RADIO SYSTEMS – ETIN15 Lecture no: 4 · 2012-03-21 Ove Edfors - ETIN15 17 COST 231-Walfish-Ikegami model Background The Okumura-Hata model is not suitable for micro cells or small

Ove Edfors, Department of Electrical and Information [email protected]

RADIO SYSTEMS – ETIN15

Lecture no:

2012-03-21 Ove Edfors - ETIN15 1

4

Channel modelsand antennas


Contents

• Why do we need channel models?• Narrowband models

– Review of properties– Okumura’s measurements– Okumura-Hata model– COST 231-Walfish-Ikegami model

• Wideband models– Review of properties– COST 207 model for GSM– ITU-R model for 3G

• Antennas– Efficiency and bandwidth– Mobile station antennas– Base station antennas– Dipole and parabolic antennas


WHY DO WE NEEDCHANNEL MODELS?


Why do we need channel models?

During system design, testing and type approval:

Simple models reflecting the important propertiesof important channels (best, average, worst case)

During network design:

More detailed models appropriate for certaingeographical areas

Models used to make sure that the system designbehaves well in typical situations.

Models used to obtain an efficient network in termsof base station locations and other parameters


NARROWBANDMODELS


Narrowband modelsReview of properties

Narrowband models contain ”only one” attenuation, which is modeledas a propagation loss, plus large- and small-scale fading.

Large-scale fading: Log-normal distribution (normal distr. in dB scale)

Small-scale fading: Rayleig, Rice, Nakagami distributions ... (not in dB-scale)

Path loss: Often proportional to 1/dn, where n is the propagation exponent. (n may be different at different distances)

NOTE: Several of these models are found in an on-line appendix of the textbook which can be downloaded from the publisher's website (see “Literature” on course web).

Printed copies of textbook appendices areallowed during Part B of the written exam.


Okumura’s measurementsBackground

Extensive measurement campaign in Japan in the 1960’s.

Parameters varied during measurements:

FrequencyDistanceMobile station heightBase station heightEnvironment

100 – 3000 MHz1 – 100 km1 – 10 m20 – 1000 mmedium-size city, large city, etc.

Propagation loss is given as a median value (50% of thetime and 50% of the area).

Results from these measurements are displayed in figures7.12 – 7.14.


Okumura’s measurementsHow to calculate the prop. loss

Free spaceattenuation

1. We start by calculating the free-space attenuation

2. Apply a frequency and distance dependent correction

3. Apply a BS-height and distance dependent correction

4. Apply a MS-height, frequency and environment dependent correction

Fig. 7.12 Fig. 7.13 Fig. 7.14

OkuL


Okumura’s measurementsExample

Propagation at 900 MHz in medium-size citywith 40 m base station antenna height and1.5 m mobile station antenna height.

Use Okumura’s curves to calculate thepropagation loss at a distance of 30 kmbetween base station and mobile station.

2012-03-21 Ove Edfors - ETIN15 10

Okumura’s measurements1. Calculate free-space loss

Attenuation between twoisotropic antennas in freespace is (free-space loss):

The obtained value does not dependon antenna heights.

900 MHz and30 km distance

=> 121 dB

Lfree∣dBd =20 log 4 d

2012-03-21 Ove Edfors - ETIN15 11

Okumura’s measurements2. Apply correction for excess loss

Exce

ss loss

[d

B]

Frequency [MHz]

Dis

tan

ce [

km

]

These curves are only for

hb=200 m and hm=3 m

900 MHz and30 km distance

=> 36.5 dB

FIGURE 7.12

2012-03-21 Ove Edfors - ETIN15 12

Okumura’s measurements3. Apply correction of BS height

BS height [m]

Dis

tance

[km

]

Corr

ect

ion

fact

or

[dB

]

40 m BS and30 km distance

=> -16 dB

Note: Lower base station means INCREASING

attenuation => subtract this number.

FIGURE 7.13

2012-03-21 Ove Edfors - ETIN15 13

Okumura’s measurements4. Apply correction of MS height

MS height [m]Fr

equency

[M

Hz]

Corr

ect

ion

fact

or

[dB

]

Note: Lower mobile station means INCREASING

attenuation => subtract this number.

