Propagation Modeling for Umts

Commercial-in-Confidence

Propagation Modeling for UMTSPropagation Modeling for UMTS

Propagation Modeling for UMTS


ContentsContents• UMTS spectrum• Model types• Map data requirements• CW measurements• Model calibration exercise



Session AimsSession Aims• This session answers the following questions:

! What are the characteristics of a UMTS radio frequency?! What is a propagation model?! What are the main types of propagation model available?! What map data is required for a propagation model?! How do I calibrate a propagation model?



3rd Generation Spectrum Allocations3rd Generation Spectrum Allocations

1800 20501900 1950 20001850 2100 2150 2200

ITU(WARC-92)

Europe

Japan

Korea

USA

1885 1980 20102025 2110 2170 2200

1920 1980 20102025 2110 2170 2200

1920 1980 2110 2170

2110 21701920 1980

1850 1910 1930 1990 2110 2200

MSS MSSIMT-2000

Land Mobile

IMT-2000

Land Mobile UL

IMT-2000

Land Mobile UL

IMT-2000

Land Mobile

IMT-2000

Land Mobile DL

IMT-2000

Land Mobile DL

UMTS

Paired UL

UMTS

Paired DLUMTS

SATUMTS

SAT

UMTS

UnpairedUMTS

Unpaired

IMT-2000

Land Mobile

PCS

ULPCS

DLReserved

1900

DECTGSM 18001880



The UMTS Radio CarrierThe UMTS Radio Carrier• 5MHz bandwidth.• 200kHz Rasterised carriers.

! Centre frequency can be adjusted in 200kHz steps.

• Adjacent Channel Performance! Base Station

"ACLR1 = 45dBm"ALCR2 = 50dBm

! Mobile"ALCR1 = 21dBm"ALCR2 = 24dBm

fn fn+1 fn+2

Power level

ALCR2ALCR1

Power level Power

levelTX Power

5MHz

5MHz

5MHz



UK Spectrum AllocationsUK Spectrum AllocationsD D DE E EC C CA A A A B B B

Hutchison

One2One

Vodafone

BT Cellnet

Orange

14.6MHz 14.8MHz10MHz 10MHz 10MHz0.3MHz0.3MHz

20MHz



What is a propagation model?What is a propagation model?• A mathematical model used by computer planning tools to

predict coverage from a radio transmitter.! Typical inputs

" 3D terrain data" Land use data (Clutter)" Building outlines" Building heights



Model Types (1)Model Types (1)Empirical (Statistical)

• Well understood. 2G standard.• Lower computational overhead.• Wide variety available.• Less accurate in small urban

cells. • Subjective clutter classification.• Only accurate in environment

where measurements were taken.

Deterministic (Ray tracing)• Few proven models available• Higher computational overhead• More map data required• More suited to small urban sites• Higher accuracy• Higher resolution


Street Canyon• More map data required• More suited to microcells


Model Types (2)Model Types (2)• In practice, few planning tools use a single mode of model and most

are a hybrid of these methods.• Macrocell models are normally based on statistical techniques.• A large variety of Microcell models have been proposed.



Statistical ModelsStatistical Models• Carrier wave measurements are

made from test transmitters.• The measurements are plotted

vs. log(distance).• A straight line is fitted through

the data.• A basic y=mx +c formula can be

used to estimate path loss.• The formula can be modified to

account for other factors eg. Txheight, Rx height & terrain effects.

Plot of measurements vs. log(distance)



Deterministic ModelsDeterministic Models• The planning tool traces rays

from each site through an accurate 3D representation of an urban area.

• Typically a major component of the calculated path loss is free space loss.

• The effect of reflections diffraction and absorption due to buildings can be incorporated into the model.

Reflected rayDiffracted ray



PopularPopular MacrocellMacrocell ModelsModels• Okumura - Hata

! Industry standard.! Not well suited for ranges under 1km.! Upper frequency limit 2GHz.

