Probability and Random Process

ECEg2110

Chapter 4
Random Process
 Outline
4.1. Introduction to Random Process
4.2. Characterization of Random Processes
4.3. Classification of Random Processes
4.4. Power Spectral Densities of Random Processes
4.5. Response of Linear Systems to Random Inputs

AASTU,ECE, 2024 2
 Introduction
• So far, we have seen random experiments whose outputs are independent of
time /other parameter.
• But in certain random experiments, the outcome is a function of time or
space. For example,
• In speech recognition systems, decisions are made on the basis of a voltage
waveform corresponding to a speech utterance. [f(t)]
• In an image processing, the intensity of the image varies over a rectangular
region. [f(area)]
• In queuing systems, the number of customers in a system varies with t.
• In some situations; 2 or more functions of time may be of our interest.
• Eg: the temperature in a certain city & the demand placed on the local electric
power utility vary together in time.

AASTU,ECE, 2024 3
 Definition of a Random Process
✓ In general, the random time functions in all of the above examples can be
viewed as numerical quantities that evolve randomly in time or space.
✓ Thus, we have a family of random variables indexed by the time or
space variable: Random Process.
✓ A random process (rp) or stochastic process is the mathematical model
of an empirical process whose development is governed by probability
laws.
✓ A stochastic process is a mathematical model of a probabilistic
experiment that evolves in time and generates a sequence of numerical
values.
✓ Each numerical value in the sequence is modeled by a random variable, so a
stochastic process is simply a (finite or infinite) sequence of random variables.
AASTU,ECE, 2024 4
 Definition of a Random Process
• Consider a random experiment specified by
the outcomes ω from sample space Ω by the
events defined on Ω & by the probabilities on
these events.
• Suppose that to every outcome ωϵΩ , we
assign a function of time according to some
rule: 𝑋(𝑡, 𝜔) , 𝒕 ∈ 𝑻
• The set of functions {𝑋(𝑡, 𝜔1), 𝑋(𝑡, 𝜔2), … , 𝑋(𝑡, 𝜔𝑁 )}
corresponding to the N outcomes of an
experiment is called an ensemble.
• Graph of the function 𝑋(𝑡, 𝜔) versus 𝑡 , for a
fixed ω is called a realization, sample path
or sample function of the random
process.
AASTU,ECE, 2024 5
 Definition of a Random Process
• For a fixed time 𝑡 = 𝑡𝑘 , 𝑋(𝑡𝑘, 𝜔) = 𝑋𝑘 (𝜔) is a random variable, which
results in indexed family (ensemble) of random variables by the
parameter 𝑡.
• This family is called a random process or stochastic process.
• Thus, random process is a family/collection of sample functions.
• We will usually leave the ω and use 𝑋 𝑡 to denote random process (rp).
• For generality, the parameter ‘t’ above can be time, distance, angle, etc. But here,
we will see considering it time.
• NB: The randomness in ω induces randomness in the observed function 𝑋 𝑡, 𝜔 & one
can deduce the probabilities of the events involving a stochastic process at various
instants of time from probabilities involving ω by using the equivalent method.
AASTU,ECE, 2024 6
 Definition of a Random Process
•

AASTU,ECE, 2024 7
 Definition of a Random Process
The concept of random variable can be extended to include time and the outcome
will be random functions of time as shown below.

AASTU,ECE, 2024 88
 Definition of a Random Process
X(t) the random process
(Ensemble)

Let x(t) be a sample of

x(t) a sample of the
the random process X(t) random Process; one
were the lower case letter realization
imply a sample function
(not random function).

AASTU,ECE, 2024 9
 Example
▪ Starting at launch time t=0. Let X(t) denote the temperature in degrees
Kelvin on the surface of a space shuttle. With each launch, we record a
temperature sequence x(t,s).
▪ For example, x(8073.68, 2)=207, indicates that the temperature is 207K at
8073.68 seconds during the second launch. X(t) is a stochastic process.

AASTU,ECE, 2024 10
 Example
▪ Suppose that at time instants T = 0, 1, 2, 3, …, we roll a die and record the
outcome NT where 1 ≤ NT ≤ 6.

▪ We then define the random process X(t) such that for T ≤ t ≤ T+1, X(t) = NT.

▪ In this case the experiment consists of an infinite sequence of rolls and a sample
function is just the waveform corresponding to a particular sequence of rolls.

