Lecture-4: Fading

1. Different types of fading.

2. Modeling of flat-fading channels.
3. Rayleigh distribution.
4. Ricean distribution.
Section-1: Different types of fading
1. Slow fading.
2. Fast fading.
3. Frequency- Flat fading.
4. Frequency- Selective fading.

Section-2: Modeling of flat-fading channels

1. Multipath fading:
(i) Rayleigh fading
(ii) Nakagami-m fading
(iii) Nakagami-n (Ricean) fading
(iv) Nakagami-q (Hoyt) fading
(v) Weibull fading
(vi) Beckmann fading
2. Log-normal shadowing.
3. Composite multipath/shadowing fading.
(i) Composite Gamma/Log-normal fading.
(ii) Composite Rayleigh/Log-normal fading.

Section-3: Rayleigh distribution

1. Characteristics of Rayleigh distribution
2. Probability density function (PDF) of Rayleigh distribution
3. Cumulative distribution function (CDF) of Rayleigh distribution
4. Moment generating function (MGF) of Rayleigh distribution

Section-4: Ricean distribution

1. Characteristics of Rayleigh distribution
2. Probability density function (PDF) of Rayleigh distribution
3. Cumulative distribution function (CDF) of Rayleigh distribution
4. Moment generating function (MGF) of Rayleigh distribution

1
 Lecture-4: Fading
(Section-1: Different types of fading)
1. Introduction
(a) Channel’s Coherence Time: The coherence time of the channel is defined as the period of time over
which the fading process of a channel is correlated. It is denoted by
TC.

(b) Symbol Time: The time duration which is required for a particular symbol to cross the distance
between transmitter and receiver is known as symbol time. It is denoted by TS.

(c) Channel’s Coherence Bandwidth: The coherence bandwidth of a channel is defined as the frequency
range over which the fading process of a channel is correlated. It is denoted by fC.

(d) Transmitted Signal Bandwidth: The bandwidth which is allocated for a particular transmitted signal
or symbol is known as transmitted signal bandwidth. It is denoted by fS.

2. Slow fading: The fading of a channel is said to be slow if the symbol time duration TS is smaller than the
channel’s coherence time TC.

In this case, different symbols experience the same fading. In the following figure, three
symbols experience the same fading.

TC = Channel’s Coherence Time

Symbol-1 Symbol-2 Symbol-3

TS-1 TS-2 TS-3

3. Fast fading: The fading of a channel is said to be fast if the symbol time duration TS is equal to or greater
than the channel’s coherence time TC.

In this case, all the bits of a single symbol experience the same fading. Different symbols
experience different fading. In the following figure, all the bits of a single symbol experience
the same fading.

TC = Channel’s Coherence Time

Symbol-1
TS-1

Sometime, the bits of a single symbol experience different fading levels. In the following figure,
some of the bits of symbol-1 experience a fading level, after that the fading level of the channel
changes and the rest of the bits of symbol-1 experience another fading level of the channel.

TC = Channel’s Coherence Time TC = Channel’s Coherence Time

Symbol-1
TS-1
2
4. Frequency-Flat fading: If all the frequency components of the transmitted signal are affected in a similar
manner, the fading is said to be frequency-flat fading.

This is the case of narrowband systems, in which the transmitted signal bandwidth, fS, is much smaller than the
channel’s coherence bandwidth, fC.

Constant gain + Constant phase shift

fC
fS

5. Frequency-Selective fading: If all the frequency components of the transmitted signal are affected by different
amplitude gains and phase shifts, the fading is said to be frequency-selective fading.

This is the case of wideband systems, in which the transmitted signal bandwidth, fS, is much bigger than the
channel’s coherence bandwidth, fC.

3
 Lecture-4: Fading
(Section-2: Modeling of flat-fading channels)
1. Introduction
(a) Modeling of flat fading channel includes the following measures;
(i) Fading Amplitude of the channel
(ii) Instantaneous signal-to-noise ratio (SNR) of the channel
(iii) Probability density function (PDF) of the channel
(iv) Cumulative distribution function (CDF) of the channel
(v) Moment generating function (MGF) of the channel
(vi) Amount of Fading in the channel

(b) Analysis of flat fading channel requires the following fundamental parameters;
(i) Probability density function (PDF) of the channel
(ii) Cumulative distribution function (CDF) of the channel
(iii) Moment generating function (MGF) of the channel
(iv) Amount of Fading in the channel

2. General description of the measures to model the flat fading channel

(a) Fading amplitude of the channel:

Let,
(i) α = Fading amplitude of the channel.

