Mobile Communications

Part IV- Propagation Characteristics

 Contents
• Radiation from Antenna
• Propagation Model (Channel Models)
– Free Space Loss
• Plan Earth Propagation Model
• Practical Models
• Summary
 Wireless Communication System
Channel code word
Message Signal
Modulated
Source Channel Mod- Transmitted
Source Signal
Encoder Encoder ulator

Wireless
Channel

Source Channel Demod-

User Received
Decoder Decoder ulator
Signal
Estimate of
Estimate of
Message signal
channel code word
 Basic Questions

Tx:
• What will happen if the transmitter
- changes transmit power ?
- changes frequency ?
- operates at higher speed ?

Channel:
What will happen if we conduct this experiment in different types of
environments?

Desert Metro Street Indoor

Rx:

What will happen if the receiver moves?

 Antenna - Ideal
• Isotropics antenna: In free space radiates power equally in all
direction. Not realizable physically

H
Radiation pattern
E EM fields around a d
d transmitting antenna
Pt
, a polar coordinate

2
𝐴𝑒−𝐼 =
4
• d- distance directly away from the antenna.
Area •  is the azimuth, or angle in the horizontal
plane.
d •  is the zenith, or angle above the horizon.
 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
 Antenna – Real - Simple Dipoles
• Not isotropic radiators, e.g., dipoles with lengths /4 on car roofs or /2 as
Hertzian dipole
Largest dimension of the antenna La = /2

/4 Feedline
/2

2 𝐺
𝐴𝑒−𝑅 =
4

y y z

simple
x z x dipole

side view (xy-plane) side view (yz-plane) top view (xz-plane)

• Example: Radiation pattern of a simple Hertzian dipole shape of antenna is

proportional to the wavelength
 Antenna – Real - Sdirected and Sectorized
• Used for microwave or base stations
for mobile phones (e.g., radio
coverage of a valley)

y y z
Directed
x z x

side view (xy-plane) side view (yz-plane)top view (xz-plane)

z
z

x Sectorized
x

top view, 3 sector top view, 6 sector

 Antenna - Ideal - contd.
• The power density of an ideal loss-less antenna at a distance d away from the
transmitting antenna is:
𝑃𝑡
𝑃𝑎 = W/m2
4𝜋𝑑 2
Pt Gt
For a directive antenna the
Pa  W/m2 Note: the area is for a
power density
4d 2 sphere.

• Gt is the transmitting antenna gain in dB

• The product PtGt : Equivalent Isotropic Radiation Power (EIRP)

ERIP is the power fed to a perfect isotropic antenna to

get the same output power of the practical antenna in hand.
 Antenna - Ideal - contd.
• The strength of the signal is often defined in terms of its Electric Field Intensity E (V/m) or Magnetic
Intensity H (A/m), because it is easier to measure it.

Note,
• The direction of E field defines the polarization of the wave.
E = Rm.H, Pa = E2/Rm = H2Rm and Pa = E.H • At a sufficiently large distance from the transmitting antenna, E and
H field
where Rm is the impedance of the medium. For free strength
space areOhms.
Rm = 377 proportional to each other. "Sufficiently large"
.
means more than 4λ (λ is the wavelength of the transmitted signal
Pa in meters). Distances from λ/2π to 4λ give good results, though
under certain circumstances the values may not be too precise.

Pt Rm Pt Rm
E  2
and E  V/m
4d 2 4d 2
 Antenna - Ideal - contd.
Example: A GSM base station about 100 m away from the
wooden pole on the left.

