mobile communications -...

61
Z. Ghassemlooy Mobile Communications Part IV- Propagation Characteristics Professor Z Ghassemlooy Faculty of Engineering and Environment University of Northumbria U.K. http://soe.unn.ac.uk/ocr

Upload: vuque

Post on 30-Mar-2018

217 views

Category:

Documents


1 download

TRANSCRIPT

Z. Ghassemlooy

Mobile Communications

Part IV- Propagation CharacteristicsPart IV- Propagation Characteristics

Professor Z Ghassemlooy

Faculty of Engineering and

Environment

University of Northumbria

U.K.

http://soe.unn.ac.uk/ocr

Professor Z Ghassemlooy

Faculty of Engineering and

Environment

University of Northumbria

U.K.

http://soe.unn.ac.uk/ocr

Z. Ghassemlooy

Contents

Radiation from Antenna

Propagation Model (Channel Models)

– Free Space Loss

Plan Earth Propagation Model

Practical Models

Summary

Z. Ghassemlooy

Wireless Communication System

UserSource

Decoder

Channel

Decoder

Demod-

ulator

Estimate of

Message signalEstimate of

channel code word

Received

Signal

Channel code word

SourceSource

Encoder

Channel

Encoder

Mod-

ulator

Message SignalModulated

Transmitted

Signal

Wireless

Channel

Basic Questions

Tx:

What will happen if the

transmitter

- changes transmit power ?

- changes frequency ?

- operates at higher speed ?

Z. Ghassemlooy

Desert Metro Street Indoor

Channel:

What will happen if we conduct this experiment in different types of

environments?

Rx:

What will happen if the receiver moves?

Z. Ghassemlooy

dd

Antenna - Ideal

Isotropics antenna: In free space radiates power equally in

all direction. Not realizable physically

dE

H

Pt

Distance d

Area• d- distance directly away from the

antenna.

• is the azimuth, or angle in the horizontal

plane.

• is the zenith, or angle above the horizon.

EM fields around a

transmitting antenna

, a polar coordinate

Z. Ghassemlooy

Antenna - Real

Not isotropic radiators, but always have

directive effects (vertically and/or

horizontally)

A well defined radiation pattern measured

around an antenna

Patterns are visualised by drawing the set of

constant-intensity surfaces

Z. Ghassemlooy

Antenna – Real - Simple Dipoles

Not isotropic radiators, e.g., dipoles with lengths /4

on car roofs or /2 as Hertzian dipole

Example: Radiation pattern of a simple Hertzian dipole

shape of antenna is proportional to the wavelength

side view (xy-plane)

x

y

side view (yz-plane)

z

y

top view (xz-plane)

x

z

simple

dipole

/4 /2

Largest dimension of the antenna La = /2

Z. Ghassemlooy

Antenna – Real - Sdirected and Sectorized

Used for microwave or base stations for mobile

phones (e.g., radio coverage of a valley)

side view (xy-plane)

x

y

side view (yz-plane)

z

y

top view (xz-plane)

x

z

Directed

top view, 3 sector

x

z

top view, 6 sector

x

z

Sectorized

Z. Ghassemlooy

Antenna - Ideal - contd.

The power density of an ideal loss-less antenna at a distance d away from the transmitting antenna is:

24 d

GPP tt

a

• Gt is the transmitting antenna gain

• The product PtGt : Equivalent Isotropic Radiation Power

(EIRP)

which is the power fed to a perfect isotropic antenna to

get the same output power of the practical antenna in hand.

Note: the area is for a

sphere.W/m2

Z. Ghassemlooy

Antenna - Ideal - contd.

The strength of the signal is often defined in terms

of its Electric Field Intensity E, because it is easier

to measure.

Pa = E2/Rm where Rm is the impedance of the medium. For

free space Rm = 377 Ohms..

22

2

44 d

RPEand

d

RPE mtmt

V/m

Z. Ghassemlooy

Antenna - Ideal - contd.

