Umts FDD Rnp

8/12/2019 Umts FDD Rnp

1/40

Radio Planning Aspects for UMTSMhlbauer Helmut

11.05.1999; 1

commercially not binding

Radio Network Planning Aspects for

UMTS Networks

Munich, June 1999

Helmut Muehlbauer

Siemens AG, Munich

Tel:+49-89-72234563

e-mail:[email protected]


2/40


11.05.1999; 2


Outline

Introduction

Basic WCDMA Aspects

Planning Aspects for WCDMA

Summary


3/40


11.05.1999; 3


CDMA Principles


4/40


11.05.1999; 4


Multiple Access Techniques

Time Time Time Time

Power Power Power Power

Freq

uenc

y

Freq

uenc

y

Freq

uenc

y

Freq

uenc

y

TDMA FDMA FDD-CDMA TDD-CDMA

FDMA - Frequency Division Multiple Access

TDMA - Time Division Multiple Access

CDMA - Code Division Multiple Access


5/40


11.05.1999; 5


Direct Sequence (DS) - CDMA

Each user gets an unique spreading code sequence

Information bandwidth R is spread with the code sequence

Transmission bandwidth W is much larger than information bandwidth R

At the receiving end same code is used for the despreading of the data(synchronization on chip level necessary)

Ratio W/R is called the processing gain

Data signal

Information

rate R

Spreading

code

Data

signal

De-spreading

code

= Spreading

code

Filter

Bandwidth

R

Wideband signal

Spread signal


6/40


11.05.1999; 6


Direct Sequence (DS) - CDMA

Data signal is directly modulated by the spreading code signal

Spread data signal is modulated with the wideband carrier

Basic properties: Multiple access due to loss cross-correlation of the different spreading

codes coherent receiver puts only a small ratio of the interfering user power into the

information bandwidth

Multipath interference due to ideal/high auto-correlation of the spreadingcodes

cross-correlation outside the time window +/-Tchip is nearly zero

signals (multipath) delayed more than 2 x Tchip are treated as interferenceand only a small ratio of the power is put into the information bandwidth

Narrowband interference is reduced by the processing gain due to thespreading of the signal onto the wideband bandwidth

interference signal power level is reduced by the spreading gain (W/R)


7/40


11.05.1999; 7


Main Aspect of CDMA

All users are transmitting in the same RF band

The users are separated via orthogonal spreading codes

Typically a CDMA network is interference limited

Capacity and Qualityis limited by the amount of

interference power in the system

Capacity and Quality

is limited by the amount of

interference power in the system

Interference power in the system

has to be minimized

Interference power in the system

has to be minimized


8/40


11.05.1999; 8


Uplink Capacity in a single Cell in a single Cell Network

Omnidirectional Cell

All users share the same power

Perfect power control assigns

to all users the same power

level at the receiver

Simple capacity equation for single cell

Introduction of energy per bit per noise

power density Eb/No

Power

Frequency

User M

User M-1User M-2

User M-3

User 1C

I

( ) ( )

( )

CIM

MI

C

kTWCMkTWCM

C

I

C

+=

=

>>+

=

13.

1

12.

11

1. ;

00

0

12.

1.

NE

RW

NE

RWM

W

R

N

E

I

C

bb

b

+=

=


9/40


11.05.1999; 9


Uplink Capacity in a single Cell in a multiple Cell Network

Frequency Reuse Efficiency

Interference increases due to the

intercell interference from the

neighboring cells

We assume a homogenous network

with equal number of users in each cell

If the number of users in neighboring

cells increases the max. achievable

capacity in the cells will decrease

A frequency reuse of 1 is planned

But typically the reuse efficiency islower;

Reuse Fraction

Reuse Factor

Reuse Efficiency

Uplink Capacity

700330ceinterferenintracell

ceinterferenintercell,...,==

+= 1F

+=

1

1eF

e

b

FNE

RWM =

0


10/40


11.05.1999; 10


Uplink Capacity in a single Cell in a multiple Cell Network

Consideration of Voice Activity and Sectorization

Voice activity factor Range of 0,50,6

Typical value: 0,57

Sectorization gain Capacity gain for one site

consisting of

3 sectors: about 2.5

6 sectors: about 5

1

0

= eb

FNE

RWM

=1

0

eb

FNE

RWM

Pole CapacityPole Capacity


11/40


11.05.1999; 11


Cell Load Evaluation

Cell load calculation for one single bearer (M is the number of

required links including blocking)

