Scribd
Upload
369 views12 pages

Random Processes

The document discusses wide-sense stationary (WSS) random processes. A process is WSS if its mean and autocorrelation function do not vary with shifts in time. The autocorrelation function measures how samples of a process are related over time. If a process is passed through a linear, time-invariant system, the output will also be WSS and its power spectral density will be equal to the input's PSD multiplied by the system's transfer function squared.

Uploaded by

Rahul Sarangle
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
369 views12 pages

Random Processes

The document discusses wide-sense stationary (WSS) random processes. A process is WSS if its mean and autocorrelation function do not vary with shifts in time. The autocorrelation function measures how samples of a process are related over time. If a process is passed through a linear, time-invariant system, the output will also be WSS and its power spectral density will be equal to the input's PSD multiplied by the system's transfer function squared.

Uploaded by

Rahul Sarangle
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 12

2.2.

Wide-Sense Stationary (WSS)


Processes

Ø Mean of the random process Ø Autocorrelation function of X(t) is


X(t) is the mean of random a function of two variables t1 = t
variable X(t) at time instant t. and t2 = t + τ,
E[ X (t )] = µ X (t ) R X (t, t + τ ) = E[ X (t ) X (t + τ )]

Let fX(t)(x) be the pdf of X(t)


at time instant t. This is a measure of the degree to
which two time samples of the same
+∞
random process are related
E[ X (t )] =
∫
−∞
xf X (t ) ( x)dx
Ø Autocorrelation function of X(t) is a function of two variables t1 = t
and t2 = t + τ,

R X (t , t + τ ) = E[ X (t ) X (t + τ )]

This is a measure of the degree to which two time samples of the


same random process are related
What is a WSS Process?

A random process X(t) is WSS if


(i ) µ X (t ) = E[ X (t )] = constant
(ii ) R X (t , t − τ ) = E[ X (t ) X (t − τ )] = R X (τ )

In other words, a random process X(t) is WSS if its two


statistics, its mean and autocorrelation, do not vary with a shift
in the time origin.
Example 2.3. Find the mean and
autocorrelation function of the RX (t, t + τ )
random process X(t) ( in E.g. 2.2), = E [ A cos( 2πf t + Θ) A cos( 2πf (t + τ ) + Θ)]
c c
X (t ) = A cos( 2πf c t + Θ) 1 1 
= A 2 E  cos(2πf cτ ) + cos(2πf c (2t + τ ) + 2Θ )
2 2 
where Θ is uniformly
distributed over [0, 2 π]. A2
= cos(2πf cτ )
1 /(2π ), 0 ≤ θ ≤ 2π 2
pdf : fΘ (θ ) = 
0, otherwise

Is X(t) a WSS? Since the mean and autocorrelation


of X(t) do not depend on time t, then
Solution. X(t) is a WSS process.
According to the definitions,
2π
1
E[ X (t )] = ∫ A cos( 2πf ct + θ )
0 2π
dθ

=0
Properties of Autocorrelation Function
of a WSS process X(t)

1. R X (τ ) = RX (−τ ) Symmetric in τ about zero

2. R X (τ ) ≤ RX (0) for all τ, Maximum value occurred at the


origin

3 . R X (τ ) ↔ S X ( f ) Autocorrleation and psd form a


pair of the Fourier transform

4. R X (0) = E[ X (t ) 2 ] The value at origin is equal to the


average power of the signal
Power Spectral Density (PSD) of a
WSS Random Process

For a given WSS process X(t), the psd of X(t) is the


Fourier transform of its autocorrelation, i.e.,

+∞
S X ( f ) = F ( R X (τ )) = ∫−∞
R X (τ ) e − j 2 π fτ d τ

+∞
∫
−1
R X (τ ) = F ( S X ( f )) = S X ( f ) e j 2π f τ df
−∞
For the random process in Example 2.3, we have
A2
R X (τ ) = cos( 2πf cτ )
2
Hence, the psd of X(t) is the Fourier transform of the
autocorrelation of X(t), given by

A2
SX ( f ) = [δ ( f − f c ) + δ ( f + f c )]
4

A2/4 SX(f)
A2/4

- fc fc f
Solution.
First we need to show that Y(t) is WSS.
Mean:

mY (t ) = E[Y (t )]
Example 2. 4 Let
= E[ X (t ) cos( 2πf ct + Θ)]
Y (t ) = X (t ) cos( 2πf ct + Θ )
where X(t) is a WSS process with = E[ X (t )]E[cos( 2πf c t + Θ)]
psd SX(f), Θ is uniformly
distributed over [0, 2 π], and X(t) (by independence )
is independent of Θ and
cos(2πf ct + Θ) = m X (t ) ⋅ 0 = 0
Find the psd of Y(t). (by Example 2.3)
Autocorrelation of Y(t):

RY (t , t + τ ) = E[Y (t )Y (t + τ )]
= E[ X (t ) cos( 2πf ct + Θ) X (t + τ ) cos( 2πf c (t + τ ) + Θ)]

= E[ X (t ) X (t + τ )] E[cos( 2πf c t + Θ) cos( 2πf c (t + τ ) + Θ)]

By Example 2.3.
1
= R X (τ ) cos(2πf cτ ) = RY (τ )
2
Hence, Y(t) is WSS. Therefore
1 j 2πf cτ − j 2πf cτ
SY (t ) = F [ RY (τ )] = F [ RX (τ )(e +e )]
4
= [S X ( f − f c ) + S X ( f + f c )]
1
4
Properties of PSD

1. S X ( f ) ≥ 0 always real valued

2. S X ( f ) = S X (− f ) for X(t) real-valued

3. S X ( f ) ↔ R X (τ ) a pair of Fourier transform

+∞
4. P = R X (0) = ∫ S X ( f )df Relationship between
−∞
average power and psd
Transmission over LTI Systems

n Response of LTI system to a Properties of the output:


random input X(t):
1) If X(t) is WSS, so does Y(t).

X(t) Y(t) 2) Mean:


h(t) µY = µX H(0)

3) Autocorrelation:
Y (t ) = X (t ) ∗ h(t ) RY (τ ) = R X (τ ) ∗ h(τ ) ∗ h (−τ )
∞
= ∫ x(t − τ )h(τ )dτ
−∞
4) PSD:

S Y ( f ) = S X ( f ) | H ( f ) |2
X(t) Y(t)
RX(τ) h(t) Ry(τ)

F F-1

SX(f) |H(f)|2 SY(f)

You might also like