Introduction to Large Scale Path

Loss – Part 1
ArunKumar Jayaprakasam
Scope: Chapter 4 (till 4.8)
 Agenda
• Introduction
• Basic Properties of ElectroMagnetic Waves
– Antenna
• Basic Methods of Propagation
– Free Space
– Reflection
– Diffraction
– Scattering
• Two Ray Reflection
• Knife Edge Diffraction
• Scattering
 Propagation of RF waves can be divided into two main study areas

Large Scale Effects

• due to terrain and the density & dimensions of objects
• includes random destructive interference and diffraction
• characterized statistically by median path loss & log normal shadowing
• behavior is slowly time varying
 important for predicting coverage areas and service availability

Small Scale Effects

• considers fading associated with
(i) superposition of many E-field components caused by local reflections
(ii) changes in interference patterns at mobile receiver
• due to environment local to the receiver and mobility
• behavior varies much faster in time
• often characterized statistically as Rayleigh fading
 important in design of modulation format & general transceiver design
3
(1) Large Scale Propagation Models: predict mean signal strength
for TX-RX pair with arbitrary separation
• useful for estimating coverage area of a transmitter
• characterizes signal strength over large distances (102-103 m)
• predict local average signal strength that decreases with distance

signal power commonly computed by averaging measurements over

a measurement track of 5  - 40 

e.g. at 2.4 GHz,  =12.5cm 

• measurement tracks range from 0.625m - 5m
• if 0.625m used, for nth point  use segment between 0.625n & 0.625(n+1)
• take measurements at intervals of 0.1  to 
• averaged measurements = nth point
0.625m 1.25m 1.875m

4
 (2) Small Scale or Fading Models: characterize rapid fluctuations of
received signal over
• short distances (few ) or
• short durations (few seconds)

with mobility over short distances

• instantaneous signal strength fluctuates
• received signal = sum of many components from different directions
• phases are random  sum of contributions varies widely
• received signal may fluctuate 30-40 dB by moving a fraction of 

5
Scales of Mobile Signal Variation
BASIC TERMINOLOGIES OF
ELECTROMAGNETIC WAVES
 Plane Wave properties
• Electric and Magnetic Fields Perpendicular to
each other and to the direction of propagation.
• The two fields are in phase at any point in
time/space.
 Various Quantities

• Electric and Magnetic Permittivity

• Shown above ‘Z_0’ is free space impedance
 Power Density

• S denotes the instantaneous value

• Average Power density: Average of the power
flow over one period
 Phase Velocity

• We mostly deal with free space

 Lossy Media

• Alpha – attenuation constant

– Delta – “Skin depth”
• Sigma – Conductivitiy of the medium
Precise Expressions
Polarisation

• E-field
• Vertical or
Horizontal
polarisation
– Most Common
• Circular and
Elliptical
Polarisation also
exist
Antenna
Radiation
Near and Far Field (Fraunhoffer) Region
Spherical Coordinates
 Far Field Radiation from wires
(Hertzian dipole)

• Shown are the near (left) and far field (right) expressions
– NOTE: The phase term exp(jwt) is dropped in the above expressions
Radiation Pattern
 Generic Radiation Pattern and Directivity

• Directivity also reflects the gain of the antenna

• HPBW – Half Power Beam Width
 Reciprocity and Aperture

η = Ae/A

• A_e - Effective Aperture

– Related to Antenna Efficiency
– A mostly denotes the physical area of the antenna
 (1) Friis free space equation: receive power at antenna separated by
distance d from transmitter

 Gt Gr 2  Pt (3.1)
Pr(d) =  
2  2
 (4 ) L  d

Pr & Pt = received & transmitted power

Gt & Gr = gain of transmit & receive antenna
 = wavelength
d = separation
L = system losses (line attenuation, filters, antenna)
- not from propagation
- practically, L  1, if L = 1  ideal system with no losses

• power decays by d 2  decay rate = 20dB/decade

23
 Antenna Gain

4
G= 2
Ae (3.2)

• Ae = effective area of absorption– related to antenna size

Antenna Efficiency η = Ae/A

A = antenna’s physical area (cross sectional)

• for parabolic antenna η  45% -
50%
• for horn antenna η  50% - 80%

24
 (2) Radiated Power

Isotropic Radiator: ideal antenna (used as a reference antenna)

• radiates power with unit gain uniformly in all directions
• surface area of a sphere = 4πd 2

2
Effective Area of isotropic antennae given by Aiso =
4

 2   1  2
Isotropic Received Power PR =    P 
2  T
PT
 4   4d  4d 2

• d = transmitter-receiver separation

PT 4d 
2

Isotropic free space path loss Lp =  2

PR 
• f 2 relationship with antenna size results from dependence of Aiso on 

25
EIRP: effective isotropic radiated power
• represents maximum radiated power available from a transmitter
• measured in the direction of maximum antenna gain as compared to
isotropic radiator

EIRP = PtGiso (3.4)

ERP: effective radiated power - often used in practice

• denotes maximum radiated power compared to ½ wave dipole
antenna
• dipole antenna gain = 1.64 (2.15dB) > isotropic antenna
• thus EIRP will be 2.15dB smaller than ERP for same system

ERP = PtGdipole

26
Eg: Area of a parabolic antenna, looking into center-axis given by
beamwidth = angular width of the antenna beam at 3dB points
main lobe = direction of strongest signal propagation
center axis: direction of peak gain
side & back lobes not considered for communications, must be considered
in interference analysis
2 2 peak
D D gain
Aparabolic =  thus Ae = 
4 4
2 3dB Beamwidth
Ae D 2 4  D 
G=    
Aiso 4  2   