The receiving antenna is characterized by its

effective aperture Ae, which describes how well an

antenna can pick up power from an incoming

electromagnetic wave

The effective aperture Ae is related to the gain Gr

as follows:Ae = Pr / Pa => Ae = Gr

2/4

which is the equivalent power absorbing area of

the antenna.

Gr is the receiving antenna gain and = c/f

Z. Ghassemlooy

Signal Propagation

(Channel Models)

Basic Digital Communication System

Model

Z. Ghassemlooy

Input

data

Pulse

generator

TX

filterchannel

x(t)

Receiver

filterA/D

+Channel noise

n(t)

Output data

+

HT(f)

HR(f)

y(t)

Hc(f)

n

bdck tnnTtthAty

tnthththty

tnthtxty

RT

)()()(

)()(*)(*)()(

)()(*)()(

0

0

0

)()( bn nTtxatx

Tb: Symbol period

Fourier Representation of Periodic

Signals

)2cos()2sin(2

1)(

11

nftbnftactgn

n

n

n

1

0

1

0t t

ideal periodic signal real composition

(based on harmonics)

Different representations of signals – amplitude (amplitude domain)

– frequency spectrum (frequency domain)

– phase state diagram (amplitude M and phase in polar coordinates)

Composite signals mapped into frequency domain using Fourier transformation

Digital signals need– infinite frequencies for perfect representation

– modulation with a carrier frequency for transmission (->analog signal!)

Signals II

f [Hz]

A

[V]

I= M cos

Q = M sin

A [V]

t[s]

Z. Ghassemlooy

Channel Models

High degree of variability (in time, space etc.)

Large signal attenuation

Non-stationary, unpredictable and random – Unlike wired channels it is highly dependent on the environment,

time space etc.

Modelling is done in a statistical fashion

The location of the base station antenna has a significant effect on channel modelling

Models are only an approximation of the actual signal propagation in the medium.

Are used for:– performance analysis

– simulations of mobile systems

– measurements in a controlled environment, to guarantee repeatability and to avoid the expensive measurements in the field.

Z. Ghassemlooy

Channel Models - Classifications

System Model - Deterministic

Propagation Model- Deterministic– Predicts the received signal strength at a distance from the

transmitter

– Derived using a combination of theoretical and empirical method.

Stochastic Model - Rayleigh channel

Semi-empirical (Practical +Theoretical) Models

Z. Ghassemlooy

Channel Models - Linear

Therefore for a linear channel is:

Linear channel

h(t) H(f)x(t) y(t) = k x(t-td)

ConstantFinite delay

d

d

tj

tj

kefHwhere

fXfH

fXke

tyFfY

)(

)()(

)(

)]([)(

dftfHkfH 2)(arg,)(

Amplitude

distortion

Phase

distortion

The phase delayDescribes the phase delayed experienced by each frequency component

ffHftd 2/)(arg)(

No Noise

This can be modelled as an

– ideal low pass filter

– non-ideal low pass filter

• This smears the transmitted signal x(t) in time causing the effect

of a symbol to spread to adjacent symbols when a sequence of

symbols are transmitted.

• The resulting interference, intersymbol interference (ISI),

degrades the error performance of the communication system.

Z. Ghassemlooy

)(tx

Channel

)(ty

Power Delay Profile

Z. Ghassemlooy

Receiv

ed S

ignal Level (d

Bm

)

-105

-100

-95

-90

-90

0 50 100 150 200 250 300 350 400 450

Excess delay (ns)

RMS Delay Spread () = 46.4 ns

Mean Excess delay () = 45 ns

Maximum Excess delay < 10 dB = 110 ns

Noise threshold

Power Delay Spread - Example

Z. Ghassemlooy

-30 dB

-20 dB

-10 dB

0 dB

0 1 2 5

Pr()

(µs)

s 38.4

]11.01.001.0[

)0)(01.0()2)(1.0()1)(1.0()5)(1(_

2

2222_2 07.21

]11.01.001.0[

)0)(01.0()2)(1.0()1)(1.0()5)(1(s

s

37.1)38.4(07.21 2

1.37 µs

4.38 µs

Z. Ghassemlooy

Channel Models – Multipath Link

The mathematical model of the multipath can be presented

using the method of the impulse response used for

studying linear systems.