1.5 m MS and900 MHz inmedium-size city=> -3 dB

FIGURE 7.14

2012-03-21 Ove Edfors - ETIN15 14

Okumura’s measurementsSummary of example

Propagation loss (between isotropic antennas) usingOkumura’s measurements:

| 121 36.5 ( 16) ( 3) 176.5 dBOku dBL = + − − − − =

Calc. step: 1 2 3 4

2012-03-21 Ove Edfors - ETIN15 15

The Okumura-Hata modelBackground

In 1980 Hata published a parameterized model, based on Okumura’smeasurements.

The parameterized model has a smaller range of validity thanthe measurements by Okumura:

FrequencyDistanceMobile station heightBase station height

150 – 1500 MHz1 – 20 km1 – 10 m30 – 200 m

2012-03-21 Ove Edfors - ETIN15 16

The Okumura-Hata modelHow to calculate prop. loss

( )|logO H kmL A B d C− = + +

Small/medium-size cities

Metropolitanareas

Suburbanenvironments

Rural areas

( )( )( )( )

0|

0|

1.1log 0.7

1.56log 0.8

MHz m

MHz

f h

f

− −

−

( )ma h =

8.29 log 1.54 hm 2−1.1

3.2 log 11.75hm 2−4.97

for f 0200 MHz

for f 0400 MHz0

0

C =

−2 log f 0∣MHz /28 2−5.4

−4.78 log f 0∣MHz 218.33 log f 0∣MHz −40.94

( ) ( ) ( )( )

0|69.55 26.16log 13.82log

44.9 6.55log

MHz b m

b

A f h a h

B h

= + − −

= −

hb and hm

in meter

2012-03-21 Ove Edfors - ETIN15 17

COST 231-Walfish-Ikegami modelBackground

The Okumura-Hata model is not suitable for micro cells or smallmacro cells, due to its restrictions on distance (d > 1 km).

The COST 231-Walfish-Ikegami model covers much smallerdistances and is better suited for calculations on small cells.

FrequencyDistanceMobile station heightBase station height

800 – 2000 MHz0.02 – 5 km1 – 3 m4 – 50 m

2012-03-21 Ove Edfors - ETIN15 18

COST 231-Walfish-Ikegami modelHow to calculate prop. loss

0 msd rtsL L L L= + +

BS

MS

d

Freespace

Roof-topto street

Building multiscreen

Details about calculations can be found in Appendix 7.B.

2012-03-21 Ove Edfors - ETIN15 19

WIDEBANDMODELS

2012-03-21 Ove Edfors - ETIN15 20

Wideband modelsReview of properties

Let’s assume the tapped delay-line model

The power-delay profile tells us how much energy the channel hasat a certain delay τ (essentially the rms values of the αi(t)’s).

The Doppler spectrum tells us how fast the channel changes in time(essentially how fast the αi(t)’s and θi(t)’s change). There can be oneDoppler spectrum for each delay.

ht ,=∑i=1

N

i t exp j i t − i

2012-03-21 Ove Edfors - ETIN15 21

Wideband modelsCOST 207 model for GSM

The COST 207 model specifies:

FOUR power-delay profiles for differentenvironments.

FOUR Doppler spectra used for differentdelays.

IT DOES NOT SPECIFY PROAGATION LOSSES FOR THEDIFFERENT ENVIRONMENTS!

2012-03-21 Ove Edfors - ETIN15 22


[ ]sτ µ

[ ]P dB0

10−20−30−

0 1 [ ]sτ µ

[ ]P dB0

10−20−30−

0 1 2 3 4 5 6 7

[ ]sτ µ

[ ]P dB0

10−20−30−

0 5 10 [ ]sτ µ

[ ]P dB0

10−20−30−

100 20

Four specified power-delay profiles

RURAL AREA TYPICAL URBAN

BAD URBAN HILLY TERRAIN

2012-03-21 Ove Edfors - ETIN15 23


Four specified Doppler spectra

maxν− maxν+0

( ),s iP ν τ

maxν− maxν+0

maxν− maxν+0 maxν− maxν+0

CLASS GAUS1

GAUS2 RICEShortestpath in

rural areas

( ),s iP ν τ

( ),s iP ν τ( ),s iP ν τ

i≤0.5 s 0.5 si≤2 s

i2 s

2012-03-21 Ove Edfors - ETIN15 24

GAUS2GAUS1CLASS


[ ]sτ µ

[ ]P dB0

10−20−30−

0 1 [ ]sτ µ

[ ]P dB0

10−20−30−

0 1 2 3 4 5 6 7

[ ]sτ µ

[ ]P dB0

10−20−30−

0 5 10 [ ]sτ µ

[ ]P dB0

10−20−30−

100 20

RURAL AREA TYPICAL URBAN

BAD URBAN HILLY TERRAIN

Doppler spectra:

First tapRICEhere

2012-03-21 Ove Edfors - ETIN15 25


There are also suggested tapped delay-line implementations, withsix Rayleigh-fading taps per channel. See Appendix 7.C (on-line).