Assuming that Pt = 20 W, that the cable has a loss of 3 dB, and

Gt = 18 dBi, we can estimate the field strength at 100 m
distance. To take into account the loss in the cable, deduce the
3 dB loss from the 18 dBi gain and use a modified antenna
gain of G = 15 dBi.
• E = 1.4 V/m, H = 3.7 mA/m and the Pa = 5.0 mW/m2. We
also find the effective isotropic radiated power PEIRP =
633 W.
 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

Ae = Pr / Pa => Ae-I-Rx = 2/4

• For a real antenna Ae-R-Rx = Gr2/4

which is the equivalent power absorbing area of the antenna.
Gr is the receiving antenna gain and  = c/f

𝑃𝑟
And 𝐸= 𝑅𝑚
𝐴𝑒−𝑅𝑥
Signal Propagation
(Channel Models)
 Basic Digital Communication System Model
x(t)
Input Pulse TX
channel
data generator filter
Hc(f)
HT(f)
+
Channel noise

x(t )  anx(t  nTb ) n(t)

+

Tb: Symbol period y(t) Receiver

A/D filter
Output data HR(f)

y (t )  x(t ) * h(t )  n 0(t )

y (t )  hT (t ) * h(t ) * hR (t )  n 0(t )

y (t )   Ak hc (t  t d  nTb )  n0 (t )
n
 Fourier Representation of Periodic
Signals
 
1
g (t )  c   an sin( 2nft)   bn cos( 2nft)
2 n 1 n 1

Amp
Amp
1 1

0 0
t t
Ideal periodic signal Real composition
(based on harmonics)
 Signals II
• Different representations of signals
– amplitude (amplitude domain)
– frequency spectrum (frequency domain)
– phase state diagram (amplitude M and phase  in polar coordinates)

A [V] A Q = M sin 
[V]
t[s] 
I= M cos 

 f [Hz]

• 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!)
 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.
 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
 Channel Models - Linear
• 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.
x(t ) y (t )

Channel
 Channel Models – Linear Time Varying
Finite delay
Constant

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

h(t) H(f)

No Noise
• Therefore for a linear channel is: Y ( f )  F [ y (t )]
H ( f )  k , arg H ( f )  2ftd  ke jtd X ( f )
 H( f )X ( f )
Amplitude Phase where H ( f )  ke jtd
distortion distortion
 The phase delay td ( f )   arg H ( f ) / 2f
Describes the phase delayed experienced by each frequency component
 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.

x(t) = (t) Linear channel y(t) = k x(t-td)

x(n)= (n) h(t) H(f) y(n)= k x(n-N)
h(n)
Channel is usually modeled as Tap-Delay-Line

x(n)
D D D

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

+
+ +
+ +

y(n)
 Channel Models – Multipath Link

 The received signal is: 𝑦 𝑡 = ෍ 𝑎𝑖 𝑡 𝑥(−𝜏𝑖 𝑡 )

𝑖

 Time invariant multi-path channel impulse response

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

h 𝑡, 𝜏 = σ𝑁−1
𝑖 𝑎𝑖 𝑡 𝑒 −𝑗𝜃𝑖 𝛿(−𝜏𝑖 𝑡 )

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)
 Channel Models – Multipath Link
 Channel transfer function

H ( f )   h(t )e  j 2 ft dt

N 1
  ai e  ji e i 2 fi
i 0

• Multipath Time
– Mostly used to denote the severity of multipath conditions.
– Defined as the time delay between the 1st and the last received impulses.

TMP   N 1  0
 Coherence bandwidth - on average the distance between two notches
Bc ~ 1/TMP
Channel Models – Multipath Link
Time domain view Freq. domain view

x(t ) X(f )

Range of freq over

which response is flat
 RMS delay spread Bc
1
Bco  For 0.9 correlation High correlation of amplitude
50.  between two different freq.
1 components
Bco  For 0.5 correlation
5. 
 Power Delay Profile
-90

RMS Delay Spread () = 46.4 ns

Received Signal Level (dBm)

-90

Mean Excess delay () = 45 ns

-95

Maximum Excess delay < 10 dB = 110 ns

-100
Noise threshold

-105

0 50 100 150 200 250 300 350 400 450

Excess delay (ns)

 Power Delay Spread - Example
Pr() 4.38 µs
1.37 µs
0 dB

-10 dB

-20 dB

-30 dB 

0 1 2 5 (µs)