AASTU,ECE, 2024 11
Types of Random Process

AASTU,ECE, 2024 12
 Types of Random Process
Deterministic Vs Non-Deterministic
• A stochastic process is deterministic if a sample function can be described by
a mathematical function such that its future values can be computed. The
randomness is in the ensemble, not in the time functions.
• For example, let the sample functions be of the form,
𝑋 𝑡 = 𝐴cos(2𝜋 𝑓0𝑡 + 𝜃) and let the parameter 𝜃 be random over the
ensemble but constant for any particular sample function.
• All other stochastic processes are non-deterministic.
• Future values cannot be determined from current ones.

AASTU,ECE, 2024 13
 Types of Random Process
Deterministic Vs Non-Deterministic

AASTU,ECE, 2024 14
 Types …
Parameters Nature Notation
t ω
random process (random
Varying Varying 𝑋(𝑡, 𝜔)
signal)
sample function/realization
Varying Fixed 𝑋𝑖 𝑡 , 𝑥𝑖 (𝑡)
(deterministic signal)
Fixed Varying random variable 𝑋 𝑡𝑘 , 𝜔 = 𝑋(𝜔)
fixed Fixed scalar (deterministic number) c
Deterministic signal>>can be reproduced exactly with repeated measurement.

Random signal>>that are not repeatable in a predictable manner.

AASTU,ECE, 2024 1515

 Applications of Random processes
✓developed in connection with the study of fluctuations and noise in
physical systems.

✓are the mathematical model of an empirical process whose

development is governed by probability laws.

✓Thus, provides useful models for the studies of such diverse fields as
statistical physics, communication and control, time series
analysis, population growth, and management sciences.

AASTU,ECE, 2024 16
 Examples of random processes

Shot noise

Stock market samples

over 3 years
Popcorn noise

AASTU,ECE, 2024 17
 Examples of random processes

Random Settings for a

DC Power Supply Random Settings for an AC Power Supply

AASTU,ECE, 2024 18
 Examples of random processes
•

pdf's for X(t) at different times fX(t)(x) for the Bernoulli Process
AASTU,ECE, 2024 19
 Characterization of Random Processes
▪ If X(t) is a random process, then for fixed t=t1, X1=X(t1) represents a
random variable.
▪ Its cdf is given by:
FX ( x1 , t1 ) = P{X (t1 )  x1}

▪ Notice that FX(x, t) depends on t, since for a different t, we obtain a

different random variable.
▪ The first-order pdf of the process X(t) is defined as:

dFX ( x1 , t1 )
f X ( x1 , t1 ) =
dx1

AASTU,ECE, 2024 20
 Characterization of Random Processes
▪ For t = t1 and t = t2, X(t) represents two different random variables X1
= X(t1) and X2 = X(t2) respectively.
▪ Their cdf is given by:
FX ( x1 , x2 , t1 , t2 ) = P{ X (t1 )  x1 , X (t2 )  x2 }
▪The second-order pdf of the random process X(t) is:
 2 FX ( x1 , x2 , t1 , t2 )
f X ( x1 , x2 , t1 , t2 ) =
x1 x2
▪ Similarly f ( x1 , x2 , xn , t1 , t2 , tn )
X represents the nth order density function
of the process X(t).

AASTU,ECE, 2024 21
 Mean, Correlation, and Covariance Functions
▪ As in the case of random variables, random processes are often described by using
statistical/ ensemble averages.
▪ The mean of X(t) is defined by:
 X (t ) = EX (t )
where X(t) is treated as a random variable for a fixed value of t.
The time average of a quantity is defined as

1 T Here A[ ] is used to denote time

A  = lim    dt average in a manner analogous to
T →  2T − T
E[ ] for the statistical average.

The time average is taken over all time because, as applied to random
processes, sample functions of processes are presumed to exist for all time.