(ii)    2 = Mean-square value of α.

(iii) f   = Probability density function (PDF) of α, which is dependent on the nature of the radio
propagation environment.

Roll of Fading Amplitude: When fading affects the transmitted signal in a narrowband system, the received
carrier amplitude is modulated by the fading amplitude α.

Here α is modeled as a random variable with the mean-square value    2 and the probability density function
f    .

(b) Instantaneous signal-to-noise ratio (SNR) of the channel:

Let,
(i) Es = Energy per symbol.

(ii) N 0 = Noise power.

Es 2
(iii)    = Instantaneous signal-to-noise ratio (SNR) of the channel.
N0

Es
(iv)    = Average SNR of the channel.
N0
Instantaneous SNR: After passing through fading channel, signal is perturbed at the receiver by the additive
white Gaussian noise (AWGN), which is statistically independent of the fading amplitude α and which is
characterized by the power spectral density N0. Equivalently, the received instantaneous signal power is
modulated by α2.

4
Hence, the instantaneous received signal-to-noise ratio (SNR) of the channel per symbol is then defined by
Es 2 E
  and the average SNR of the channel per symbol is defined by   s  . Also, the PDF of  is
N0 N0
obtained by

  
f  
  
f    

2


(c) Probability density function (PDF) of the channel:

The PDF of channel is equivalently the PDF of  which is derived from the PDF of α using the following
expression

  
f  
  
f    

2


(d) Cumulative distribution function (CDF) of the channel :

The CDF of channel is equivalently the CDF of  which is derived from the PDF of  using the
following expression

F      f   d 
0

(e) Moment generating function (MGF) of the channel:

The MGF of channel is equivalently the MGF of  which is derived from the PDF of  using the
following expression

M   s    f    e s d 
0

(f) Amount of Fading in the channel:

The amount of fading in the channel is derived from the MGF of  using the following expression
2
d 2 M  (s)  dM  ( s ) 
2   
E[ 2 ]   E  
2
ds s 0  ds s  0 
AF  2
 2
 E     dM  ( s ) 
 
 ds s  0 

5
 Section-2: Modeling of flat-fading channels

1. Modeling of Rayleigh fading channel

(i) α = Fading amplitude of the Rayleigh fading channel.
(ii)    2 = Mean-square value of α of the Rayleigh fading channel.
(iii) f   = The PDF of α of the Rayleigh fading channel (Derived).

  
f  
  
(iv) f     = The PDF of the instantaneous SNR of the Rayleigh fading channel,  .

2

(Derived from the expression of f   ).


(v) F      f   d  = The CDF of the instantaneous SNR of the Rayleigh fading channel,  .
0

(Derived from the expression of f    )


s
(vi) M   s    f   e d  = The MGF of the instantaneous SNR of the Rayleigh fading channel,  .
0

(Derived from the expression of f    )

2
d 2 M  (s)  dM  ( s ) 
2   
E[ 2 ]   E   ds 2 s  0  ds s  0 
(vii) AF  2
 2
= The amount of fading in the Rayleigh
 E    dM  ( s ) 
 
 ds s  0 
fading channel. (Derived from the expression of M  ( s ) )

6
2. Modeling of Nakagami-m fading channel
(i) α = Fading amplitude of the Nakagami-m fading channel.
(ii)    2 = Mean-square value of α of the Nakagami-m fading channel.
(iii) f   = The PDF of α of the Nakagami-m fading channel (Derived).

  
f  
  
(iv) f     = The PDF of the instantaneous SNR of the Nakagami-m fading channel,  .

2

(Derived from the expression of f   ).