Cell load calculation for different bearers i

=e

buser

FRW

NEM

10

==

i

ii

i

biuser

e

ieeiiei

bi i

iuseri

RN

EM

FW

FFFRW

NEM

1 0,

,,

0

1 1

,

1

;1


12/40


11.05.1999; 12


Downlink Capacity

The downlink capacity is defined by the intercell

interference, the intracell interference and the

orthogonality of the spreading codes at the receiver

( ) 01 NII

RWS

I

C

erra ++

=

intint


13/40


11.05.1999; 13


Example: Voice 8kbps

Calculation of number of

simultaneous links on the uplink

for voice 8kbps

Ideal conditions

ideal power control

voice activity of 0,5

100% load

W/kcps 4096

frequency efficiency 0,58

soft handover 0,4

Voice

Bearer/kbps 8

Eb/No in dB (UL) 5,4

Eb/No 3,47link activity 0,5

cell load 1

# simult. links 172,3

# simult. links incl.

SHO103,4


signaling overhead(15%)

87,9

# links incl. 2%

blocking (ETSI)76

W/kcps 4096

frequency efficiency 0,58

soft handover 0,4

Voice

Bearer/kbps 8

Eb/No in dB (UL) 5,4

Eb/No 3,47link activity 0,5

cell load 1

# simult. links 172,3


SHO103,4


signaling overhead(15%)

87,9

# links incl. 2%

blocking (ETSI)76


14/40


11.05.1999; 14


Utran Network

Core Network

RNC RNC RNC

Node B Node B Node B Node B Node B Node B

Iu Iu Iu

Iur Iur

Iub Iub Iub IubIub Iub

Iur

UTRAN

UE


15/40


11.05.1999; 15


Softer Handover

One mobile station is

connected to several sectors

belonging to the same site

(NodeB)

NodeB: The combining of the

signals are performed by the

fingers of the rake receiver

MS: The combining of the

signals are performed by the

fingers of the rake receiver

Macro diversity Number of MS in softer

handover: 1015%

Softer HO Area


16/40


11.05.1999; 16


Soft Handover

The MS is connected to more

than one NodeB simultaneously

Uplink:

The signal combining is

performed via the RNC

Typically a selection combiningis used

Macro diversity

Downlink:

Different NodeBs are

transmitting same information

Macro diversity versusinterference increase

Number of MS in Soft HO:

30%...40%

RNC

Soft HO Area


17/40


11.05.1999; 17


Network Design and Dimensioning

Cell Range Estimation

Capacity Estimation

Capacity versus Coverage Optimization


18/40


11.05.1999; 18


Initial Network Dimensioning

Input Parameter

Traffic requirements:

Type of services: HMM, MMM,

Amount of traffic in Erlang or Mbyte

Traffic distribution

User distribution

Coverage requirements

coverage probability

penetration loss

site (GSM900/1800) reuse


19/40


11.05.1999; 19


UMTS Radio Network Planning Process

Network SimulationLoad maximization

Coverage versus capacity optimization

Quality optimization

Input data

Traffic

requirements

Coverage

requirements

Quality

requirements

Link budget

requirements

Prediction

model

Traffic model

Initial Cell

Plan

Site-to-site

distance

Simplified traffic model

Number of sites

Number of

carriers

Standard site

configuration

Final Cell

PlanUser & service & traffic

distribution

User mobility

Final site locationOpt. site configuration

System features


20/40


11.05.1999; 20


Bearer Services

Service User data rate Required SF Symbol rate Transport block s ize

Speech 8 kbps 128 (UL & DL) 32 ksps 80 bits (interl. 1 frame)

160 bits (interl. 2 frames)

LCD data 64 kbps

144 kbps

384 kbps

16/32 (UL/DL)

8/16

4/8

256 ksps

512 ksps

1024 ksps

640 bits (5120 bits - 80 ms)

1440 bits (11520 bits - 80 ms)

20480 bits (1633840 bits - 80 ms)

UDD data 30.4 kbps

60.8 kbps

243.2 kbps

32/64 (UL/DL)

16/32

4/8

128 ksps

256 ksps

1024 ksps

304 bits per frame

304 bits *2 per frame

304 bits * 8 per frame


21/40


11.05.1999; 21


Estimation of Cell Load - offered Traffic

Circuit switched data

separately performed for each bearer

overall required number of links is the sum of the different required

links for each bearer

assumption due to lack of more accurate traffic modeling (multi

dimensional Erlang formula) more accurate modeling will be available end of 09/99