Linear channel

h(t) H(f)

h(n)

x(t) = (t)

x(n)= (n)

y(t) = k x(t-td)

y(n)= k x(n-N)

D D D

h(0) h(1) h(2) h(N-1) h(N)

y(n)

x(n)

++

++

+

Channel is usually modeled as Tap-Delay-Line

Z. Ghassemlooy

Channel Models – Multipath Link

Time variable multi-path channel impulse response

))(())],()(2(exp[),(),(1

0

tttfjtath iii

N

i

cib

Time invariant multi-path channel impulse response

– Each impulse response is the same or has the same statistics, then

1

0

)()(N

i

i

j

iieah

Where aie(.) = complex amplitude (i.e., magnitude and phase) of the

generic received pulse.

= propagation delay generic ith impulse

N = number signal arriving from N path

(.) = impulse signal (delta function)

Where a(t-) = attenuated signal

Z. Ghassemlooy

Channel Models – Multipath Link

Multipath Time

– Mostly used to denote the severity of multipath conditions.

– Defined as the time delay between the 1st and the last received

impulses.01 NMPT

1

0

2

2)()(

N

i

fij

i

ftj

ii eea

dtethfH

Channel transfer function

Coherence bandwidth - on average the distance between two notches

Bc ~ 1/TMP

Z. Ghassemlooy

)(tx

Time domain view

High correlation of amplitude

between two different freq.

components

Range of freq over

which response is flat

Bc RMS delay spread

)( fX

Freq. domain view

.50

1coB For 0.9 correlation

.5

1coB For 0.5 correlation

Power Delay Spread - Example

Z. Ghassemlooy

-30 dB

-20 dB

-10 dB

0 dB

0 1 2 5

Pr()

(µs)

s 38.4_

2

_2 07.21 s s

37.1

1.37 µs

4.38 µs

kHzBcoherence c 146.5

1)%50(

Signal bandwidth for Analog Cellular = 30 KHz

Signal bandwidth for GSM = 200 KHz

Multipath Delay

With a 52 meter separation, in a LOS environment, a large 753.5 ns

excess delay

NLOS excess delay over 423 meters extended to 1388.4 ns.

Office building: RMS delay spread = 10-60 ns

Z. Ghassemlooy

Cell size (km) Max Delay Spread

Pico cell 0.1 300 nn

Micro cell 5 15 us

Macro cell 20 40 us

Inter-symbol interference (ISI)

Channel is band limited innature

– Limited frequency response

unlimited time response

Channel has multipath

Tx filter –

There are two major ways to

mitigate the detrimental effect of ISI.

– The first method is to design

bandlimited transmission pulses

(i.e., Nyquist pulses) which

minimize the effect of ISI.

– Equalization

Z. Ghassemlooy

BS

MU

Build

ing

01

2

ISI

Z. Ghassemlooy

-30 dB

-20 dB

-10 dB

0 dB

0 1 2 5

Pr()

(µs)

4.38 µs

0 1 2 5 (µs)

Symbol time

4.38

Symbol time > 10* --- No equalization required

Symbol time < 10* --- Equalization will be required to deal with ISI

In the above example, symbol time should be more than 14µs to avoid ISI.

This means that link speed must be less than 70Kbps (approx)

Z. Ghassemlooy

Propagation Path Loss

The propagation path loss is

LPE = LaLlf Lsf

where

La is average path loss (attenuation): (1-10

km),

Llf - long term fading (shadowing): 100 m

ignoring variations over few wavelengths,

Lsf - short term fading (multipath): over fraction

of wavelength to few wavelength.