(v) F      f   d  = The CDF of the instantaneous SNR of the Nakagami-m fading channel,  .
0

(Derived from the expression of f    )


s
(vi) M   s    f  e d  = The MGF of the instantaneous SNR of the Nakagami-m fading channel,  .
0

(Derived from the expression of f    )

2
d 2 M  (s)  dM  ( s ) 
2  
E[ 2 ]   E   ds 2  ds
(vii) AF   s 0  s0 
= The amount of fading in the Nakagami-m
2 2
 E     dM  ( s ) 
 
 ds s  0 
fading channel. (Derived from the expression of M  ( s ) )

7
 3. Modeling of Nakagami-n (Ricean) fading channel
(i) α = Fading amplitude of the Nakagami-n (Ricean) fading channel.
(ii)    2 = Mean-square value of α of the Nakagami-n (Ricean) fading channel.
(iii) f   = The PDF of α of the Nakagami-n (Ricean) fading channel (Derived).

  
f  
  
(iv) f     = The PDF of the instantaneous SNR of the Nakagami-n (Ricean) fading channel,

2

 . (Derived from the expression of f   ).


(v) F      f   d  = The CDF of the instantaneous SNR of the Nakagami-n (Ricean) fading channel,
0

 . (Derived from the expression of f   )


s
(vi) M   s    f  e d  = The MGF of the instantaneous SNR of the Nakagami-n (Ricean) fading
0

channel,  . (Derived from the expression of f    )

2
d 2 M  (s)  dM  ( s ) 
2   
E[ 2 ]   E   ds 2 s 0  ds s  0 
(vii) AF  2
 2
= The amount of fading in the Nakagami-n
 E     dM  ( s ) 
 
 ds s  0 
(Ricean) fading channel. (Derived from the expression of M  ( s ) )

8
4. Modeling of Nakagami-q (Hoyt) fading channel
(i) α = Fading amplitude of the Nakagami-q (Hoyt) fading channel.
(ii)    2 = Mean-square value of α of the Nakagami-q (Hoyt) fading channel.
(iii) f   = The PDF of α of the Nakagami-q (Hoyt) fading channel (Derived).

  
f  
  
(iv) f     = The PDF of the instantaneous SNR of the Nakagami-q (Hoyt) fading channel,  .

2

(Derived from the expression of f   ).


(v) F      f   d  = The CDF of the instantaneous SNR of the Nakagami-q (Hoyt) fading channel,  .
0

(Derived from the expression of f    )


s
(vi) M   s    f  e d  = The MGF of the instantaneous SNR of the Nakagami-q (Hoyt) fading
0

channel,  . (Derived from the expression of f    )

2
d 2 M  (s)  dM  ( s ) 
2   
E[ 2 ]   E   ds 2 s 0  ds s  0 
(vii) AF  2
 2
= The amount of fading in the Nakagami-q
 E     dM  ( s ) 
 
 ds s  0 
(Hoyt) fading channel. (Derived from the expression of M  ( s ) )

9
5. Modeling of Weibull fading channel
(i) α = Fading amplitude of the Weibull fading channel.
(ii)    2 = Mean-square value of α of the Weibull fading channel.
(iii) f   = The PDF of α of the Weibull fading channel (Derived).

  
f  
  
(iv) f     = The PDF of the instantaneous SNR of the Weibull fading channel,  .

2

(Derived from the expression of f   ).


(v) F      f   d  = The CDF of the instantaneous SNR of the Weibull fading channel,  .
0

(Derived from the expression of f    )


s
(vi) M   s    f  e d  = The MGF of the instantaneous SNR of the Weibull fading channel,  .
0

(Derived from the expression of f    )

2
d 2 M  (s)
 dM  ( s ) 
2   
E[ 2 ]   E   ds 2
s 0  ds s  0 
(vii) AF  2
 2
= The amount of fading in the Weibull fading
 E     dM  ( s ) 
 
 ds s  0 
channel. (Derived from the expression of M  ( s ) )

10
6. Modeling of Beckmann fading channel
(i) α = Fading amplitude of the Beckmann fading channel.
(ii)    2 = Mean-square value of α of the Beckmann fading channel.
(iii) f   = The PDF of α of the Beckmann fading channel (Derived).

  
f  
  
(iv) f     = The PDF of the instantaneous SNR of the Beckmann fading channel,  .

2

(Derived from the expression of f   ).