Offered

Load

Bearer i

Blocking

Requirement

Soft

Handover

Signaling

Overhead

Required

Links


22/40


11.05.1999; 22


Estimation of Cell Load - offered Traffic

Packet switched data

approximation: packet switched data are regarded as circuit

switched data with mean data throughput and a mean call

duration time

mean call duration=data volume/mean data throughput

separately performed for each bearer

overall required number of links is the sum of the different requiredlinks for each bearer

assumption due to lack of more accurate traffic modeling

more accurate modeling will be available end of 09/99

Offered

Load

Bearer i

Soft

Handover

Signaling

Overhead

Required

Links


23/40


11.05.1999; 23


Offered traffic per user

Circuit switched traffic is considered with blocking (Erlang B)

Packet switched traffic is assumed to be circuit switched

with a mean data throughput but without blockingrequirement

Offered Traffic

service bearer symmetriemean data

rate/kbps

call

duration

in s

amount of

data in

Mbyte

BCHAequivalent

merl

voice voice symmetric 8 120 - 0,8 26,67videophonie LCD64 symmetric 64 240 - 0,02 1,33

slow internet UDD144 asymmetric 64 250 2 0,05 3,47


24/40


11.05.1999; 24


Subscriber Distribution

Subscriber distribution per clutter and service type

area in sqkm 10000

population 5000000

environment penetration pop.distribution

areadistribution

pop/sqkm sub/sqkm voice video-phonie

slowinternet

user/voice/sqkm

user/videophonie/sqkm

user/internet/sqkm

urban 30% 40% 10% 2000 600 50% 20% 20% 300,00 120,00 120,00

suburban 25% 30% 20% 750 188 50% 5% 10% 93,75 9,38 18,75

rural 25% 30% 70% 214 54 30% 1% 5% 16,07 0,54 2,68

area in sqkm 10000

population 5000000

environment penetrationpop.

distribution

area

distributionpop/sqkm sub/sqkm voice

video-

phonie

slow

internet

user/voice/

sqkm

user/video

phonie/sqkm

user/internet/s

qkm

urban 30% 40% 10% 2000 600 50% 20% 20% 300,00 120,00 120,00

suburban 25% 30% 20% 750 188 50% 5% 10% 93,75 9,38 18,75

rural 25% 30% 70% 214 54 30% 1% 5% 16,07 0,54 2,68


25/40


11.05.1999; 25


Offered Traffic and Capacity Evaluation

Offered traffic for urban environment per sqkm

Cell load = 56% for cells with area of 1 sqkm

service

equivalent

merl

user/service

/sqkm

offered

load in erl blocking

required

links

req. links

incl. SHO

&

signalling

load %

voice 26,67 300,00 8,00 2% 14 26 19%

video honie 1,33 120,00 0,16 2% 2 4 33%

slow internet 3,47 120,00 0,42 - 0,42 1 4%


26/40


11.05.1999; 26


Coverage Planning - Cell Breathing Effect

All user share the available power and interferes each other

Increasing number of active user increases the interference level in the

system

High load results in a high

interference and in ashrinking cell size

Cell load is considered

via an interference

margin in the link budget

I=10log(1-load)

Reduced through adaptive

antennas and multiuser detection 02

4

6

8

10

12

14

16

18

20

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%

Load Factor


27/40


11.05.1999; 27


Coverage Estimation - Link Budget

Output power and RxSensitivity of NodeB/UE

Diversity gain, Soft/softer handover gain

Penetration loss, body loss

Interference margin - load factor on uplink

Location probability at cell border or within cell area

Site configuration (antenna height/gain NodeB/UE, feeder loss)

Uplink:

Downlink:

problocLossGainGainloadfactor

bE

WR

WNodeBfeederLNodeBantGUEfeederLUEantGUEPL

bodySHOdiversityreq .Margin1

1log10)

0

(

)0log(10log10,,,,max

++

+++=

problocLossGainGainbE

WRWUEfeederLGNodeBfeederLNodeBantGNodeBPL

bodySHOdiversityreq

UEant

.Margin)

0

(

)0log(10log10,,,max ,

++

+++=


28/40


11.05.1999; 28


Coverage Estimation

Cell Sizes (uplink) for Voice and LCD Bearer

Cell Range for WCDMA(urban area (inbuilding @15dB pen.loss); hant=30; 90% area loc. prob.)