• Wavecall! Not so well accepted.! Fully deterministic no calibration required.! Suitable for urban macro and microcells.! Available for many common planning systems.



Popular Popular Microcell Microcell ModelsModelsPropagation Modeling for UMTS

Prediction Model Method Used Parameters 2D 3D Over Roof-top Required Data ResultsUniversity of Lund Dual Slope

Empirical− √ × × 2D Building

LayoutPath Loss

University of Karlsruhe Ray Launching Reflected +Penetrated Rays

√ × × 2D BuildingLayout

Path Loss + CIR

CNET Ray Tracing9 Reflections +Ground + Street

Diffraction√ × ×

2D BuildingLayout Path Loss

Swiss Telecom Ray Tracing Reflection +Diffraction

√ × × 2D BuildingLayout

Path Loss + CIR

Ericsson Mathematical − √ × WalfischIkegami Method

2D BuildingLayout

Path Loss

Telecom Street Canyon − √ × Walfisch BertoniMethod

2D BuildingLayout

Path Loss

Walfisch Ikegami Semi -Empirical − √ × WalfischIkegami Method

2D BuildingClasses

Path Loss

University of Valencia Ray TracingScattering +Diffraction

√ × Walfisch BertoniMethod

2D BuildingLayout +

Building Height

Path Loss

CNET Ray Launching Reflection +Diffraction

× √ × 3D BuildingLayout

Path Loss + CIR

ASCOM ETH Ray TracingReflection fromWall and Street × √ ×

2D BuildingLayout +

Building HeightPath Loss + CIR

University of Stuttgard Ray Launching6 Reflections + 2

Diffractions × √ ×2D Building

Layout +Building Height

Path Loss + CIR

University of Karlsruhe Ray TracingThe Most

Dominant Raysare Considered

× √ ×2D Building

Layout +Building Height

Path Loss + CIR


Map data requirements (1)Map data requirements (1)• The minimum requirements for

map data are:! Height! Clutter

• Some propagation models also require building information:

! Building vectors! Building rasters



Map data requirements (2)Map data requirements (2)• The resolution of map data will depend on several things.

! Budget" You can pay as much as you like for map data." Better data requires deeper pockets.

! Area type" Urban areas warrant higher resolution data (1 - 50 m)." Rural areas warrant lower resolution data (50 - 200 m).

! Model type" Microcell models require better data than Macrocell models." Macrocell models typically use 20 - 200 m." Microcell models typically require 1 - 20 m.



Carrier Wave (CW) MeasurementsCarrier Wave (CW) Measurements• CW measurements are

accurate radio measurements used to calibrate propagation models.

• A number of temporary test sites are used for the test transmissions.

• Signal strength measurements and GPS fixes are made along predefined routes.

• These measurements must be averaged before they can be used for model calibration.

GPS



CW Test SitesCW Test Sites• A typical network would require 3 different propagation models eg.

! Urban! Suburban! Rural

• Each model would require measurements from 10 - 15 test sites for calibration.

• The test sites used for propagation modeling should:! Be representative of typical cellular sites.! Should be free of obstructions.



CW Test RoutesCW Test Routes• CW test routes must be planned carefully.• Routes must avoid:

! Elevated sections of road.! Cuttings.! Tunnels.! Bridges.

• Sufficient measurements must be made in each clutter type for the model to be valid.

• Typically the distances driven for each site would be in the order of! 80km per urban test site.! 160km per rural test site.



CW Measurement EquipmentCW Measurement Equipment• Measurements should be distance based.

! Measurements should be taken no closer than 0.38λ.

• Equipment can be:! Distance triggered.! Time triggered.

• GPS outputs position every 1 second.! Position interpolation is required for each measurement.

• Test mobile measurements are NOT suitable.


Propagation Modeling for Umts

Documents

umtspropagation modeling

mathematical model

heightspropagation modeling

confidencepropagation

3mhz20mhzpropagation

land mobileimt

land mobile ulimt

single mode of model