AASTU,ECE, 2024 22
 Mean, Correlation, and Covariance Functions
Autocorrelation function >>A measure of dependence among the random
variables of X(t) & is defined by:
RXX (t1 , t2 ) = E  X (t1 ) X (t2 ) 
▪ Note that:
R XX (t1 , t 2 ) = R XX (t 2 , t1 ) and R XX (t , t ) = E X 2
(t ) 
The autocovariance function of X(t) is defined by:

 XX (t1 , t 2 ) = CovX (t1 ) , X (t 2 ) = EX (t1 ) −  X (t1 )X (t 2 ) −  X (t 2 )

= RXX (t1 , t 2 ) −  X (t1 )  X (t 2 )

AASTU,ECE, 2024 23
 Mean, Correlation, and Covariance Functions
▪ It is clear that if the mean of X(t) is zero, then:

 XX (t1 , t2 ) = RXX (t1 , t2 )

▪ Note that the variance of X(t) is given by:


 X 2 (t ) = VarX (t ) = E X (t ) −  X (t )2 
▪ If X(t) is a complex random process, then its autocorrelation function RXX(t1, t2)
and autocovariance function CXX(t1, t2) are defined, respectively, by:

 
RXX (t1 , t 2 ) = E X (t1 ) X * (t 2 ) and

 XX (t1 , t2 ) = E X (t1 ) −  X (t1 )X (t2 ) −  X (t2 )
*

AASTU,ECE, 2024 24
 Classification of Random Processes
i. Stationary Processes
▪ A random process {X(t), tϵT} is said to be stationary or strict-sense
stationary (SSS) if, for all n and for every set of time instants (ti ϵT, i = 1,2, .
. . , n),
FX ( x1 ,........, xn , t1 , ....., t n ) = FX ( x1 , ........, xn , t1 +  , ......, t n +  )

▪ Hence, the distribution of a stationary process will be unaffected by a shift in the

time origin, and X(t) and X(t+τ) will have the same distributions for any τ.
▪ Nonstationary processes are characterized by distributions depending on the
points t1, t2, . . . , tn.

AASTU,ECE, 2024 25
 Classification of Random Processes
ii. Wide-Sense Stationary Processes
▪ A random process X(t) is wide-sense stationary (WSS) if:

1. E  X (t ) =  X (constant)

2. R XX (t1 , t 2 ) = E  X (t1 ) X (t 2 ) = R XX ( t 2 − t1 )

▪ Note that a strict-sense stationary process is also a WSS process, but,

in general, the converse is not true.

AASTU,ECE, 2024 26
 Classification of Random Process
▪ Other types of random processes include:
✓ Independent Processes
✓ Markov Processes
✓ Normal Processes
✓ Ergodic Processes
✓ Poisson Processes

AASTU,ECE, 2024 27
Power Spectral Densities of Random Processes
▪ The autocorrelation function of a continuous-time random process X(t) is defined
as:

RXX ( ) = EX (t ) X (t +  )
▪ Properties of RXX(τ):

1. R XX (− ) = R XX ( )

2. R XX ( )  R XX (0)


3. R XX (0) = E X (t )  0
2

AASTU,ECE, 2024 28
Power Spectral Densities of Random Processes
▪ In case of a discrete-time random process X(n), the autocorrelation function of
X(n) is defined by:

RXX (k ) = EX (n) X (n + k )

▪ Properties of RX(k):

1. R XX (− k ) = R XX (k )

2. R XX (k )  R XX (0)


3. R XX (0) = E X 2 (n)  0 
AASTU,ECE, 2024 29
Power Spectral Densities of Random Processes
▪ Two processes X(t) and Y(t) are called (mutually) orthogonal if:

RXY ( ) = 0 , for all 

▪ Similarly, the cross-correlation function of two discrete-time jointly WSS
random processes X(n) and Y(n) is defined by:

RXY (k ) = EX (n)Y (n + k )

▪ The various properties of RXY(k) similar to those of RXY(τ) can be obtained by
replacing τ by k in the above equations

AASTU,ECE, 2024 30
Power Spectral Densities of Random Processes
▪ The power spectral density (or power spectrum) SXX(ω) of a continuous-time
random process X(t) is defined as the Fourier transform of RXX(τ), i.e. ,

S XX ( ) =  R XX ( )e − j  d
−

▪ Thus, taking the inverse Fourier transform of SX(ω), we obtain:


R XX ( ) =
1
2 
−
S XX ( )e j  d

▪ The above equations are known as the Wiener-Khinchin relations.