G depends on dimensions of antenna and frequency

assume D = 0.6m, fc = 12GHz, and η = 0.5

2
  0.6m  0.36 2
G =    2842.4 35dB
 c / 12GHz  0.025
27
if d  0  large scale models use d0 (received power reference point)
• for any d > d0  Pr(d) can be related to Pr(d0)

Pr(d0) predicted using 3.1 or average of many measurements at distance d0

• d0 must be selected to lie in far-field region d0  df
• d0 selected to be smaller than any practical d in mobile system
• practical choices for received power reference point are:
- indoors d0 ≈ 1m
- outdoors d0 ≈ 102m - 103m

2
 d0 
Pr(d) (watts) = Pr(d0)   d  d 0  df (3.8)
 d 
28
 Measuring in dB

dbm & dBW relative power measurements

• Pr in mobile systems can change by many dB in a coverage area ≈ 1km2
• dBm or dBW units are used to express power levels
• conversion from watts – take log of both sides & multiply by 10

 Pr (d 0 )   d0  d  d 0  df (3.9)
Pr(d) dBm = 10 log   20 log 
 0.001W  d 

e.g. Pr = 20 mW  Pr (dBm) = 10 log(20 mW/1mW) = 13 dBm

dBi – antenna gain with respect to isotropic source

dBd – antenna gain with respect to dipole antenna

29
India Spectrum auction – 900 MHz

• ~434
Cr per
MHz
India Spectrum auction – 1800 MHz

• ~110 Cr
per MHz
• 1800
MHz
attenuat
ion is
higher
than
900
MHz
BASIC PROPAGATION MECHANISMS
3.4 Basic Propagation Mechanisms & Transmission Impairments

(1) Reflection: propagating wave impinges on object with size >> 

• examples include ground, buildings, walls

(2) Diffraction: transmission path obstructed by objects with edges

• 2ndry waves are present throughout space (even behind object)
• gives rise to bending around obstacle (NLOS transmission path)

(3) Scattering propagating wave impinges on object with size < 

• number of obstacles per unit volume is large (dense)
• examples include rough surfaces, foliage, street signs, lamp posts

33
at high frequencies  diffraction & reflections depend on
• geometry of objects
• EM wave’s, amplitude, phase, & polarization at point of intersection

Models are used to predict received power or path loss (reciprocal)

based on refraction, reflection, scattering
• Large Scale Models
• Fading Models

34
3.5 Reflection: EM wave in 1st medium impinges on 2nd medium
• part of the wave is transmitted
• part of the wave is reflected

(1) plane-wave incident on a perfect dielectric (non-conductor)

• part of energy is transmitted (refracted) into 2nd medium
• part of energy is transmitted (reflected) back into 1st medium
• assumes no loss of energy from absorption (not practically)

(2) plane-wave incident on a perfect conductor

• all energy is reflected back into the medium
• assumes no loss of energy from absorption (not practically)

35
(3)  = Fersnel reflection coefficient relates Electric Field intensity
of reflected & refracted waves to incident wave as a function of:
• material properties,
• polarization of wave
• angle of incidence
• signal frequency

incident wave
reflected wave

boundary between dielectrics refracted wave

(reflecting surface)

36
(4) Polarization: EM waves are generally polarized
• instantaneous electric field components are in orthogonal directions
in space represented as either:
(i) sum of 2 spatially orthogonal components (e.g. vertical
& horizontal)

(ii) left-handed or right handed circularly polarized components

• reflected fields from a reflecting surface can be computed using

superposition for any arbitrary polarization

E||

E

37
3.5.1 Reflection from Dielectrics
• assume no loss of energy from absorption

EM wave with E-field incident at i with boundary between 2 dielectric media
• some energy is reflected into 1st media at r
• remaining energy is refracted into 2nd media at t
• reflections vary with the polarization of the E-field

plane of incidence

reflecting surface= boundary between i

dielectrics r
t

plane of incidence = plane containing incident, reflected, & refracted rays

38
Two distinct cases are used to study arbitrary directions of polarization

(1) Vertical Polarization: (Evi) E-field polarization is

• parallel to the plane of incidence
• normal component to reflecting surface

(2) Horizontal Polarization: (Ehi) E-field polarization is

• perpendicular to the plane of incidence
• parallel component to reflecting surface

Evi plane of incidence

Ehi

i
r
t
boundary between dielectrics
(reflecting surface)
39
Vertical Polarization: E-field in the plane of Horizontal Polarization: E-field normal to
incidence plane of incidence

Ei Er

Hi Ei Er Hr
Hi Hr
1,1, 1 i r 1,1, 1
i r
t 2,2, 2 t 2,2, 2

Et Et

• Ei & Hi = Incident electric and magnetic fields

• Er & Hr = Reflected electric and magnetic fields
• Et = Transmitted (penetrating) electric field

40
(1) EM Parameters of Materials
•  = permittivity (dielectric constant): measure of a materials ability
to resist current flow
•  = permeability: ratio of magnetic induction to magnetic field
intensity
•  = conductance: ability of a material to conduct electricity,
measured in Ω-1

dielectric constant for perfect dielectric (e.g. perfect reflector of lossless material) given
by

0 = 8.85 10-12 F/m

41
often permittivity of a material,  is related to relative permittivity r

 = 0  r

lossy dielectric materials will absorb power  permittivity described with complex dielectric
constant

