Wireless Channels

Path Loss and Shadowing

A. Özgür Yılmaz - METU

EE 728 METU AOY 1

 Wireless channel susceptible to
 Noise
 Interference
 Channel impediments
 Impediments change over time unpredictably
due to
 User movement
 Environment dynamics
 Channel impediments
 Path loss and shadowing (~deterministic, large
scale)
 Multipath (~statistical, small scale)

EE 728 METU AOY 2

 Path Loss and Shadowing
 Free space propagation, line of sight (LOS)
attenuation
 An isotropic tx antenna with power P Watts
t
 Power density at distance d

d
Pt
W / m2
4d 2

 If tx antenna has directivity Gt Pt 2 (field radiation

pattern) 4d

 Rx antenna gathers a portion of the radiated power

proportional to its cross-sectional area.
Pt
Gt Ar
4d 2

EE 728 METU AOY 3

 2
 From EMT Ar  Gr
4

 Friis transmission formula

c2
Pr  Pt Gt Gr
(4fd ) 2

Received
power

time
EE 728 METU AOY 4
 Remarks
 Gains depend on antenna physical properties.
 Other losses (atmosperic absorption) sometimes
effective 1
 2 
 
Path loss PL   
2 
 (4d ) 
Pr  Pt Gt Gr / PL

 Usually in dB
( Pr ) dB  ( Pt ) dB  (Gt ) dB  (Gr ) dB  ( PL ) dB  (Pa ) dB 

 Used directly for satellite communications and

radio links

EE 728 METU AOY 5

 Transmitted signal

 Equivalent lowpass representation of

bandpass signals
 u(t ) complex envelope, equivalent lowpass signal
 Received signal

EE 728 METU AOY 6

 c2
Pr  Pt Gt Gr
(4fd ) 2

 With free space path loss

c Gl
r (t )  Re{ u(t   )e j 2f ( t  ) }
4fd
 Link gain comprised of transmit and receive
antenna gains
Gt
Gr
Gl

 Delay due to the distance traveled by the EM

waves

EE 728 METU AOY 7

 Path loss (usually antenna gains excluded)
2
Gl Pt  c 
PL    
Pr  4fd 
 c 
PL ,dB  20 log10  
 4fd 

EE 728 METU AOY 8

 Example

EE 728 METU AOY 9

 Atmospheric attenuation
 Oxygen, water
 Rain, fog
 Height dependent

http://www.rfcafe.com/references/electrical/atm_absorption.htm

EE 728 METU AOY 10

 http://www.tscm.com

EE 728 METU AOY 11

Ray Tracing
Multipath
signal
components

 Many objects in surroundings

 Reflection (EMW on an object larger than wavelength)
 Diffraction (path obstructed by a surface with sharp
irregularities)
 Scattering (medium densely consists of objects smaller than
wavelength)
EE 728 METU AOY 12
 Solve Maxwell’s equations with boundary
conditions
 Too complex
 Everything should be perfectly known
 Simplification necessary
 Ray tracing
 Assume a finite number of reflectors with known
location and dielectric properties

EE 728 METU AOY 13

 Two-ray model

x
x'

Ground reflection coefficient


  Gl u (t   )e  j 2l /  R Gr u (t   )e  j 2 ( x  x ') /   j 2f t 

r (t )      e c

 4
  l x  x '  

l x  x
 ,  Gl  Ga Gb Gr  Gc Gd
c c

EE 728 METU AOY 14

 If transmitted signal is slowly changing in
relation to  , u(t   )  u(t ), u(t   )  u(t )
2
2  j
   Gl R Gr e
Pr  Pt    ,   2 ( x  x'l ) / 
 4  l x  x'

x  x'l  ht  hr 2  d 2  ht  hr 2  d 2

2
1 x2  1 x , 0  x  1
2
2ht hr
x  x'l 
d
4ht hr
 
d

EE 728 METU AOY 15

 Asymptotic case
 d is large x  x'  l  d ,   0
 Gl  Gr
 For earth and road surfaces R  1

 Receiver power falls off as d 4

 Independent of frequency since combination of
two rays effectively forms an antenna array
(antenna array gain does not necessarily
decrease with frequency)

EE 728 METU AOY 16

900MHz
Gl  Gr  1
R  1
ht  50m, hr  2m

EE 728 METU AOY 17

 Example

EE 728 METU AOY 18

 Ten-ray model
 For urban microcells
 Flat city with 90 degrees intersecting linear streets
(rectilinear streets)
 Buildings along both sides of streets
 Building-lined streets act as dielectric canyon to
the propagating signal.
 Since signal energy is dissipated with each
reflection, more than 3 reflections can be
generally ignored.
EE 728 METU AOY 19
 *

2
 Typical power falloff  d
 Some empirical studies obtained power falloff

proportional to d ,2 6

EE 728 METU AOY 20

 Generalized ray tracing
 Diffracted and scattered rays also taken into
account
 Leads to a complicated path loss model.
 Simplifications needed
 Local mean received power (LMRP)
 Ray tracing depends on exact tx/rx locations
(phase)
 Only a mean received power usually required for
link quality
 Cellular systems utilize LMRP for power control
and handoff

EE 728 METU AOY 21

Empirical Path-Loss Models
 Most wireless systems operate in complex
propagation environments
 Cannot be accurately modeled by free space
prop. or ray tracing
 Path-loss models developed over the years
to predict path loss in typical wireless
environments
 Large urban macrocells
 Urban microcells
 Inside buildings, …
 Never forget: these are just models!