Metrics (dBm, mW)

[P(dBm) = 10 * log[ P(mW) ]

Z. Ghassemlooy

Propagation Path Loss – Free Space

Power received at the

receiving antenna

at far field only

2

4

dGGPP rttr

• Isotropic antenna has unity gain (G = 1) for both

transmitter and receiver.

Thus the free space propagation path loss is defined

as:

2

2

1010)4(

10Log10d

GGLog

P

PL rt

t

rf

Z. Ghassemlooy

Propagation - Free Space–contd.

The difference between two received signal powers in

free space is:

dB2

1log20log10 10

2

110

d

d

P

PP

r

r dB2

1log20log10 10

2

110

d

d

P

PP

r

r

If d2 = 2d1, the P = -6 dB i.e 6 dB/octave or 20 dB/decade

Z. Ghassemlooy

Propagation - Non-Line-of-Sight

• Generally the received power can be expressed as:

Pr d-v

• For line of sight v = 2, and the received power

Pr d-2

• Average v over all NLOS could be 4 - 6.

• For non-line of sight with no shadowing, received

power at any distance d can be expressed as:

ref

rrd

dvdmPdP 10ref10 log20)(log20)1()(

100 m< dref < 1000 m

Z. Ghassemlooy

Propagation - Non-Line-of-Sight

Log-normal Shadowing

Where Xσ: N(0,σ) Gaussian distributed random variable

X

d

dvdmPdP

ref

rr 10ref10 log20)(log10)1()(

Z. Ghassemlooy

Received Power for Different Value of Loss

Parameter v

Distance (km)0 10 20 30 40 50

-70

-80

-90

-100

-110

-120

-130

-135

v = 2, Free space

v = 3 Rural areas

v = 4,

City and urban

areas

Z. Ghassemlooy

Propagation Model- Free Space

dB)(log20)(log2044.32 1010 dfL f dB)(log20)(log2044.32 1010 dfL f

Expressing frequency in MHz and distance d in km:

dB)(log20)(log20)4/3.0(log20

)(log20)(log20)4/(log20

101010

101010

df

dfcL f

dB)(log20)(log20)4/3.0(log20

)(log20)(log20)4/(log20

101010

101010

df

dfcL f

In terms of frequency f and the free space velocity of electromagnetic wave c = 3 x 108 m/s it is:

dB4

/log20 10

d

fcL f

Z. Ghassemlooy

Propagation Model- Free Space (non-ideal,

path loss)

• Non-isotropic antenna gain unity, and there are

additional losses Lad , thus the power received is:

ad

trtr

Ld

PGGP

1

)4( 2

2

ad

trtr

Ld

PGGP

1

)4( 2

2

d > 0 and L 0

Thus for Non-isotropic antenna the path loss is:

dB)(log10)(log20)(log20

)4/(log20)(log10)(log10

101010

101010

ad

rtnif

Ldf

cGGL

dB)(log10)(log20)(log20

)4/(log20)(log10)(log10

101010

101010

ad

rtnif

Ldf

cGGL

BS MU

Note: Interference margin can also be added

Considering loss S due to shadowing

Z. Ghassemlooy

SLGGPP frttr

Propagation Model- Free Space (non-ideal,

path loss)

Z. Ghassemlooy

Propagation Model - Mechanisms

Reflection

Diffraction

Scattering

Source: P M Shankar

Z. Ghassemlooy

Channel Model- Plan Earth Path Loss - 2

Ray Reflection

In mobile radio systems the height of both antennas (Tx.

and Rx.) << d (distance of separation)

From the geometry dd = [d2 + (hb - hm )2]

hb

dd

dr

hm

Direct path (line of sight)

Ground reflected

path

d

Z. Ghassemlooy

Channel Model- Plan Earth Path Loss -

contd.

Using the binomial expansion

Note d >> hb or hm.