 = 0 r -j’ (3.17)

where ’ =

(3.18)
2f
• 0 and r are generally constant
•  may be sensitive to operating frequency


f
highly conductive materials   0 r
• r &  are generally insensitive to operating frequency

42
 Material r  /r0 f (Hz)
Poor Ground 4 0.001 2.82  107 108
Typical Ground 15 0.005 3.77  107 108
Good Ground 25 0.02 9.04  107 108
Sea Water 81 5 6.97  109 108
Fresh Water 81 0.001 1.39  106 108
Brick 4.44 0.001 2.54  107 4 109
Limestone 7.51 0.028 4.21  108 4 109
Glass, Corning 707 4 0.00000018 5.08  103 106
Glass, Corning 707 4 0.000027 7.62  105 108
Glass, Corning 707 4 0.005 1.41  108 1010

43
(2) Reflections, Polarized Components & Fresnel Reflection
Coefficients

• because of superposition – only 2 orthogonal polarizations need be

considered to solve general reflection problem

Maxwell’s Equation boundary conditions used to derive (3.19-3.23)

Fresnel reflection coefficients for E-field polarization at reflecting

surface boundary
•  || represents coefficient for || E-field polarization

•   represents coefficient for  E-field polarization

44
Fersnel reflection coefficients given by

(i) E-field in plane of incidence (vertical polarization)

 || =
Er  2 sin t  1 sin i
 (3.19)
Ei  2 sin t  1 sin i

(ii) E-field not in plane of incidence (horizontal polarization)

Er  2 sin i  1 sin t
=  (3.20)
Ei  2 sin i  1 sin t

i = intrinsic impedance of the ith medium

• ratio of electric field to magnetic field for uniform plane wave in
ith medium

• given by i = i  i
45
velocity of an EM wave given by   
1

boundary conditions at surface of incidence obey Snell’s Law

 11 sin(90  1 )   2 2 sin(90   2 ) (3.21)

 1 
 t 90  sin  1
sin( 90   i ) 
  
 2 
i = r (3.22)

Er =  Ei (3.23a)

Et = (1 +  )Ei (3.23b)

 is either  || or   depending on polarization

• |  |  1  for a perfect conductor, wave is fully reflected
• |  |  0  for a lossy material, wave is fully refracted

46
 Simplification of reflection coefficients for vertical and horizontal
polarization assuming:

• radio wave propagating in free space (1st medium is free space)

• 1 = 2

  r sin i   r  cos2 i
 || = (3.24)
2
 r sin i   r  cos i

sin i   r  cos2 i
 = (3.25)
2
sin i   r  cos i

Elliptically Polarized Waves have both vertical & horizontal components

• waves can be depolarized (broken down) into vertical & horizontal
E-field components
• superposition can be used to determine transmitted & reflected
waves
47
Plot of Reflection Coefficients for Parallel Polarization for  r= 12, 4

1.0
||||
0.8
vertical polarization 0.6
(E-field in plane of incidence) 0.4 r=12
0.2 r=4
0.0
Brewster Angle (B) 0 10 20 30 40 50 60 70 80 90
for r=12 angle of incidence (i)

for i < B: a larger dielectric constant  smaller || & smaller Er
for i > B: a larger dielectric constant  larger || & larger Er

48
Plot of Reflection Coefficients for Perpendicular Polarization for  r= 12, 4

1.0 ||
0.9
horizontal polarization 0.8
(E-field not in plane of 0.7 r=12
0.6
incidence) 0.5 r=4
0.4
0.3
0 10 20 30 40 50 60 70 80 90
angle of incidence (i)

for given i: larger dielectric constant  larger  and larger Er

49
e.g. let medium 1 = free space & medium 2 = dielectric
• if i  0o (wave is parallel to ground)
• then independent of r, coefficients | |  1 and | |||  1

  r sin  i   r  cos2  i  r  cos2 i

 || =  1
2 2
 r sin i   r  cos i i 0
 r  cos i

sin i   r  cos2 i   r  cos2 i

 =   1
2 2
sin i   r  cos  i  i 0
 r  cos i

thus, if incident wave grazes the earth

• ground may be modeled as a perfect reflector with | | = 1
• regardless of polarization or ground dielectric properties
• horizontal polarization results in 180 phase shift

50
3.5.2 Brewster Angle = B

• Brewster angle only occurs for vertical (parallel) polarization

• angle at which no reflection occurs in medium of origin
• occurs when incident angle i is such that  || = 0  i = B

1
B satisfies sin(B) = (3.27)
1   2

• if 1st medium = free space & 2nd medium has relative permittivity  r
then (3.27) can be expressed as

r  1
sin(B) = (3.28)
 r2  1

51
e.g. 1st medium = free space

Let  r = 15 
15  1
sin(B) = 2 = 0.25
15  1
B = sin-1(0.25) = 14.5o

Let  r = 4 

sin(B) =
4 1 = 0.44
16  1
B = sin-1(0.44) = 26.6o

52
3.6 Ground Reflection – 2 Ray Model

Free Space Propagation model is inaccurate for most mobile RF

channels

2 Ray Ground Reflection model considers both LOS path & ground
reflected path
• based on geometric optics
• reasonably accurate for predicting large scale signal strength for
distances of several km
• useful for
- mobile RF systems which use tall towers (> 50m)
- LOS microcell channels in urban environments

