Digital Communication

Unit No. 3
Random Processes

Dr. Suvarna S. Patil

Professor in E&TC Engg., BVCOEW, Pune-43
Mail id: sschorage@gmail.com
 Unit 3: Contents
• Introduction
• Mathematical definition of a random process.
• Stationary processes.
• Mean, Correlation and Covariance Functions
• Ergodic process
• Transmission of a random process through a LTI filter
• Power spectral density (PSD)
• Gaussian process
• Narrow band Noise
• In-phase and Quadrature components of noise

12/08/20 Random Processes 2

 References

• T1: Digital Communication Systems, by

Simon Haykin 4th Ed.

• R3: Modern Digital & Analog Communication

Systems, by B. P. Lathi 4th Ed.

12/08/20 Random Processes 3

 Unit 3: Objectives
• To understand random signals and random process.
• To define and calculate mean, correlation and covariance
functions of random process.
• To describe and define various types of random process
e.g. Ergodic and Gaussian process etc.
• To understand transmission of a random process through
a linear filter.
• To understand concept of narrow-band noise and study
its in-phase and quadrature components.

12/08/20 Random Processes 4

 Pre-requisites
• Signals and their classification.
• Concepts of deterministic and random processes
stationarity, ergodicity
• Basic properties of a single random process
mean, standard deviation, auto-correlation, spectral
density
• Joint properties of two or more random processes
correlation, covariance, cross spectral density, simple
input-output relations.
• LTI system
• Noise and its types.
12/08/20 Random Processes 5
 Importance of Random Processes
• Random variables and processes talk about
quantities and signals which are unknown in advance
• The data sent through a communication system is
modeled as random variable
• The noise, interference, and fading introduced by the
channel can all be modeled as random processes
• Even the measure of performance (Probability of Bit
Error) is expressed in terms of a probability

12/08/20 Random Processes 6

 Deterministic and random processes :
• both continuous functions of time (usually)

• Deterministic processes :
physical process is represented by explicit
mathematical relation
• Random processes :
result of a large number of separate causes.
Described in probabilistic terms and by properties
which are averages

12/08/20 Random Processes 7

 Random Processes: significance
• In practical problems, we deal with time varying
waveforms whose value at a time is random in nature.
Randomness or unpredictability is a fundamental property
of information.
For example,
• The speech waveform recorded by a microphone,
• The signal received by communication receiver
• The temperature of a certain city at noon or
• The daily record of stock-market data represents
random variables that change with time. 
• How do we characterize such data? Such data are
characterized as random or stochastic processes.
12/08/20 Random Processes 8
 Random Processes: significance contd.

• A random variable maps each sample point in the

sample space to a point in the real line.

• A random process maps each sample point to a

waveform.
•Thus a random process is a function of the sample
point  ‘s’ and index variable ’t’  and may be written as 
X(t,s)
for fixed t=t0 X(t0, s) is a random variable.

12/08/20 Random Processes 9

e.g. Temperature records for the day
X (t,  )
S1

S2

S3
X (t,  n )

S4 X (t,  k ) 

X (t,  2 ) 
X (t, 1 )
t
0 t1 t2
12/08/20 Random Processes 10
 Ensemble and sample function
• The collection of all possible waveform is
known as Ensemble. (sample space in random
variable)

• Each waveform in the collection is a sample

function (sample point)
• Amplitudes of all the sample functions at t=t0
is ensemble statistics.

12/08/20 Random Processes 11

 Classification of random processes

• Stationary and Nonstationary

• Wide-Sense or Weakly Stationary

• Ergodic

12/08/20 Random Processes 12

 Strictly Stationary random process :
• Ensemble averages do not vary with time
• The statistical characterization of the
process is time invariant.
• The PDFs obtained at any instants must be
identical.
• The Autocorrelation function must be
Rx (t1 , t2 )  Rx (t 2  t1 )

12/08/20 Random Processes 13

 Wide sense Stationary random process :

• Mean value is constant

• Autocorrelation function is independent of
the shifts
Rx (t1 , t2 )  Rx (t 2  t1 )  Rx ( )

• All stationary processes are wide-sense

stationary but converse is not true.
12/08/20 Random Processes 14
 Stationarity

12/08/20 Random Processes 15

 Ergodic random process :
• Ensemble averages are equal to time averages of
any sample function.
T 2 T 2
1 1
E   X (T )  x(t )  lim  x(t )dt  lim   X dt   X
T  T T  T
T 2 T 2

• stationary process in which averages from a single

record are the same as those obtained from
averaging over the ensemble
• Most stationary random processes can be treated
as Ergodic
12/08/20 Random Processes 16
Terminology Describing Random Processes
• A stationary random process has statistical
properties which do not change at all time
• A wide sense stationary (WSS) process has a
mean and autocorrelation function which do not
change with time
• A random process is Ergodic if the time average
always converges to the statistical average
• Unless specified, we will assume that all random
processes are WSS and Ergodic
12/08/20 Random Processes 17
 Ergodic

12/08/20 Random Processes 18

Mean, Correlation and covariance function
• Mean value :

x(t)

x

time, t T


 X (t )  E  X (t )   xf X (t ) ( x) dx


12/08/20 Random Processes 19

Mean, Correlation and covariance function
• Mean value of a stationary random
process is a constant.  (t )  
X X

• Autocorrelation function of a random

process X(t)is given as

RX (t1 , t 2 )  E  X (t1 ) X (t 2 )

12/08/20 Random Processes 20

12/08/20 Random Processes 21
12/08/20 Random Processes 22
 • Autocorrelation :
x(t)

time, t T

• The autocorrelation, or autocovariance,

describes the general dependency of x(t)
with its value at a short time later, x(t+)

12/08/20 Random Processes 23

 Autocorrelation properties

Symmetry

2. Power of W.S.S. process

3. Maximum value

12/08/20 Random Processes 24

Autocorrelation function of slowly and
rapidly fluctuating random processes

Correlation and decorrelation time concept

12/08/20 Random Processes 25
Mean, Correlation and covariance function
• Autocovariance function of a stationary
random process X(t)is given as

C X (t1 , t 2 )  E  ( X (t1 )   X )( X (t 2 )   X )
 RX (t 2  t1 )   2
X

12/08/20 Random Processes 26

 Probable Questions on this topic
• Numerical based on calculation of
expectation of a given random process.
• Properties of Autocorrelation
• Find mean, autocorrelation and covariance
of a given random process.