AASTU,ECE, 2024 31
Power Spectral Densities of Random Processes
▪ Properties of SXX(ω):

1. S XX ( ) is real and S XX ( )  0

2. S XX (− ) = S XX ( )

 
3. E X (t ) = R XX (0) =
2 1
2 −

S XX ( )d

▪ Similarly, the power spectral density SXX(Ω) of a discrete-time random process

X(n) is defined as the Fourier transform of RXX(k):

S XX ( ) =  XX
R ( k ) e − j k

k = −

AASTU,ECE, 2024 32
Power Spectral Densities of Random Processes
▪ Thus, taking the inverse Fourier transform of SXX(Ω), we obtain:
1 
R XX (k ) =
2 
−
S XX ()e jk d

▪ Properties of SXX(Ω):

1. S XX ( + 2 ) = S XX ()

2. S XX () is real and S XX ()  0

3. S XX (−) = S XX ()

 
3. E X (n) = R XX (0) =
2 1
2

S
−
XX (  ) d

AASTU,ECE, 2024 33
Power Spectral Densities of Random Processes
▪ The cross power spectral density (or cross power spectrum) SXY(ω) of two
continuous-time random processes X(t) and Y(t) is defined as the Fourier
transform of RXY(τ):

S XY ( ) =  R XY ( )e − j 
d
−

▪ Thus, taking the inverse Fourier transform of SXY(ω), we get:


R XY ( ) =
1

j 
S XY ( )e d
2 −

AASTU,ECE, 2024 34
Power Spectral Densities of Random Processes
▪ Properties of SXY(ω):
▪ Unlike SXX(ω), which is a real-valued function of ω, SXY(ω), in general, is a
complex-valued function.
1. S XY ( ) = S YX ( − )

2. S XY ( − ) = S XY ( )
*

▪ Similarly, the cross power spectral density SXY(Ω) of two discrete-time random
processes X(n) and Y(n) is defined:

S XY () =  XY
R (
k = −
k ) e − j k

AASTU,ECE, 2024 35
Power Spectral Densities of Random Processes
▪ Taking the inverse Fourier transform of SXY(Ω), we get:
1 
S
jk
R XY (k ) = ()e d
2
XY
−

▪ Properties of SXY(ω):
▪ Unlike SXX(Ω), which is a real-valued function of Ω, SXY(Ω), in general, is a
complex-valued function.
1. S XY ( + 2 ) = S XY ()

2. S XY () = S YX (−)

3. S XY (−) = S XY ()
*

AASTU,ECE, 2024 36
 Example
Consider a random process X(t) defined by
X (t ) = A cos( 0 t +  )

where  0 and A are constants and  is a uniform

random variable over the interval (0, 2 )

a. Find the mean  X (t ).

b. Find the autocorrelation function R XX (t1 , t 2 ).

c. Find the autocovariance function C XX (t1 , t 2 ).

d . Determine whether X (t ) is WSS random process or not.

e. Find the power spectral density of X (t ).

AASTU,ECE, 2024 37
 Example
Solution:
a.  X (t ) = E  X (t ) = E  A cos( 0 t +  ) = AE cos( 0 t +  )

But, cos( 0 t +  ) = cos( 0 t)cos - sin( 0 t)sin

  X (t ) = E  X (t ) = AE cos( 0 t)cos - sin( 0 t)sin 

= A cos( 0 t) E cos  − A sin( 0 t) E sin  

2
E cos  =
1
2 0
cosd = 0

2
Similarly , E sin   =
1
2 0
sin d = 0

  X (t ) = E  X (t ) = 0

AASTU,ECE, 2024 38
 Example
Solution:
b. R XX (t1 , t 2 ) = E  X (t1 ) X (t 2 )

= E  A cos( 0 t1 +  ) A cos( 0 t 2 +  )

= A 2 E cos( 0 t1 +  ) cos( 0 t 2 +  )

A2
= E cos 0 (t 2 − t1 ) + cos( 0 (t1 + t 2 ) + 2 )
2
But , E cos 0 (t 2 − t1 ) = cos 0 (t 2 − t1 ) and

E cos( 0 (t1 + t 2 ) + 2 ) = 0

A2
 R XX (t1 , t 2 ) = cos 0 (t 2 − t1 )
2

AASTU,ECE, 2024 39
 Example
Solution:

c. C XX (t1 , t 2 ) = R XX (t1 , t 2 ) −  X (t1 )  X (t 2 )

A2
= cos 0 (t 2 − t1 ) − 0
2
A2
 C XX (t1 , t 2 ) = cos 0 (t 2 − t1 )
2
d . Since the mean is constant and the autocorrelation function

depends on time differenceonly, X (t ) is a WSS random process.