Assume
• maximum LOS distances d  10km
• earth is flat

53
(1) Determine Total Received E-field (in V/m) ETOT

ETOT is combination of ELOS & Eg

• ELOS = E-field of LOS component
ELOS
• Eg = E-field of ground reflected component
Ei
• θi = θr Eg
Er =
i 0

d
Let E0 = free space E-field (V/m) at distance d0

• Propagating Free Space E-field at distance d > d0 is given by

E0 d 0   d 
E(d,t) = cos wc  t    (3.33)
d   c 
• E-field’s envelope at distance d from transmitter given by

|E(d,t)| = E0 d0/d
54
 E-field for LOS and reflected wave relative to E0 given by:

E0 d 0   d'
ELOS(d’,t) = cos wc  t    (3.34)
d'   c 

E0 d 0   d"  
Eg(d”,t) = Γ cos wc  t   (3.35)
d"   c 
and ETOT = ELOS + Eg
ELOS d’

ht Ei
hr
i Eg 0
d”

assumes LOS & reflected waves arrive at the receiver with

- d’ = distance of LOS wave
- d” = distance of reflected wave

55
From laws of reflection in dielectrics (section 3.5.1)

i = 0 (3.36)

Eg =  Ei (3.37a)

Et = (1+ ) Ei (3.37b)

 = reflection coefficient for ground

ELOS d’

Ei
Eg
i 0
d”

Et

56
Assume
i. perfect horizontal E-field Polarization
ii. perfect ground reflection
iii. small i (grazing incidence)   ≈ -1 & Et ≈ 0
• reflected wave & incident wave have equal magnitude
• reflected wave is 180o out of phase with incident wave
• transmitted wave ≈ 0

resultant E-field is vector sum of ELOS and Eg

• total E-field Envelope is given by |ETOT| = |ELOS + Eg| (3.38)
• total electric field given by

E0 d 0   d '   Ed   d"  
ETOT(d,t) = cos wc  t     ( 1) 0 0 cos wc  t    (3.39)
d'   c  d"   c 
57
 (2) Compute Phase Difference & Delay Between Two Components

• path difference  = d” – d’ determined from method of images

 = h t  hr   d 2 
2
h
t  hr   d 2
2
(3-40)

ELOS

h r h
Ei
d’
ht Eg
i d” 0
ht +hr

d
if d >> hr + ht  Taylor series approximations yields (from 3-40)

  2ht hr (3-41)
d 58
once  is known  we can compute

• phase difference  = 2  wc (3-42)


 c

Δ
0 π 2π if Δ = /n   = 2π/n

 
• time delay d =  (3-43)
c 2f c
As d becomes large   = d”-d’ becomes small
• amplitudes of ELOS & Eg are nearly identical & differ only in phase

E0 d 0 E0 d 0 E0 d 0
  (3.44)
d d' d"
59
(3) Evaluate E-field when reflected path arrives at receiver

reflected path arrives at receiver at t = d”/c

E0 d 0   d " d '   E0 d 0
ETOT(d,t)|t=d”/c = cos wc     ( 1) cos0  (3.45)
d'   c  d"

E0 d 0    

 cos  wc   1
d   c 
E0 d 0
= cos   1
d
E0 d 0
=    1
d

60
 (4) Determine exact E-field for 2-ray ground model at distance d

Use phasor diagram to find resultant E-field from combined direct & ground reflected rays:

E0 d 0
d'
ETOT

E0 d 0
d"
2
cos 
2 2
 E0 d 0  2  E d   E0 d 0 
|ETOT(d)|=   cos    1   0 0  sin   =
2
    1
2
 sin 2

 d   d   d 
(3.46)

 E0 d 0 
=   2  2 cos  (3.47)
 d 

 E d   
= 2 0 0  sin    (3.48)
 d   2
61
As d increases  ETOT(d) decreases in oscillatory manner
• local maxima 6dB > free space value
• local minima ≈ - dB (cancellation)
• once d is large enough  θΔ < π & ETOT(d) falls off asymtotically
with increasing d

Propagation Loss ht = hr = 1, Gt = Gr = 0dB

-50
fc = 3GHz
-60
fc = 7GHz
-70
fc = 11GHz
-80
-90
-100
-110
-120
-130
-140
101 102 103 104 m
62
For phase difference,  < 0.6 radians (34o) sin(0.5 )  

 E0 d 0     
|ETOT(d)|  2  
 d  2 
 1 2 2ht hr
   0.3rad (3.49)
2 2  d

this implies d> 20ht hr 20ht hr (3.50)


3 
if d satisfies 3.50  total E-field can be approximated as:

ETOT(d)  2 E0 d 0  2ht hr  k V/m (3.51)

  2
d  d  d
k is a constant related to E0 ht,hr, and 

e.g. at 900MHz  if  < 0.03m  total E-field decays with d2

63
Received Power at d is related to square of E-field by 3.2, 3.15, & 3.51

ht2hr2
Pr(d) = Pt Gt Gr 4 (3.52a)
d

E0 (d ) 2 ER (d ) 2  Gr 2 
Pr(d) = Ae    (3.52b)
120 120  4 

if d >> ht hr
• received power falls off at 40dB/decade
• receive power & path loss become independent of frequency

64
Path Loss for 2-ray model with antenna gains is expressed as:

1
PL = Pt  ht2hr2 
 Gt Gr 
Pr  d4 
PL(dB) = 40log d - (10logGt + 10logGr + 20log ht + 20 log hr ) (3.53)

• 3.50 must hold

• for short Tx-Rx distances use (3.39) to compute total E field

• evaluate (3.42) for  =  (180o)  d = 4hthr/ is where the ground

appears in 1st Fresnel Zone between Tx & Rx
- 1st Fresnel distance zone is useful parameter in microcell path
loss models