2

5.01d

hhdd mb

d

The path difference d = dr - dd = 2(hbhm )/d

Similarly

2

5.01d

hhdd mb

r

The phase differenced

hh

d

hh mbmb

422

Z. Ghassemlooy

Channel Model- Plan Earth Path

Loss– contd.

Total received power

22

14

j

rttr ed

GGPP

Where is the reflection coefficient.

For = -1 (low angle of incident) and .

)2/(sin4

)cos1(2sin)cos1(1

sincos11

2

222

j

j

eHence

je

Z. Ghassemlooy

Channel Model- Plan Earth Path

Loss– contd.

Therefore:

d

hh

dGGPP mb

rttr

2sin4

4

2

2

d

hh

dGGPP mb

rttr

2sin4

4

2

2

Assuming that d >> hm or hb, then 12

d

hh mb

sin x = x for small x

Thus2

2

d

hhGGPP mb

rttr

2

2

d

hhGGPP mb

rttr

which is 4th power law

Z. Ghassemlooy

Channel Model- Plan Earth Path Loss– contd.

In a more general form (no fading due to multipath),

path attenuation is

dΒlog40log20

log20log10log10

1010

101010

dh

hGGL

m

brtPE

dΒlog40log20

log20log10log10

1010

101010

dh

hGGL

m

brtPE

• LPE increases by 40 dB each time d increases by 10

Propagation path

loss (mean loss)

2

2log10log10

d

hhGG

P

PL mb

rt

t

rPE

Compared with the free space = Pr= 1/ d2

Z. Ghassemlooy

Channel Model- Plan Earth Path Loss– contd.

Including impedance mismatch, misalignment of antennas,

pointing and polarization, and absorption The power ration

is:

where

Gt(θt,φt) = gain of the transmit antenna in the direction (θt,φt) of receive

antenna.

Gr(θr,φr) = gain of the receive antenna in the direction (θr,φr) of transmit

antenna.

Γt and Γr = reflection coefficients of the transmit and receive antennas

at and ar = polarization vectors of the transmit and receive antennas

α is the absorption coefficient of the intervening medium.

d

rtrtrrrttt

t

r eaad

GGP

P

2*22

2

114

),(),(

Z. Ghassemlooy

LOS Channel Model - Problems

Simple theoretical models do not take into account

many practical factors:

– Rough terrain

– Buildings

– Refection

– Moving vehicle

– Shadowing

Thus resulting in bad accuracy

Solution: Semi- empirical Model

Z. Ghassemlooy

Sem-iempirical Model

Practical models are based on combination of measurement and theory. Correction factors are introduced to account for:– Terrain profile

– Antenna heights

– Building profiles

– Road shape/orientation

– Lakes, etc.

Okumura model

Hata model

Saleh model

SIRCIM model

Outdoor

Indoor

Y. Okumura, et al, Rev. Elec. Commun. Lab., 16( 9), 1968.

M. Hata, IEEE Trans. Veh. Technol., 29, pp. 317-325, 1980.

Z. Ghassemlooy

Okumura Model

Widely used empirical model (no analytical basis!) in macrocellular environment

Predicts average (median) path loss

“Accurate” within 10-14 dB in urban and suburban areas

Frequency range: 150-1500 MHz

Distance: > 1 km

BS antenna height: > 30 m.

MU antenna height: up to 3m.

Correction factors are then added.

Z. Ghassemlooy

Hata Model

Consolidate Okumura’s model in standard formulas for macrocells in urban, suburban and open rural areas.

Empirically derived correction factors are incorporated into the standard formula to account for:– Terrain profile

– Antenna heights

– Building profiles

– Street shape/orientation

– Lakes

– Etc.

Z. Ghassemlooy

Hata Model – contd.

The loss is given in terms of effective heights.

The starting point is an urban area. The BS antennae is mounted on tall buildings. The effective height is then estimated at 3 - 15 km from the base of the antennae.