  (1)(5) (0.1)(1) (0.1)(2)(0.01)(0)  4.38s

_

[0.01 0.1 0.11]

(1)(5) 2  (0.1)(1) 2  (0.1)(2) 2  (0.01)(0) 2

_
 
2
 21.07s 2 
[0.01 0.1 0.11]

  21.07 (4.38) 2 1.37s

 Power Delay Spread - Example
Pr() 4.38 µs _ _
1.37 µs   4.38s   21.07s 2 
2
0 dB

-10 dB
  1.37s
-20 dB

1
(50%  coherence ) Bc   146kHz
-30 dB 

0 1 2 5 (µs) 5. 

Indoor 10  50 n sec
Suburbs 2  101  2  sec
Urban 1  3  sec
Hilly 3-10  sec
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

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  BS
unlimited time response
• Channel has multipath
• Tx filter – 1 0 2
Buildi
• There are two major ways to mitigate
ng
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
MU
 ISI

Symbol time Pr() 4.38 µs

0 dB

-10 dB

-20 dB

 -30 dB 
0 1 2 5 (µs) 0 1 2 5 (µs)
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)
 Propagation Path Loss
• The propagation path loss 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) ]
 Propagation Path Loss – Free Space
• Power received at the receiving antenna at far field
only

  
2

Pr  Pt Gt Gr  
 4d 
• Isotropic antenna has unity gain (G = 1) for both transmitter and receiver.

• If the Pr = Pr-min (i.e., the minimum signal required for the system), then the
maximum range is:

0.5
𝑃𝑡 𝐺𝑡 𝐺𝑟 2
𝑑𝑚𝑎𝑥 =
4𝜋 2 𝑃𝑟−𝑚𝑖𝑛
 Propagation Path Loss – Free Space
Thus the free space propagation path loss is defined as:

Pr  Gt Gr 2 
L f  10 Log10  10 Log10  2
Pt  ( 4d ) 

The difference between two received signal powers in

free space is:
 Pr1   d1 
P  10 log 10    20 log 10   dB
 Pr 2   d2 

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

 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: 4 – 6
– Urban: 2.7-3.5
– Indoor (LOS): 1.6-1.8

 For non-line of sight with no shadowing, received power at

any distance d can be expressed as:
𝑃𝑟 𝑑 =10 log𝑃𝑟 𝑑𝑟𝑒𝑓 + 10𝑣 log 𝑑
ൗ𝑑𝑟𝑒𝑓
1-10 m indoor < dref < 10-100 m outdoor
 Propagation - Non-Line-of-Sight

• Log-normal Shadowing

 d 
Pr (d )  Pr (1 m)  10 log 10 (d ref )  20v log 10    X
d 
 ref 

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

Received Power for Different Value of Loss
Parameter v
-70
-80

-90
v = 2, Free space
-100

-110
v = 3 Rural areas City and urban
-120 areas

v = 4,
-130
-135
0 10 20 30 40 50
Distance (km)
 Propagation Model- Free Space
In terms of frequency f and the free space velocity of
electromagnetic wave c = 3 x 108 m/s it is:
c/ f 
L f  20 log 10   dB
 4d 

Expressing frequency in MHz and distance d in km:

L f  20 log 10 (c / 4)  20 log 10 ( f )  20 log 10 (d )

 20 log 10 (0.3 / 4)  20 log 10 ( f )  20 log 10 (d ) dB

L f  32.44  20 log 10 ( f )  20 log 10 (d ) dB

 Propagation Model- Free Space (non-ideal, path
loss)

BS MU

• Non-isotropic antenna gain  unity, and there are additional

losses Lad , thus the power received is:

Pt  1 2
Pr  Gt Gr  d > 0 and L 0
(4d ) Lad
2

Thus for Non-isotropic antenna the path loss is:

L f ni  10 log 10 (Gt )  10 log 10 (Gr )  20 log 10 (c / 4)
 20 log 10 ( f )  20 log 10 (d )  10 log 10 ( Lad ) dB
Note: Interference margin can also be added
 Propagation Model- Free Space (non-ideal, path
loss)
• Considering loss S due to shadowing

Pr  Pt Gt Gr L f S

S is a normally (Gaussian) distributed random variable (in dB) with

standard deviation
 Propagation Model - Mechanisms

• Reflection
• Diffraction
• Scattering
 Propagation Model- Path loss +

Path loss, shadowing, and Multipath [Goldsmith2005]

 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)

d
dd
Direct path (line of sight)
hb
hm

Ground reflected
dr path

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

 Channel Model- Plan Earth Path Loss -
contd.