(v) F      f   d  = The CDF of the instantaneous SNR of the Beckmann fading channel,  .
0

(Derived from the expression of f    )


s
(vi) M   s    f  e d  = The MGF of the instantaneous SNR of the Beckmann fading channel,  .
0

(Derived from the expression of f    )

2
d 2 M  (s)
 dM  ( s ) 
2   
E[ 2 ]   E   ds 2
s 0  ds s  0 
(vii) AF  2
 2
= The amount of fading in the Beckmann
 E     dM  ( s ) 
 
 ds s  0 
fading channel. (Derived from the expression of M  ( s ) )

11
7. Modeling of Log-normal fading channel
(i) α = Fading amplitude of the Log-normal fading channel.
(ii)    2 = Mean-square value of α of the Log-normal fading channel.
(iii) f   = The PDF of α of the Log-normal fading channel (Derived).

  
f  
  
(iv) f     = The PDF of the instantaneous SNR of the Log-normal fading channel,  .

2

(Derived from the expression of f   ).


(v) F      f   d  = The CDF of the instantaneous SNR of the Log-normal fading channel,  .
0

(Derived from the expression of f    )


s
(vi) M   s    f  e d  = The MGF of the instantaneous SNR of the Log-normal fading channel,  .
0

(Derived from the expression of f    )

2
d 2 M  (s)
 dM  ( s ) 
2   
E[ 2 ]   E   ds 2
s 0  ds s  0 
(vii) AF  2
 2
= The amount of fading in the Log-normal
 E     dM  ( s ) 
 
 ds s  0 
fading channel. (Derived from the expression of M  ( s ) )

12
8. Modeling of Composite Gamma/Log-normal fading channel
(i) α = Fading amplitude of the Composite Gamma/Log-normal fading channel.
(ii)    2 = Mean-square value of α of the Composite Gamma/Log-normal fading channel.
(iii) f   = The PDF of α of the Composite Gamma/Log-normal fading channel (Derived).

  
f  
  
(iv) f     = The PDF of the instantaneous SNR of the Composite Gamma/Log-normal fading

2

channel,  . (Derived from the expression of f   ).


(v) F      f   d  = The CDF of the instantaneous SNR of the Composite Gamma/Log-normal fading
0

channel,  . (Derived from the expression of f    )


s
(vi) M   s    f  e d  = The MGF of the instantaneous SNR of the Composite Gamma/Log-normal
0

fading channel,  . (Derived from the expression of f    )

2
d 2 M  (s)  dM  ( s ) 
2   
E[ 2 ]   E   ds 2 s 0  ds s  0 
(vii) AF  2
 2
= The amount of fading in the Composite
 E     dM  ( s ) 
 
 ds s  0 
Gamma/Log-normal fading channel. (Derived from the expression of M  ( s ) )

13
9. Modeling of Composite Rayleigh/Log-normal fading channel
(i) α = Fading amplitude of the Composite Rayleigh/Log-normal fading channel.
(ii)    2 = Mean-square value of α of the Composite Rayleigh/Log-normal fading channel.
(iii) f   = The PDF of α of the Composite Rayleigh/Log-normal fading channel (Derived).

  
f  
  
(iv) f     = The PDF of the instantaneous SNR of the Composite Rayleigh/Log-normal

2

fading channel,  . (Derived from the expression of f   ).


(v) F      f   d  = The CDF of the instantaneous SNR of the Composite Rayleigh/Log-normal
0

fading channel,  . (Derived from the expression of f    )


s
(vi) M   s    f  e d  = The MGF of the instantaneous SNR of the Composite Rayleigh/Log-normal
0

fading channel,  . (Derived from the expression of f    )

2
d 2 M  (s)  dM  ( s ) 
2   
E[ 2 ]   E   ds 2 s 0  ds s  0 
(vii) AF  2
 2
= The amount of fading in the Composite
 E     dM  ( s ) 
 
 ds s  0 
Rayleigh/Log-normal fading channel. (Derived from the expression of M  ( s ) )

14
 Lecture-4: Fading
(Section-3: Rayleigh distribution)
1. Characteristics of Rayleigh distribution
(i) It has no direct line-of-sight (LOS) path between transmitter and receiver.
(ii) It supports the experimental data for mobile communication where no LOS path exists between transmitter and
receiver.
(iii) It is used to explain the propagation of reflected and refracted paths through the troposphere and ionosphere.