65
3.7 Diffraction

• allows RF signals to propagate to obstructed (shadowed) regions

- over the horizon (around curved surface of earth)
- behind obstructions
• received field strength rapidly decreases as receiver moves into
obstructed region
• diffraction field often has sufficient strength to produce useful signal

Segments
3.7.1 Fresnel Zone Geometry

66
Huygen’s Principal
• all points on a wavefront can be considered as point sources for
producing 2ndry wavelets
• 2ndry wavelets combine to produce new wavefront in the direction
of propagation
• diffraction arises from propagation of 2ndry wavefront into
shadowed area
• field strength of diffracted wave in shadow region =  electric field
components of all 2ndry wavelets in the space around the obstacle

slit knife edge

67
3.7.1 Fresnel Zone Geometry
• consider a transmitter-receiver pair in free space
• let obstruction of effective height h & width   protrude  to page
- distance from transmitter = d1
- distance from receiver = d2
- LOS distance between transmitter & receiver = d = d1+d2

Excess Path Length = difference between direct path & diffracted path

 = d – (d1+d2) 2
d =  d1+  d2, where ,  di = h  d i2
h d
= h d +
2 2 2
h d 2 – (d1+d2)
1 2
TX RX
d1 d2
ht hobs hr

Knife Edge Diffraction Geometry for ht = hr

68
 Assume h << d1 , h << d2 and h >>  then by substitution and Taylor
Series Approximation
2
  h  d1  d 2  3.54
 
2  d1d 2 
Phase Difference between two paths given as

2 2  h 2 d1  d 2   2  2d1  d 2  3.55

=    = h  
   2 d1d 2  2  d1d 2 

 h h’
TX
d1
d2 
RX
ht hobs
hr

Knife Edge Diffraction Geometry ht > hr

69
Equivalent Knife Edge Diffraction Geometry with hr subtracted from all other heights

 180-
TX
 hobs-hr
ht-hr 
RX
d1 d2

when tan x  x   =  + 

h
tan  =  tan(x)
 d1
x = 0.4 rad  tan(x) = 0.423
h
tan  =  (0.4 rad ≈ 23o )
 d2

h h  d1  d 2 
  h 
d1 d 2  d1d 2 
x
70
Eqn 3.55 for  is often normalized using the dimensionless Fresnel-Kirchoff diffraction
parameter, v

2(d1  d 2 ) 2d1d 2
v= h  (3.56)
d1d 2  (d1  d 2 )

when  is in units of radians  is given as

= 
v2 (3.57)
2

from equations 3.54-3.57   , the phase difference, between LOS & diffracted path is
function of
• obstruction’s height & position
• transmitters & receivers height & position

simplify geometry by reducing all heights to minimum height

71
(1) Fresnel Zones
• used to describe diffraction loss as a function of path difference, 
around an obstruction
• represents successive regions between transmitter and receiver
• nth region = region where path length of secondary waves is n /2
greater than total LOS path length
• regions form a series of ellipsoids with foci at Tx & Rx

d
λ/2 + d

λ+d

at 1 GHz λ = 0.3m 1.5λ + d

72
Construct circles on the axis of Tx-Rx such that  = n/2, for given integer n
• radii of circles depends on location of normal plane between Tx and Rx
• given n, the set of points where  = n/2 defines a family of ellipsoids
• assuming d1,d2 >> rn R

n  h 2 d1  d 2 
= 
2  2 d1d 2 
T

slice an ellipsoid with a plane  yields circle with radius rn given as

nd1d 2
h = rn =
d1  d 2

then Kirchoff diffraction parameter is given as

2d1  d 2  nd1d 2 2d1  d 2 
v= h  = 2n
d1d 2 d1  d 2 d1d 2

thus for given rn  v defines an ellipsoid with constant  = n/2 73

nth Fresnel Zone is volume enclosed by ellipsoid defined for n and is defined
as  relative to LOS path

n  1 n
≤Δ≤
2 2
• 1st Fresnel Zone is volume enclosed by ellipsoid defined for n = 1

Phase Difference,  pertaining to nth Fresnel Zone is

(n-1) ≤  ≤ n

• contribution to the electric field at Rx from successive Fresnel Zones

tend to be in phase opposition  destructive interference
• generally must keep 1st Fresnel Zone unblocked to obtain free space
transmission conditions

74
 For 1st Fresnel Zone, at a distance d1 from Tx & d2 from Rx
• diffracted wave will have a path length of d

d

Tx d1 d2 Rx

destructive interference 
•  = /2
• d = /2 + d1+d2

For 2nd Fresnel Zone

constructive interference:
• d =  + d1+d2
• =

75
Fresnel Zones
• slice the ellipsoids with a transparent plane between transmitter &
receiver – obtain series of concentric circles
• circles represent loci of 2ndry wavelets that propagate to receiver
such that total path length increases by  /2 for each successive circle
• effectively produces alternatively constructive & destructive
interference to received signal

Q R
h
d2
O
T d1

• If an obstruction were present, it could block some of the Fresnel

zones
76
Assuming, d1 & d2 >> rn  radius of nth Fresnel Zone can be given in terms of n, d1,d2, 

nd1d 2
rn = (3.58)
d1  d 2

• radii of concentric circles depends on location between Tx & Rx

- maximum radii at d1 = d2 (midpoint), becomes smaller as plane
moves towards receiver or transmitter
- shadowing is sensitive to obstruction’s position and frequency