QUICK QUIZ: The system bit-rate of GSM is 271 kbit/s.

How long is one bit in time?

How long are the different COST 207 channels,measured in bit-times?

2012-03-21 Ove Edfors - ETIN15 26

Wideband modelsITU-R model for 3GThe ITU-R model specifies:

SIX different tapped delay-line channels forthree different scenarios (indoor, pedestrian, vehicular).

TWO channels per scenario (one short and onelong delay spread).

TWO different Doppler spectra (uniform & classical),depending on scenario.

THREE different models for propagation loss (onefor each scenario).

The standard deviation of the log-normal shadowfading is specified for each scenario.

The autocorrelation of the log-normal shadowfading is specified for the vehicular scenario.

2012-03-21 Ove Edfors - ETIN15 27

Wideband modelsITU-R model for 3G

ns

2012-03-21 Ove Edfors - ETIN15 28

ANTENNAS

2012-03-21 Ove Edfors - ETIN15 29

AntennasEfficiency

The antenna efficiency measures “how efficiently” an antennaconverts the input power into radiation. This translates directlyinto power consumption and battery life.

Antenna efficiency of mobiles has decreased mainly due tocosmetic restrictions.

What cosmetic restrictions?

2012-03-21 Ove Edfors - ETIN15 30

AntennasBandwidth

We can say that the bandwidth of an antenna is the width ofthe frequency range over which it fulfills some specification.

Most cellular systems have a bandwidth requirement in therange of 10% of the carrier frequency.

Example: 900 MHz GSM needs an antenna that can transmit/receive well in a total bandwidth of about 100 MHz.

What happens when we have dual- (900/1800) or triple-band(900/1800/1900) GSM phones ... or phones with 3G andBluetooth (2.4 GHz) as well?

It is difficult to make small and efficient broadband antennas!

2012-03-21 Ove Edfors - ETIN15 31

AntennasMobile station antennas

Monopole Helix Patch

2012-03-21 Ove Edfors - ETIN15 32

AntennasMobile station antennas

The efficiency depends on many parameters, but a very importantone is its environment. Below you can see differences in antennaefficiency for 42 test persons holding the mobile.

Up to around 10 dB difference,

depending on person.

2012-03-21 Ove Edfors - ETIN15 33

AntennasBase station antennas

Narrowmast

5 cmdiam.mast

10 cmdiam.mast

Base station antenna pattern affected by the mast (30 cm from antenna).

2012-03-21 Ove Edfors - ETIN15 34

AntennasBase station antennas

Base station antenna pattern affected by a concrete foundation.

2012-03-21 Ove Edfors - ETIN15 35

AntennasThe dipole antenna

[Figure from Ericsson Radio School documentation]

2012-03-21 Ove Edfors - ETIN15 36

AntennasThe parabolic antenna

Opening area:

Effective area:

Antenna gain:

3dB beamwidth: ≈200

Ga

[degrees]25o

[Figure from Ericsson Radio School documentation]

Aeff≈0.55 A

A= d 2

4

Ga=4

2Aeff≈0.55

2 d 2

2

2012-03-21 Ove Edfors - ETIN15 37

Summary

• Narrowband models: Okumura´s measurements, Okumura-Hata, COST 231-Ikegami-Walfish. Mainly models for propagation loss. Fading has to be added.

• Wideband models: COST 207 for GSM & ITU-R for 3G. Mainly specification of power-delay profile and doppler spectrum (IRT-R also gives e.g. path loss).

• Antennas: Efficiency has decreased for mobile antennas. Antenna environment changes their properties. Some specific properties for dipole and parabolic antennas.

Documents

RADIO SYSTEMS – ETIN15 Lecture no: 4 · 2012-03-21 Ove Edfors - ETIN15 17 COST 231-Walfish-Ikegami model Background The Okumura-Hata model is not suitable for micro cells or small