12/08/20 Random Processes 27

12/08/20 Random Processes 28
 _________________

   
 
Rxy  t1 , t2   x t y t    x1 y2 f xy  x1 , y 2  dx1dy 2
1 2  

12/08/20 Random Processes 29

 Jointly Stationary Properties
• Properties

• Uncorrelated:

• Orthogonal:

12/08/20 Random Processes 30

12/08/20 Random Processes 31
If τ=0 then cross-correlation results in zero i.e. the random variables
are orthogonal

12/08/20 Random Processes 32

12/08/20 Random Processes 33
 Spectral density

12/08/20 Random Processes 34

 Transmission of a random process through a
X(t) w.s.s Linear filter
Y(t) Y (t )  E Y (t )
 

I Impulse

 E   h( ) X (t   )d 
Response h(t)

  


random process 

The mean of the output  h( ) E[ X (t   )]d

Y(t) is given as 
 X 

h ( ) d   X H ( 0)
12/08/20 Random Processes 35
 Points to remember

• Mean of random process Y(t) produced at the

output of a LTI system in response to input
random process X(t) equals to the mean of
X(t) multiplied by the dc response of the
system.
• Autocorrelation function of the random
process Y(t) is a constant.

12/08/20 Random Processes 36

 Filtering of random signals

12/08/20 Random Processes 37

12/08/20 Random Processes 38
 Gaussian process
• A stochastic process is said to be a normal or Gaussian process if its
joint probability distribution is normal.

• A Gaussian process is strictly and weakly stationary because the

normal distribution is uniquely characterized by its first two
moments.

1  ( y  Y )2  1  y2 
fY ( y )  exp  exp 
2  Y  2 Y 
2
2  2

12/08/20 Random Processes 39

 • Central limit theorem
– The sum of a large number of independent and
identically distributed random variables getting
closer to Gaussian distribution
• Thermal noise can be closely modeled by Gaussian
process

• The theorem states that the probability distribution

of N random variables approaches a normalized
Gaussian distribution in the limit as N approaches
infinity.

12/08/20 Random Processes 40

 Gaussian Random Processes
• Gaussian random processes have some special properties

- If a Gaussian random process is wide-sense stationary,

then it is also stationary
- If the input to a linear system is a Gaussian random
process, then the output is also a Gaussian random
process
- if the random variable X(t1), X(t2),…..X(tn) obtained by
sampling a Gaussian process X(t) at times t1, t2, ……tn are
uncorrelated then these random variables are statistically
independent.

12/08/20 Random Processes 41

 Noise
• Unwanted waves that tend to disturb the
transmission and processing of signals in
communication systems.

Types of noise

• Shot noise as a random process X (t )   h(t  
k  
k )
current pulse generated at time instant

(t 0 ) k   t 0
Poisson distributi on  e
k!
Sample function of Poisson counting process and autocovaiance plot

12/08/20 Random Processes 42

 WHITE NOISE
• The primary spectral characteristic of thermal noise is that its power
spectral density is the same for all frequencies of interest in most
communication systems
• A thermal noise source emanates an equal amount of noise power per
unit bandwidth at all frequencies—from dc to about 1012 Hz.
• Power spectral density G(f)
• Autocorrelation function of white noise is

• The average power P of white noise if infinite

12/08/20 Random Processes 43

 WHITE NOISE

12/08/20 Random Processes 44

 Narrowband Noise Representation
• The noise process appearing at the output of a
narrowband filter is called narrowband noise.
• With the spectral components of narrowband noise
concentrated about some midband frequency  f c
• We find that a sample function n(t) of such a process
appears somewhat similar to a sine wave of frequency f c

• Representations of narrowband noise

– A pair of component called the in-phase and quadrature
components.
– Two other components called the envelop and phase.
 Representation of Narrowband Noise in Terms of In-Phase and
Quadrature Components
• Consider a narrowband noise n(t ) of bandwidth 2B centered on
frequency f c, as illustrated in Figure
• We may represent n(t) in the canonical (standard) form:
n(t )  nI (t ) cos(2f c t )  nQ (t ) sin(2f c t )
where, nI (t ) is in-phase component of n(t ) and nQ (t ) is quadrature
component of n(t. )
 PSD of narrow-band noise

12/08/20 Random Processes 47

Representation of Narrowband Noise in Terms
of Envelope and Phase Components
 Properties of narrow-band noise
• The in-phase and quadrature components of
narrow-band noise have zero mean.
• If the narrow-band noise is Gaussian then its in-
phase component and quqdrature component are
jointly Gaussian.
• If the narrow-band noise is wide sense stationary
(w.s.s.), then its in-phase component and
quqdrature component are jointly w.s.s.
• PSDs of these quadrature components are same as
that of PSD of original narrow-band noise.
• The quadrature components have same variance as
the narrow-band noise.
12/08/20 Random Processes 49
 Thank You

12/08/20 Random Processes 50