Excess Total Path Length,  for each ray passing through nth circle

n  =n/2 Rx
1 /2
2 
3 3/2 Tx
77
(2) Diffraction Loss caused by blockage of 2ndry (diffracted) waves

partial energy from 2ndry waves is diffracted around an obstacle

• obstruction blocks energy from some of the Fresnel zones
• only portion of transmitted energy reaches receiver

received energy = vector sum of contributions from all unobstructed

Fresnel zones
• depends on geometry of obstruction
• Fresnel Zones indicate phase of secondary (diffracted) E-field

Obstacles may block transmission paths – causing diffraction loss

• construct family of ellipsoids between TX & RX to represent
Fresnel zones
• join all points for which excess path delay is multiple of /2
• compare geometry of obstacle with Fresnel zones to determine
diffraction loss (or gain)

78
Diffraction Losses
Place ideal, perfectly straight screen between Tx and Rx
(i) if top of screen is well below LOS path  screen will have little effect
- the Electric field at Rx = ELOS (free space value)

(ii) as screen height increases E will vary up & down as screen blocks more
Fresnel zones below LOS path
amplitude of oscillation increases until just in line with Tx and Rx
 field strength = ½ of unobstructed field strength

Rx

Tx

79
Fresnel zones: ellipsoids with foci at transmit & receive antenna
• if obstruction does not block the volume contained within 1st Fresnel
zone  then diffraction loss is minimal
• rule of thumb for LOS uwave:
if 55% of 1st Fresnel zone is clear  further Fresnel zone clearing
does not significantly alter diffraction loss

e.g. 2(d1  d 2 ) 2d1d 2

v= h 
d1d 2  (d1  d 2 )

 excess path length

RX  /2
TX h 
3 /2
d1 d2

 and v are positive, thus h is positive

80
 2(d1  d 2 ) 2d1d 2
v= h 
d1d 2  (d1  d 2 )

d1 d2
TX RX

h = 0   and v =0

TX RX
d1 h d2


 and v are negative h is negative

81
3.7.2 Knife Edge Diffraction Model

Diffraction Losses
• estimating attenuation caused by diffraction over obstacles is
essential for predicting field strength in a given service area
• generally not possible to estimate losses precisely
• theoretical approximations typically corrected with empirical
measurements

Computing Diffraction Losses

• for simple terrain  expressions have been derived
• for complex terrain  computing diffraction losses is complex

82
Knife-edge Model - simplest model that provides insight into order of magnitude for
diffraction loss
• useful for shadowing caused by 1 object  treat object as a knife edge
• diffraction losses estimated using classical Fresnel solution for field
behind a knife edge

Consider receiver at R located in shadowed region (diffraction zone)

• E- field strength at R = vector sum of all fields due to 2ndry Huygen’s
sources in the plane above the knife edge

Huygens 2nddry
source
T h’
d1 R
d2

Knife Edge Diffraction Geometry, R located in shadowed region 83

Electric field strength, Ed of knife-edge diffracted wave is given by:

Ed 1  j   exp  jt 2 dt

E0
= F(v) =
2 
v

 2 

(3.59)

F(v) = Complex Fresnel integral

• v = Fresnel-Kirchoff diffraction parameter
• typically evaluated using tables or graphs for given values of v

E0 = Free Space Field Strength in the absence of both ground

reflections & knife edge diffraction

84
Gd(dB) = Diffraction Gain due to knife edge presence relative to E0

• Gd(dB) = 20 log|F(v)| (3.60)

Graphical Evaluation

5
0
-5
-10
Gd(dB)

-15
-20
-25
-30-3 -2 -1 0 1 2 3 4 5 v

85
 Table for Gd(dB)

Gd(dB) v
0  -1
20 log(0.5-0.62v) [-1,0]
20 log(0.5 e- 0.95v) [0,1]
20 log(0.4-(0.1184-(0.38-0.1v)2)1/2) [1, 2.4]
20 log(0.225/v) > 2.4

86
e.g. Let:  = 0.333 (fc = 900MHz), d1 = 1km, d2 = 1km, h = 25m
Compute Diffraction Loss at h = 25m
1. Fresnel Diffraction Parameter
2(d1  d 2 ) 2(2000)
v= h 25 = 2.74
d1d 2 0.333(106 )

2. diffraction loss
• from graph is Gd(dB)  -22dB
• from table Gd(dB)  20 log (0.225/2.74) = - 21.7dB

3. path length difference between LOS & diffracted rays

h 2  d1  d 2  252  2000 
      6  0.625m
2  d1d 2  2  10 

4. Fresnel zone at tip of obstruction (h=25)

• solve for n such that  = n /2
• n = 2· 0.625/0.333 = 3.75
• tip of the obstruction completely blocks 1st 3 Fresnel zones
87
e.g. Let:  = 0.333 (fc = 900MHz), d1 = 1km, d2 = 1km, h = 25m
Compute Diffraction Loss at h = -25m
1. Fresnel Diffraction Parameter
2(d1  d 2 ) 2(2000)
v= h  25 6 = -2.74
d1d 2 0.333(10 )