2. Probability density function (PDF) of Rayleigh distribution

Let P 2,
 h
N
where  = Instantaneous signal-to-noise ratio (SNR) per symbol of the Rayleigh fading channel.
P = Transmit signal power.
N = Noise power.
2
h = Channel gain of the Rayleigh fading SISO channel.

Then, the probability density function (PDF) of  is given by,


1 

f ( )  e ,  0 (1)

2
where   E  h  = Average SNR.
 

3. Cumulative distribution function (CDF) of Rayleigh distribution

If denotes the instantaneous SNR per symbol of the Rayleigh fading channel. Then the cumulative distribution
function of  is defined by

  
1 
F ( )   f ( )d    e  d 
0 0 
 
1    1   
     e        e  
   0    0



 1 e (2)

4. Moment generating function (MGF) of Rayleigh distribution

If  denotes the instantaneous SNR per symbol of the Rayleigh fading channel. Then the moment generating
function (MGF) of  is defined by

    1    1 s  
1  1    s 
1   
s  s     
M  ( s )   f ( )  e d    e  e d   e d  e d
0 0   0  0


1    1 s    1 1
   e     
 1  s   1  s
 1  s   (3)
 0

15
 Lecture-4: Fading
(Section-4: Ricean distribution)
1. Characteristics of Nakagami-n (Ricean) distribution
(i) It is used to model propagation paths consisting of one strong direct LOS component and many random
weaker components.
(ii) Practically, Ricean distribution is observed in the microcellular urban and suburban land-mobile, picocellular
indoor and factory environments.

2. Probability density function (PDF) of Ricean distribution

Let P 2,
 h
N
where  = Instantaneous SNR per symbol of the Ricean fading channel.
P = Transmit signal power.
N = Noise power.
2
h = Channel gain of the Ricean fading SISO channel.

Then, the probability density function (PDF) of  is given by

1  n  e 2  n2
 1  n 2     1  n    ,
2

f ( )  exp    I 0  2n  0 (1)
      
   
2
where n = Ricean fading parameter and   E  h  = Average SNR.
 

3. Cumulative distribution function (CDF) of Ricean distribution

If denotes the instantaneous SNR per symbol of the Ricean fading channel. Then the cumulative distribution
function of  is defined by

 1  n 2 e  n2

F ( )   f ( ) d     
  1  n 2  
exp  
 
 I 0  2n
1  n   d 
2

(2)
       
0 0
    

4. Moment generating function (MGF) of Ricean distribution

If  denotes the instantaneous SNR per symbol of the Ricean fading channel. Then the moment generating
function (MGF) of  is defined by

 
1  n  e
2  n2
 1  n 2     1  n     e
2

M  ( s )   f ( )  e d    s
exp    I 0  2n s
d
      
0 0    


1  n  exp  n s 
2 2

(3)
1  n   s  1  n   s 
2 2

16
References:
1. Simon Haykin and Michael Moher, “Modern Wireless Communications”, 1st Edition.
2. Marvin K. Simon and Mohamed-Selim Alouni, “Digital Communication over Fading Channels”, 2nd Edition.
3. John G. Proakis, “Digital Communication”, 4th Edition.
4. Bernard Sklar, “Digital Communication-Fundamentals and Applications”, 2nd Edition.
5. Leon W. Couch, “Digital and Analog Communication Systems”, 5th Edition.
6. B. P. Lathi, “Modern Digital and Analog Communication Systems”, 3rd Edition.
7. Edward A. Lee and David G. Messerschmitt, “Digital Communication”, 2nd Edition.
8. Rudiger Urbanke, “Advanced Digital Communications”, 2nd Edition.
9. Michael P. Fitz, “An Introduction to Digital Communication Theory”, 2nd Edition.
10. Upamanyu Madhow, “Fundamentals of Digital Communication”, 3rd Edition.
11. Andrew J. Viterbi and Jim K. Omura, “Principles of Digital Communication and Coding”, 3rd Edition.

17