EE 728 METU AOY 22

 These models based on empirical
measurements
 Over a given distance
 In a given frequency range
 For a particular geographical area or building
 One must be careful in using these models
for other scenarios.
 Path loss, shadowing, multipath all contribute
to received power in empirical measurements
 Averaging to remove multipath effects
 Local mean attenuation over several 
 Repetitions throughout the environment
 Repetitions in similar environments
EE 728 METU AOY 23
 Okumura Model
 Large urban macrocells, 1-100kms
 150-1500MHz
 Base station-to-mobile measurements in Tokyo

Path Median Antenna Gain due to

loss attenuation height gains type of
environment
 Empirical formulas

 Others obtained from Okumura’s empirical plots

 Corrections proposed later
EE 728 METU AOY 24
 Hata model
 Closed-form formula for Okumura’s model

 Frequency in MHz, distance in km

 Correction factor for the mobile antenna height
based on the size of the coverage area
 Small-to-medium size cities

 Larger cities f c  300MHz

 Other environments

 K ranges from 39.54(countryside) to 40.94(desert)

EE 728 METU AOY 25

  Hata well approximated Okumura for d  1km
 Hata does not model well the current cellular
systems with smaller cell sizes and higher
frequencies, indoor environments
 COST 231 Extension to Hata
 1.5-2GHz, 1-20kms
 30m  h t  200m, 1m  h r  10m

 C M 0dB for medium-sized cities, suburbs; 3dB for

metropolitan areas
 Piecewise Linear (Multislope) Model
 Empirical measurements are fitted to piecewise
linear functions
EE 728 METU AOY 26
EE 728 METU AOY 27
 Indoor Attenuation Factors
2
 Penetration through
1
 Walls
Floor 0
 Floors
 Objects
 Glass, …
 All these factors significantly affect indoor path
loss
 Floors
 Depends on building material
 Attenuation largest for the 1st passed floor (10-20dB)
 Decreases with subsequent floors (6-10dB, a few dB for
larger than 4 floors)
 Rappaport has details

EE 728 METU AOY 28

 Partition losses (walls)

 Losses by different studies vary widely

 Very hard to make generalizations

 FAF: floor attenuation factor

 PAF: partition attenuation factor

EE 728 METU AOY 29

 Simplified Path-Loss Model
 Simplified models necessary for system design

 K unitless constant depending on antenna

characteristics and average channel attenuation
 d 0 reference distance for antenna far field
  path-loss coefficient

EE 728 METU AOY 30

Shadowing
 Obstacles between transmitter and
receiver
 Signal attenuated

EE 728 METU AOY 31

 Bad reception

Good reception

 Random variation of received power due to

blockage from objects in signal path
 The exact locations and impact of the
blocking objects are usually unknown =>
statistical models
 The most common model: log-normal pdf
 Gain in dB is normal.

EE 728 METU AOY 32

 Pt Pt
  ,  dB  10 log10  N (  dB , 2 dB )
Pr Pr

   1  Pt  Pr physically impossible

 Log-normal model captures the underlying

physical model most correctly when  dB  0

EE 728 METU AOY 33

 Mathematical justification for log-normal
 Attenuation due to an object e  i   i
 Attenuation due to many objects e i
 CLT after taking logarithm
 Shadowing is a random process
 Assumption: WSS
 Covariance between shadow fading at two
points separated by distance 

EE 728 METU AOY 34

 Decorrelation distance X c is where
correlation drops to 1/e of max
 Shadowing r.p.
 White noise passed through a first-order IIR filter
(AR)

EE 728 METU AOY 35

Combined Path Loss and Shadowing

 Combination of
 simplified path loss  dB
 Zero mean shadow fading creating variations in
received power
 Pr  d
   10 log10 K  10 log10   ,  ~ N ( 0,  2
 dB )
P  dB dB
 t  dB d0

Slowly Rapidly
changing changing

EE 728 METU AOY 36

 Outage probability under path loss and
shadowing
 Outage: event that the received power falls below
a predetermined power level

 Outage probability idea can be used to find the

cell coverage area in cellular networks.

EE 728 METU AOY 37

 Example
900MHz,   3.71 , K=-31.54dB, varianceof log-normal shadowing 13.29

EE 728 METU AOY 38