelse network design has to be improved





Planning Area



4.5 “Traffic emulation approach” or “fixed load approach”?

Objective:

to be able to describe the different approaches which lead to an acceptance test


4.5 “Traffic emulation approach” or “fixed load approach”? Traffic emulation approach(1)

Traffic map (§2.2)

Traffic simulations (§4.3)

Predictions (§4.4)

in line with RNP

requirements?

Result1

Change Network Design

Parameter(s)

Field traffic

emulation

Field measurement

s

Result2

Acceptance TestResult1=Result2?

yes

no

Fixed DL(power)/UL load factors per cell

RNP tool Field


4.5 “Traffic emulation approach” or “fixed load approach”? Traffic emulation approach(2)

Advantages: accurate (but the accuracy depends on the accuracy of traffic

map)

Disadvantages: complex:

traffic forecast and traffic map for the coming years must be provided by the operator

traffic simulations must be performed with RNP tool and if any parameter is changed, it is necessary to recalculate traffic simulations before recalculating coverage predictions

no acceptance test possible, because it is not realistic to emulate the traffic map in the field.


4.5 “Traffic emulation approach” or “fixed load approach”? Fixed load approach(1)

Default DL(power)/UL load factors values for

each cell”Fixed load”

Predictions (§4.4)

in line with RNP

requirements?

Result1

Change Network Design Parameter(s)

Field Fixed load emulation

Field measurement

s

Result2


yes

no

RNP tool Field



Advantages: simple: no need of traffic map and traffic simulations acceptance test can be realized, because “fixed load” can be

emulated and measured in the field (at least in DL, see further)

Disadvantages: inaccurate (no traffic map considered) all planning efforts targeting to optimize the network by

reducing traffic per cell can not be modeled by this approach (“Fixed Load Trap” effect): adding cells/sites

real effect: big enhancement of the total network capacity

modeled effect: little enhancement of the network capacity indeed, as the same load is mandatory for all cells (“fixed load”), the new cell/site will add (artificial) load and therefore bring a lot of (artificial) interference and only very little new capacity

downtilting antenna for one cell real effect: cell load decrease (because it makes the

cell area smaller) modeled effect: no cell load decrease (due to “fixed

load”)



How to emulate DL “fixed load” in the field? DL load can be emulated

with the OCNS (Orthogonal Code Noise Simulator) feature of the Alcatel NodeB: It generates artificial

interference in downlink

It is used to emulate downlink load and perform tests with a reduced number of UEs

Typical default value: 50% for DL (power) load factor

NodeB

Common channels

OCNS channels

Dedicated channels

AvailablepowerTXDLMaximumUETracepowerTXOCNS

loadDL powerDL

TX

__

(%)_

Virtualmobiles(due to OCNS)

Tracemobile

Realtraffic

Simulatedtraffic

Maximumoutput power



UE

AttTx

RxTx

Rx

RxTx

How to emulate UL fixed load in the field? UL load could be emulated by generating artificial interference

at the NodeB receiver (a kind of “UL OCNS feature”): such a feature is not provided by Alcatel NodeB.

Workaround: UL load can be emulated at the MS side by placing an Attenuator (Att) in the MS transmit pathTypical default value: 50% for UL load factor (ie 3dB Noise Rise, ie 3dB Attenuation)


4.5 “Traffic emulation approach” or “fixed load approach”? A medium approach(1)

Traffic map (§2.2)

Traffic simulations (§4.3)

Predictions (§4.4)

in line with RNP

requirements?

Result1

Change Network Design Parameter(s)

Field fixed load

emulation

Field measurement

s

Result2


yes

no

Fixed DL(power)/UL load factors per cell

RNP tool Field

Default UL load factor values for each

cell”Fixed load”

DL(power) load factor per cell


4.5 “Traffic emulation approach” or “fixed load approach”? A medium approach(2)

Alcatel strategy is to use the fixed load approach as it is measurable on the field and less ambiguous if commitments have to be fulfilled.

Nevertheless, a medium approach can be considered to overcome the disadvantages of the fixed load approach (see previous slide): Advantages:

accurate (but the accuracy depends on the accuracy of traffic map)

acceptance test can be realized Constraints:

traffic forecast and traffic map for the coming years must be provided by the operator

traffic simulations must be performed with RNP tool DL: the operator shall agree that the DL field traffic

emulation is realized from the traffic simulation outputs of the RNP tool

UL: default value for UL load factor must be taken for the whole network (no “UL OCNS feature”)

153




5. Basic Radio Network Parameter Definition Session presentation

Objective: to be able to define the basic radio network

parameters (neighborhood planning and code planning parameters)

Program: 5.1 Neighborhood planning5.2 Scrambling code planning



5.1 Neighborhood planning

Objective:

to be able to describe the criteria and methods used to perform neighborhood planning.


5.1 Neighborhood planning Overview

The purpose of neighborhood planning is to define a neighbor set (or monitored set) for each cell of the planning area The neighbor set is broadcasted in each cell in the P-CCPCH

and can therefore be accessed by each UE Each UE monitors the neighbor set to prepare a possible cell re-

selection or handover The neighbor set may contain:

Intra-frequency neighbor list : cells on the same UMTS carrier Inter-frequency neighbor list: cells on other UMTS carrier Inter-system neighbor lists: for each neighboring PLMN a separate list is

needed. Note: it is NOT the aim of neighborhood planning to define a ranking of the cells inside the neighbor set. This ranking is performed by the UE using UE measurements and criteria defined by UTRAN radio algorithms.