AASTU,ECE, 2024 40
 Example
Solution:
e. Since X (t ) is a WSS random process,the autocorrelation

function can be simply written as :

A2
R XX ( ) = cos( 0 )
2
The power spectral density of X (t ) is given by :

S XX ( ) =  R XX ( )e − j  d
−

But from Fourier transform pair table, we have :

FT cos( 0 t ) =  ( −  0 ) +  ( +  0 )

A 2 A 2
 S XX ( ) =  ( −  0 ) +  ( +  0 )
2 2
AASTU,ECE, 2024 41
 Response of Linear Systems to Random Inputs
▪ If a WSS random process X(t) with autocorrelation function RXX(τ) is applied to
a linear system with impulse response h(t), then the cross correlation function
RXY(τ) and the output autocorrelation function RYY(τ) are given as follows.

X(t) h(t) Y(t)

R XY ( ) = R XX ( ) * h * (− )

And ,

RYY ( ) = R XY ( ) * h( )

= R XX ( ) * h * (− ) * h( )

AASTU,ECE, 2024 42
 Response of Linear Systems to Random Inputs
▪ Using properties of Fourier transform, we get:
f (t ) ⎯→
FT
F ( ) and g (t ) ⎯→
FT
G ( )

 f (t ) * g (t ) ⎯→
FT
F ( )G ( )
▪ Then using the above property, the cross and output power spectral densities can be
evaluated as:

 
S XY ( ) = FT R XX ( ) * h * (− ) = S XX ( ) H * ( )

And ,

S YY ( ) = FT RYY ( ) = FT R XY ( ) * h( )

= S XY ( ) H ( ) = S XX ( ) H ( )
2

AASTU,ECE, 2024 43
 Example
Consider a WSS random process X(t) with autocorrelation
function given by:
−a 
R XX ( ) = e , where a is a real positive constant

Let the random process X(t) be applied to the input of an LTI system with
impulse response given by:
h(t ) = e −bt u (t ) , where b is a real positive constant

Find the autocorrelation function of the output Y(t) of the

system.

AASTU,ECE, 2024 44
 Example
Solution:
The frequency response H ( ) of the system is :

H ( ) = FT h(t ) =
1
j + b
The power spectral density of X (t ) is :

S XX ( ) = FT R XX ( ) = 2
2a
 + a2
Then, the power spectral density of Y (t ) is given by :

 1  2a 
S YY ( ) = S XX ( ) H ( ) =  2
2
2  2 
  + b   + a 
2

AASTU,ECE, 2024 45
 Example
Solution:
a  2b  b  2a 
 S YY ( ) = 2  2 2 
− 2  2 2 
( )
a −b b  + b  a −b b  + b 
2 2
( )
Taking the inverse Fourier transform of both sides of the above

equation, we obtain :

RYY ( ) = 2
(
1
a − b2 b )
ae(−b 
− be
−a 
)

AASTU,ECE, 2024 46
 Exercise
1. Consider a random process X(t) defined by

X (t ) = A cos(0t +  )
where 0 and  are constants and A is a uniform
random variable over the interval (0, 2)
a. Find the mean  X (t ).
b. Find the autocorrelation function RXX (t1 , t 2 ).
c. Find the autocovariance function C XX (t1 , t 2 ).
d . Determine whether X (t ) is WSS random process or not.
e. Find the power spectral density of X (t ).

AASTU,ECE, 2024 47
 Exercise
2. Consider a random process X(t) defined by:
X (t ) = A sin( 0 t +  )
where A and  are independent random variables which are
  
uniformly distributed over the intervals [0, 1] and − , 
 2 2
respectively and  0 is a constant.
a. Find the mean  X (t ).
b. Find the autocorrelation function R XX (t1 , t 2 ).
c. Find the autocovariance function C XX (t1 , t 2 ).
d . Determine whether X (t ) is WSS random process or not.
e. Find the power spectral density of X (t ).

AASTU,ECE, 2024 48
 Exercise
3. Two random processes X(t) and Y(t) are given by:

X (t ) = A cos(0t +  ) and Y (t ) = A sin(0t +  )

where A and 0 are constants and  is a uniform random
variable over the interval (0, 2 ).
a. Find the cross correlation function of X (t ) and Y (t ).
b. Verify that R XY (- ) = RYX ( )

4. Consider a discrete-time WSS random process X(n) with autocorrelation function

given by: −|k |
RXX (k ) = 2e
Find the power spectral density of X(n).
AASTU,ECE, 2024 49
Questions ?

AASTU,ECE, 2024 50