0,40

0,60

0,80

1,00

1,20

20% 40% 60% 80%Load Factor

Speech vehicular

Speech GSM1800

LCD64 vehicular

LCD144 vehicular

LCD384 vehicular

Speech

LCD144

LCD64

LCD384

Speech

GSM1800


29/40


30/40


11.05.1999; 30


Balancing of coverage and capacity

high traffic density

cell design for maximized capacity

high cell loading

small cell area

HCS network; simulations necessary; hard hand off!? (critical inIS-95)

low traffic density

cell design for maximized capacity

low cell loading (more hardware)

large cell area restriction on availability of certain high bit rate services


31/40


11.05.1999; 31


0,00

0,10

0,20

0,30

0,40

0,50

0,60

0,70

0,80

0,90

1,00

30% 40% 50% 60% 70%cell load

cell range

max. capacity regardingthe cell load due tooffered traffic

max. coverageregarding the cell loaddue to offered traffic

Optimization of Coverage versus Capacity

optimizedoperating point

Optimization of coverage versus capacity based on a given traffic distribution


32/40


11.05.1999; 32


Interference reduction

Similar problem as in GSM

optimum would be a homogeneous network

reality: inhomogeneous network

topography, morphology traffic distribution, subscriber mobility and different cell loadings

site-to-site distance and grid

antenna heights, alignment and types


33/40


11.05.1999; 33


Site

Configuration

Cell Layout


TrafficType and

Distribution

Analysis and

Optimization Tool

Interference

balanced

Network

Homogeneous

Network

Maximum

Capacity

Topography

Morphology


34/40


11.05.1999; 34



Antenna types:

omnidirectional

effective during network launch period with low traffic load

but: no possibilities for network tuning; only change of antennas

interference reduction has to be considered carefully during siteselection (height and location)

120 antennas:

suburban and rural areas with low load

fixed sector to sector distance of 120

65 antennas:

high flexibility during planning and tuning period but requires dense grid planning


35/40


11.05.1999; 35


Tools for Initial Cell Planning

Link Budget and Cell Range Calculation Tool

Traffic Analysis Tool


36/40


11.05.1999; 36


Link Budget and Cell Range Calculation Tool

Based on ETSI Simulations

Cost-231-Hata model seems to be appropriate

Log normal fading is about 7...8dB

Additional gains and margins will be defined in the systemlevel simulation process - not yet finished


37/40


11.05.1999; 37


Traffic Evaluation Tool

Number of

basic channels

(technology)

Scaling due to

interference

signaling

load

Number of available

basic channels

per area unit

Subscriberprofile

Traffic model

Blocking

Offered traffic

in basic channels

per area unit

Cell range

per

clutterMapping ofservices

on basic

channels

Number of Carriers


38/40


11.05.1999; 38


Basic Channel Concept

Basic channel:

voice bearer of 8kbps

spectrum efficiency: 71kbps/MHz/cell

cell capacity: 71/8kbps x 5MHz/0,5=89erl/cell

Circuit switched data

example: LCD144


cell capacity: 210/144kbps x 5MHz=7,3erl/cell

#basic channel: 89/7,3=12,2 basic channel


39/40


11.05.1999; 39


Basic Channel Concept

Packet switched data

UDD144 planned with 60,8kbps max. info data rate


cell capacity: 290/17,1kbps x 5MHz=85erl/cell

mean throughput of 17,1kbps during whole session time

#basic channel: 89/85=1,05 basic channel

Model is a first estimation

till now no exact simulations available

queuing has to be modeled

admission control and congestion control has to be modeled

approximately available 10/99


40/40


11.05.1999; 40

commercially notbinding

Summary

Initial Cell

Plan

Site-to-site

distance

Simplified traffic model

Number of sites

Number of

carriers

Standard site

configuration

Input

Parameter

Traffic/user

distribution

Coverage requirement

QoS

requirement

System

parameter

Initial Cell

Plan can

be defined

Uncertainties

& Problems

Modeling of

UDD data

Propagation model adjust-

ment to wide-

band application

System parameter

(Eb/No, SHO

gain, ...)

Admission,

Congestion,

Load control

Impact on

Initial

Cell Plan

Umts FDD Rnp

Documents