2. diffraction loss from graph is Gd(dB)  1dB

3. path length difference between LOS & diffracted rays

h 2  d1  d 2   252  2000 
      0.625m
2  d1d 2  2  106 

4. Fresnel zone at tip of the obstruction (h = -25)

• solve for n such that  = n /2
• n = 2· 0.625/0.333 = 3.75
• tip of the obstruction completely blocks 1st 3 Fresnel zones
• diffraction losses are negligible since obstruction is below LOS path
88
find diffraction loss
f = 900MHz   = 0.333m T
 = tan-1(75-25/10000) = 0.287o R
50m 100m
25m
 = tan-1(75/2000) = 2.15o
10km 2km
 =  +  = 2.43o = 0.0424 radians
2d1d 2 
v=  T
 (d1  d 2 ) 
75m
25m  R
2(10000)(2000)
= 0 .0424 4.24 10km 2km
0.333(12000)

from graph, Gd(dB) = -25.5 dB

find h if Gd(dB) = 6dB

T  =0
• for Gd(dB) = 6dB  v ≈ 0  h
25m  R
• then  = 0 and  = - 
• and h/2000 = 25/12000  h = 4.16m 10km 2km
89
3.7.3 Multiple Knife Edge Diffraction
• with more than one obstruction  compute total diffraction loss

(1) replace multiple obstacles with one equivalent obstacle

• use single knife edge model
• oversimplifies problem
• often produces overly optimistic estimates of received signal
strength

(2) wave theory solution for field behind 2 knife edges in series

• Extensions beyond 2 knife edges becomes formidable

• Several models simplify and estimate losses from multiple obstacles

90
3.8 Scattering

RF waves impinge on rough surface  reflected energy diffuses in all directions

• e.g. lamp posts, trees  random multipath components
• provides additional RF energy at receiver
• actual received signal in mobile environment often stronger than
predicted by diffraction & reflection models alone

91
Reflective Surfaces
• flat surfaces has dimensions >> 
• rough surface often induces specular reflections
• surface roughness often tested using Rayleigh fading criterion
- define critical height for surface protuberances hc for given
incident angle i


hc = (3.62)
8 sin i

Let h = maximum protuberance – minimum protuberance

• if h < hc  surface is considered smooth
• if h > hc  surface is considered rough h

92
h = standard deviation of surface height about mean surface height

stone – dielectric properties

•  r = 7.51
•  = 0.028
•  = 0.95

rough stone parameters

• h = 12.7cm
• h = 2.54

93
For h > hc  reflected E-fields can be solved for rough surfaces using modified reflection
coefficient

rough = s  (3.65)

(i) Ament, assume h is a Gaussian distributed random variable with a

local mean, find s as:

   h sin i  2 
s = exp    (3.63)
   


(ii) Boithias modified scattering coefficient has better correlation

with empirical data

   h sin i  2     h sin i  2 
s = exp  8   I 0  8   (3.64)
       l  

I0 is Bessel Function of 1st kind and 0 order
94
Reflection Coefficient of Rough Surfaces

(1)  polarization (vertical antenna polarization)

• ideal smooth surface

• Gaussian Rough Surface
• Gaussian Rough Surface (Bessel)
• Measured Data forstone wall h = 12.7cm, h = 2.54

| |

1.0
0.8
0.6
0.4
0.2
0.0

0 10 20 30 40 50 60 70 80 90
angle of incidence

95
Reflection Coefficient of Rough Surfaces

(2) || polarization (horizontal antenna polarization)

• ideal smooth surface

• Gaussian Rough Surface
• Gaussian Rough Surface (Bessel)
• Measured Data forstone wall h = 12.7cm, h = 2.54

||
1.0
0.8
0.6
0.4
0.2
0.0

0 10 20 30 40 50 60 70 80 90
angle of incidence

96
3.8.1 Radar Cross Section Model (RCS)
• if a large distant objects causes scattering & its location is known
 accurately predict scattered signal strengths

power density of signal scattered in direction of the receiver

RCS =
power density of radio wave incident upon scattering object

• units = m2

• determine signal strength by analysis using

- geometric diffraction theory
- physical optics

97
Urban Mobile Radio
Bistatic Radar Equation used to find received power from
scattering in far field region
• describes propagation of wave traveling in free space that
impinges on distant scattering object
• wave is reradiated in direction of receiver by:

Pr(dBm) = Pt (dBm) + Gt(dBi) + 20 log( ) + RCS [dB m2]

– 30 log(4 ) -20 log dT - 20log dR

• dT = distance of transmitter from the scattering object

• dR = distance of receiver from the scattering object
• assumes object is in the far field of transmitter & receiver

98
RCS can be approximated by surface area of scattering object (m 2)
measured in dB relative to 1m2 reference
• may be applied to far-field of both transmitter and receiver
• useful in predicting received power which scatters off large
objects (buildings)
• units = dB m2
• [Sei91] for medium and large buildings, 5-10km

14.1 dB  m2 < RCS < 55.7 dB  m2

99
Thank You
 Acknowledgements

The contents of the slides have been taken

from various sources for explanation purposes
only. No copyright violation intended.
Backup Slides
For parabolic antenna: increasing D  increases gain, decreases beamwidth
2
 D 
Gain on center axis, G =   
  

Gain off-center axis varies with angle 

normalized
G(dB)
2 2 0
 2 J D /  sin    D 
G() =  1    -10
 sin     
-20
• normalized to unity center axis gain
-30
• J1(x) is Bessel function of 1st kind
-40
-30 -20 -10 0 10 20 30 
3dB beamwidth = 58.4  /D
• proportional to wavelength
• inversely proportional to diameter

103
3.2 Free Space Propagation Model
used to predict signal strength for LOS path
• satellites
• LOS uwave
• power decay  d –n (d = separation)

Subsections
(1) Friis Equation
(2) Radiated Power
(3) Path Loss
(4) Far Field Region