Using the binomial expansion 

  hb  hm  
2

Note d >> hb or hm. d d  d 1  0.5  

  d   

Similarly
  b
h  h
2
m  

d r  d 1  0.5  
  d  

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

2 2hb hm 4hb hm
The phase difference    
 d d
 Channel Model- Plan Earth Path Loss–
contd.

Total received power

2
   j
2
Pr  Pt Gt Gr    1  e
 4d 

Where  is the reflection coefficient.

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

1  e  j  1  cos   j sin 

 j 2
Hence 1  e  (1  cos ) 2  sin 2   2(1  cos )
 4 sin 2 ( / 2)
 Channel Model- Plan Earth Path Loss–
contd.

   2  2hb hm 
2

Therefore: Pr  Pt Gt Gr    4 sin  
 4d   d 

 2hb hm 
Assuming that d >> hm or hb, then    1
 d 
sin x = x for small x

2
 hb hm 
Thus Pr  Pt Gt Gr  2 
 d 
which is 4th power law
 Channel Model- Plan Earth Path Loss– contd.

Propagation path  Pr    b m 
h h
2

LPE  10 log    10 log Gt Gr  2  

loss (mean loss)  Pt    d  
Compared with the free space = Pr = 1/ d2

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

path attenuation is

LPE  10 log 10 Gt  10 log 10 Gr  20 log 10 hb

 20 log 10 hm  40 log 10 d dΒ
• LPE increases by 40 dB each time d increases by 10
 Channel Model- Plan Earth Path Loss– contd.

• Including impedance mismatch, misalignment of antennas,

pointing and polarization, and absorption The power ration is:

  
 1   a  a
2 2
Pr
 Gt (t , t )Gr (r , r )  1  t e d
2 2 *

 4d 
r t r
Pt
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.
 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

 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 Outdoor

• Saleh model
• SIRCIM model Indoor
Y. Okumura, et al, Rev. Elec. Commun. Lab., 16( 9), 1968.
M. Hata, IEEE Trans. Veh. Technol., 29, pp. 317-325, 1980.
 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.
 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.
 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
 Hata Model - Limits
• Frequency range: 150 - 1500 MHz
• Distance: 1 – 20 km
• BS antenna height: 30- 200 m
• MU antenna height: 1 – 10 m
 Hata Model – Standard Formula for Average
Path Loss for Urban Areas
L pl u  69.55  26.16 log 10  f   44.9  6.55 log 10 hb  log 10 d
 13.82 log 10 hb  ahmu  (dB)
Correction Factors are:
• Large cities

   
a hmu  8.3 log 10 1.5hmu 1.1  f
 2
 200MHz  dB

ahmu   3.2log 10 11.75hmu   4.97 f  400MHz  dB

2

• Average and small cities

   
ahmu   1.1log 10  f   0.7 hmu  1.56 log 10  f   0.8 dB
 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
Hata Model – Average Path Loss for Suburban and
Open Areas

• Suburban Areas
2
  f 
Lpl  su  Lpl  u  2Log10    5.4
  28 

 Open Areas

Lplo  Lplu  4.78(Log10 f ) 18.33 Log f  40.94

2
Hata Model - Average Path Loss

S. Loyka, 2003,
Introduction to Mobile Communications
 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 = 100 m
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.
 Improved Model
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
• 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
 Losses
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.
 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

• Next lecture: Multi-path Propagation- Fading