P M Shankar

Z. Ghassemlooy

Hata Model - Limits

Frequency range: 150 - 1500 MHz

Distance: 1 – 20 km

BS antenna height: 30- 200 m

MU antenna height: 1 – 10 m

Z. Ghassemlooy

Hata Model – Standard Formula for

Average Path Loss for Urban Areas

Correction Factors are:

• Large cities

(dB)log82.13

loglog55.69.44log16.2655.69

10

101010

mub

bupl

hah

dhfL

dBMHz400 97.475.11log2.32

10 fhha mumu

dB 8.0log56.17.0log1.1 1010 fhfha mumu

dBMHz200 1.15.1log3.82

10 fhha mumu

• Average and small cities

Z. Ghassemlooy

Hata Model – Average Path Loss for Urban

Areas contd.

Carrier frequency

• 900 MHz,

BS antenna height

• 150 m,

MU antenna height

• 1.5m.

P M Shankar

Z. Ghassemlooy

Hata Model – Average Path Loss for

Suburban and Open Areas

Suburban Areas

Open Areas

94.40Log33.18)(Log78.4 2

10 ffLL uplopl 94.40Log33.18)(Log78.4 2

10 ffLL uplopl

4.528

Log2

2

10

fLL uplsupl 4.5

28Log2

2

10

fLL uplsupl

Z. Ghassemlooy

Hata Model - Average Path Loss

S. Loyka, 2003,

Introduction to Mobile Communications

Z. Ghassemlooy

Improved Model

Hata-Okumura model are not suitable for lower BS antenna heights (2 m), and hilly or moderate-to-heavy wooded terrain.

To correct for these limitations the following model is used [1]:

For a given close-in distance dref. the average path loss is:

Lpl = A + 10 v log10 (d / dref) + s for d > dref, (dB)

where

A = 20 log10(4 π dref / λ)

v is the path-loss exponent = (a – b hb + c / hb)

hb is the height of the BS: between 10 m and 80 m

dref = 100m and

a, b, c are constants dependent on the terrain category

s is representing the shadowing effect

[1] V. Erceg et. al, IEEE JSAC, 17 (7), July 1999, pp. 1205-1211.

Z. Ghassemlooy

Improved Model

Terrain A: The maximum path loss category is hilly terrain with moderate-to-heavy tree densities .

Terrain B: Intermediate path loss condition

Terrain B: The minimum path loss category which is mostly flat terrain with light tree densities

Terrains

Model Type A Type B Type C

parameter

a 4.6 4 3.6

b 0.0075 0.0065 0.005

c 12.6 17.1 20

The typical value of the standard deviation for s is between 8.2

And 10.6 dB, depending on the terrain/tree density type

Losses

Z. Ghassemlooy

Summary of penetration losses through

various common building materials at 28

GHz. Both of the horn antennas have 24.5

dBi gains with 10�

half power beamwidth [28].

Penetration losses for multiple indoor obstructions in an office

environment at 28 GHz. Weak signals are denoted by locations

where the SNR

was high enough to distinguish signal from noise but not enough for

the signal to be acquired, i.e. penetration losses were between 64

dB to 74 dB

relative to a 5 m free space test. No signal detected denotes an

outage, where penetration loss is greater than 74 dB relative to a 5

m free space test [28].

Reflection Coefficients

Comparison of reflection coefficients for various common building materials at 28 GHz.

Both of the horn antennas have 24.5 dBi gains with 10 half power beam-width.

Z. Ghassemlooy

Z. Ghassemlooy

Summary

• Attenuation is a result of reflection, scattering, diffraction and

reflection of the signal by natural and human-made structures

• The received power is inversely proportional to (distance)v,

where v is the loss parameter.

• Studied channel models and their limitations

Z. Ghassemlooy

Questions and Answers

Tell me what you think about this lecture

[email protected]

Next lecture: Multi-path Propagation- Fading