104
Directional Radiation
practical antennas have gain or directivity that is a function of
• θ = azimuth: look angle of the antenna in the horizontal plane
•  = elevation: look angle of the antenna above the horizontal plane


Let Φ = power flux desnity
θ
transmit antenna gain is given by:
Φ in the direction of (θ, )
GT(θ, ) =
Φ of isotropic antenna

receive antenna gain is given by:

Ae in the direction of (θ, )
GR(θ, ) =
Ae of isotropic antenna

105
Principal Of Reciprocity:
• signal transmission over a radio path is reciprocal
• the locations of TX & RX can be interchanged without changing
transmission characteristics

signals suffers exact same effects over a path in either direction in a

consistent order  implies that GT(θ, ) = GR(θ, )

thus maximum antenna gain in either direction is given by

Ae 4
G=  Ae
Aiso  2

106
(3) Path Loss (PL)

Pt  (4 ) 2 d 2 
PL = =  
Pr  G G 2 
 t r 
 Gt Gr 2  1
PL (dB) = 10 log 10 (Pt /Pr) =  10 log10   (3.5)
 (4 ) 2  d 2
 

if G is assumed unit gain:

 2  1
PL (dB) =  10 log10  
2 2
(3.6)
 (4 )  d

107
 (4) Far-field region of an antenna

eqn 3.1 only valid for d in the far-field of transmitting antenna

• d  df (far-field distance)
• far-field distance aka Fraunhofer region

2D 2
df = (3.7)


• D = largest physical linear dimension of transmitters antenna aperture

• further df >> D and df >>  must hold

108
 3.3 Power and Electric Fields
• radiating structures produce electric fields (E) & magnetic fields (H)
• consider small linear radiator of length L, where L <<
• current flows through antenna  resulting electrical & magnetic
fields expressed as
i0
i0 L cos   1 c   jwc ( t  d / c ) 
Er =  2  e
3 
(3.10) P
2 0 c d jwc d 
L d

i0 L sin  jwc c   jwc (t  d / c )

E = 
2 
 e
3
(3.11)
4 0 c  d jwc d 
z
i0 L sin   jwc c   jwc (t  d / c )
H =   2 e (3.12)  r
4c  d d  
y
H , E , Hr all = 0 x
109
 term component represented
d-1 radiation field
d-2 induction field
d-3 electrostatic field

• electrostatic (d-3) & induced (d-2) fields decay much faster than
radiated field
• far from transmitter (d > df)  electrostatic & inductive fields ≈ 0

• if d > df only radiated field components are considered E & H

Er ≈ 0

i0 L sin  jwc   jwc (t  d / c )

E ≈ 2  e
4 0 c  d 
i0 L sin  jwc  jwc ( t  d / c )
H ≈  e
4c  d 
110
(1) Power flux density Pd in free space (W/m2)

• |E | = magnitude of radiating portion of electric field in far-field (V/m)

• Pd disperses in free space from an isotropic point source

• Pd ≈ EIRP  surface area of sphere of radius d 1m 1m

• surface are of a sphere = 4πd2

d PtGt

EIRP Pt Gt E2 E2
Pd =
2
 2
  (W/m2) (3.13)
4d 4d R fs 

Rfs = intrinsic impedance of free space given by  = 120Ω

2 2
Pd = E E (W/m2) (3.14)

120  377
111
 (2) Received Power, Pr(d)

Pr(d) = power flux density  receive antenna effective aperture

• received power at distance d from transmitter
• can be related to electric field using (3.1), (3.2), (3.13),(3.14)

Induced electric field (V/m) at receiver vs received power (W)

• assume L = 1 (system loss)

2 2
Pr(d) =
E PG G  2
E Gr 2 (3.15)
Pd Ae  Ae  t t 2r 2  (W )
120 ( 4 ) d 480

• |E | = magnitude of radiating portion of electric field in far-field

112
(3) Receiver Input Voltage And Receive Power Level
• model receive antenna as matched resistive load, Rant
• receiver antenna will induce rms voltage, V, into receiver
• induced voltage = ½ open circuit voltage at antenna: V = ½ Vant
• Rant = antenna resistance
open circuit Rant
to matched
Vant V receiver

Pr(d) =
V / 22  V2
(3.16)
Rant 4 Rant

Induced Electric Field, E vs Receiver Input Voltage, V

2
Pr(d) = E V2
Ae 
120 4 Rant
113
(3) General Case of reflection or transmission
• horizontal & vertical axes of spatial coordinates may not coincide
with || &  axes of propagating waves
• for wave propagating out of the page  define angle  
measured counter clock-wise from horizontal axes

spatial vertical axis

||
orthogonal components
of propagating wave 
 spatial horizontal axis

114
relationship of vertical & horizontal field components at the dielectric boundary

vertical & horizontal components perpendicular

polarized components  & parallel to plane of incidence

EiH , EiV EdH , EdV

 E Hd  T  E i

 E d  R DC R  E i 
H (3.26)
 v   v

EdH, EdV EiH , EiV = Time Varying Components of E-field

• EdH , EdV = depolarized field components along the horizontal &

vertical axes
• EiH , EiV = horizontal & vertical polarized components of incident
wave

- E-field components may be represented by phasors

115
R = transformation matrix that maps E-field components

 cos sin  
R= ,  = angle between two sets of axes
  sin  cos 
 

DC = depolarization matrix

 D 0 
DC =  0
 D|| || 
for case of reflection:
• D =  
• D|| || =  ||

for case of refraction (transmission):

• D = 1+  
• D|| || = 1+  ||

116