The neighborhood planning plays a key role in UMTS. Indeed, as UMTS is strongly interference limited, a wrong neighbors plan will bring interference increase and therefore capacity decrease.

e.g. if a possible soft handover candidate is not selected, because it is not in the neighbor list, it is fully working as “Pilot Polluter”


5.1 Neighborhood planning Criteria and methods

Criteria:Let’s consider one cell (called cell A). One or several of the following criteria can be used to decide to take a candidate cell as neighbor of cell A : the distance between cell A and the candidate cell is less than a

given maximum inter-site distance. the overlap area between cell A and the candidate cell is more

than a given minimum value. Note: overlap area between cell A and cell B = intersection between SA and SB, withSA[km2]=area where

(CPICH RSCP)cellA and (CPICH Ec/Io)cellA better than given minimum values (CPICH Ec/Io)cell A is the best

SB[km2]=area where (CPICH RSCP)cellB better than given minimum value (CPICH Ec/Io)cell B>(CPICH Ec/Io)cell A – (a given margin)

the candidate cell is a co-site cell (=cell of the same NodeB). cell A is neighbor of the candidate cell (neighbor symmetry).

Methods: manually (not possible to consider the overlap area criterion) with an RNP tool see example with A9155 tool on next slides


5.1 Neighborhood planning Automatic neighborhood allocation with

A9155(1)Neighborhood parameters

Typical value Comment

Minimum CPICH RSCP -105 dBm

parameters used for overlap area criterion

Minimum CPICH Ec/Io -18 dBEc/Io margin 8 dBReliability level 87%Minimum covered area 2%

Maximum inter-site distance

between 8km and 25km

8 km for dense urban and urban, 10 km for sub-urban and around 25 km for rural areas

Force co-site cells as neighbors Yes co-site cells=cells of the same

NodeB

Force neighbor symmetry Yes e.g. if cell A is neighbor of cell B, cell B will be neighbor of cell A

Max number of neighbors 14

Step1: enter input parameters


5.1 Neighborhood planning Automatic neighborhood allocation with

A9155(2) Step2: for each cell, A9155 RNP tool calculates the neighbor list as

follows if “Force co-site cells as neighbors=Yes”, co-sites cells are taken

first in the neighbor list. cells which fulfill the following criteria are taken in the neighbor

list: the maximum inter-site distance criterion the overlap area criterionNote: if the maximum number of neighbors in the list is

exceeded, only the cells with the largest overlap area are kept.

if “Force neighbor symmetry”=Yes, cells with a neighbor symmetry are taken in the neighbor list, under the condition that the maximum number of neighbors has not already been exceeded.



5.2 Scrambling code planning

Objective:

to be able to describe the criteria and the methods used to perform the scrambling code planning


Scrambling code planning in UMTS FDD is similar to frequency planning in GSM. However it is not such a key performance factor:

it concerns only DL scrambling code (channelization codes and UL scrambling codes are automatically assigned by the RNC)

In contrast to frequency planning, it is not crucial which scrambling codes are allocated to neighbors as long as they are not the same code.

5.2 Scrambling code planning Overview


DL scrambling codes: used to separate cells restricted to 512 (primary) scrambling codes (easy planning)

Criteria: the reuse distance between two cells using the same

scrambling code inside one frequency shall be higher than 4 x inter-site distance

(preferable) the same scrambling code should not be used in two cells of the same sector

Methods manually with a RNP tool (see see example with A9155 tool on next slide)

5.2 Scrambling code planning DL scrambling code planning (1)


Method with a RNP tool:Note: Neighborhood planning (see §5.1) must be performed before performing scrambling code planning, because neighborhood relationships are used in the following method.

1. define the set of allowed codes for each cell (there can be some restrictions for cells at country borders)

2. (optional) define the set of allowed codes per domain (one domain per frequency)

3. define the minimum reuse distance

4. define forbidden pairs (for known problems between two cells)

5. run automatic code allocation and check consistency A9155 assigns different primary scrambling codes to a given cell i and to its

neighbors. For a cell j which is not neighbor of the cell i, A9155 gives it a different code:

If the distance between both cells is lower than the manually set minimum reuse distance,

If the cell i / j pair is forbidden (known problems between cell i and cell j). A9155 allocates scrambling codes starting with the most constrained cell and

ending with the lowest constrained one. The cell constraint level depends on its number of neighbors and whether the cell is neighbor of other cells.

5.2 Scrambling code planning DL scrambling code planning (2)


5.2 Scrambling code planning Definition of UL scrambling code pool for a RNC

UL scrambling codes: used to separate UEs more than one million of codes available (very easy planning) 2 different UEs mustn’t have the same code (inside one

frequency)

Criterion for definition of UL scrambling code pools: 2 RNC mustn’t have the same scrambling code in their pool

Method: each RNC is assigned manually a unique pool of codes (e.g. 4096 codes in R2)

Note: when a UE performs a connection establishment to UTRAN (RRC connection), the Serving RNC will assigned dynamically an UL scrambling code out of its pool to the UE. The code is released after RRC connection release.

Module 6 Planning

Documents