Ee 704 - Control Systems Ii Anith Krishnan 1
Ee 704 - Control Systems Ii Anith Krishnan 1
Ee 704 - Control Systems Ii Anith Krishnan 1
EE 704 - CONTROL
SYSTEMS II
ANITH KRISHNAN 1
MODULE IV
EE 704 CONTROL SYSTEMS II
EE 704 - CONTROL
SYSTEMS II
ANITH KRISHNAN 2
Random Process
2
EE 704 - CONTROL
SYSTEMS II
ANITH KRISHNAN 3
Expected Value
The weighted average of all possible values that the random
variable can take.
Weight
pdf
Also called Ensemble average.
EE 704 - CONTROL
SYSTEMS II
ANITH KRISHNAN 4
Expected Value (contd.)
| |
k k
p x p x p x p x X E + + + + =
3 3 2 2 1 1
Discrete RV
Suppose X can take value x
k
with probability p
k
Here
1
3 2 1
= + + + +
k
p p p p
Therefore
| |
k
k k
p p p p
p x p x p x p x
X E
+ + + +
+ + + +
=
3 2 1
3 3 2 2 1 1 Weighted
average
3
EE 704 - CONTROL
SYSTEMS II
ANITH KRISHNAN 5
Expected Value (contd.)
Continuous RV
Suppose X can take value x with pdf f(x)
| | ( ) x x xf X E d
}
=
EE 704 - CONTROL
SYSTEMS II
ANITH KRISHNAN 6
Expected Value (contd.)
6
1
Example: Let X be the outcome of rolling a six sided fair
die.
The possible values of X are {1,2,3,4,5,6}
Each have the probability
| | 5 . 3
6
1
6
6
1
5
6
1
4
6
1
3
6
1
2
6
1
1 = + + + + + = X E
4
EE 704 - CONTROL
SYSTEMS II
ANITH KRISHNAN 7
Expected Value and Mean
0 200 400 600 800 1000
2.2
2.4
2.6
2.8
3
3.2
3.4
3.6
3.8
4
Roll
M
e
a
n
o
f
X
EE 704 - CONTROL
SYSTEMS II
ANITH KRISHNAN 8
Mean
( ) ( )
=
=
N
i
i m
t x
N
t X
1
1
5
EE 704 - CONTROL
SYSTEMS II
ANITH KRISHNAN 9
Properties of Expected Value
EE 704 - CONTROL
SYSTEMS II
ANITH KRISHNAN 10
Stationary Random Process
Statistics do not change with time.
The behaviour is time-invariant, even though the
process is random.
1. Strict Sense Stationary (SSS)
2. Wide Sense Stationary (WSS)
6
EE 704 - CONTROL
SYSTEMS II
ANITH KRISHNAN 11
Stationary Random Process
Figure 1: Non-stationary process Figure 2: Stationary process
EE 704 - CONTROL
SYSTEMS II
ANITH KRISHNAN 12
SSS
( ) ( ) ( )
( )
( ) ( ) ( )
( )
n t X t X t X n t X t X t X
x x x f x x x f
n n
, , , , , ,
2 1 , , , 2 1 , , ,
2 1 2 1
=
+ + +
A process X(t) is said to be SSS if X(t) and X(t+) have
the same statistics (mean, variance etc.) for any .
For a stationary random process, the pdf does not
change over time.
( )
( )
( )
( )
=
+
x f x f
t X t X
Which gives
( )
( ) | | ( ) constant d = = = =
}
x x xf t X E
X X t X
7
EE 704 - CONTROL
SYSTEMS II
ANITH KRISHNAN 13
SSS
A process X(t) is said to be SSS if X(t) and X(t+) have
the same statistics for any .
( ) ( ) ( )
( )
( ) ( ) ( )
( ) n x x x f x x x f
n t X t X t X n t X t X t X
n n
=
+ + +
, , , , , ,
2 1 , , , 2 1 , , ,
2 1 2 1
A SSS process is also called a nth order stationary
process.
EE 704 - CONTROL
SYSTEMS II
ANITH KRISHNAN 14
Stationary Process
First Order Stationary Process
( )
( )
( )
( )
=
+
x f x f
t X t X
pdf is independent of time.
8
EE 704 - CONTROL
SYSTEMS II
ANITH KRISHNAN 15
Stationary Process
Second Order Stationary Process
( ) ( )
( )
( ) ( )
( )
2 1 , 2 1 ,
, ,
2 1 2 1
x x f x x f
t X t X t X t X + +
=
Also called WSS
EE 704 - CONTROL
SYSTEMS II
ANITH KRISHNAN 16
Auto-correlation
( ) ( ) ( ) | |
( )
2 1 2 1 2 1
2 1 2 1
d d ,
, ,
x x x x f x x
t X t X E t t R
X
} }
=
=
Similarity between observations as a function of the time
separation between them.
Or
How much two observations taken at different times can
vary.
9
EE 704 - CONTROL
SYSTEMS II
ANITH KRISHNAN 17
Auto-correlation (contd.)
Figure 3: Interpretation of the autocorrelation function
EE 704 - CONTROL
SYSTEMS II
ANITH KRISHNAN 18
Auto-correlation (contd.)
Properties
Let X(t) be a stationary random process.
1. The average power of the random process is
( ) ( ) | | t X E R
X
2
0 =
Proof:
( ) ( ) ( ) | |
( ) ( ) ( ) | |
( ) ( ) | | t X E R
t X t X E R
t X t X E t t R
X
X
X
2
2 1 2 1
0
,
=
+ =
=
10
EE 704 - CONTROL
SYSTEMS II
ANITH KRISHNAN 19
Auto-correlation (contd.)
( ) ( ) =
X X
R R
Properties
2. The auto-correlation function has even symmetry
Proof:
( ) ( ) ( ) | |
( ) ( ) ( ) | |
( ) ( ) ( ) | |
( )
X
X
X
X
R
a X a X E R
t X t X E R
t X t X E R
=
+ =
=
+ =
a t = Let
EE 704 - CONTROL
SYSTEMS II
ANITH KRISHNAN 20
Auto-correlation (contd.)
( ) ( ) 0
X X
R R s
Properties
3. The auto-correlation function is at a maximum at t=0
Proof:
( ) ( ) ( )
( ) ( ) ( ) | |
( ) ( ) ( ) ( ) | |
( ) ( ) ( )
( ) ( )
( ) ( )
X X
X X
X X X
R R
R R
R R R
t X t X t X t X E
t X t X E
t X t X
>
>
> +
> + + +
> +
> +
0
0 0
0 0 2 0
0 2
0
0
2 2
2
2
11
EE 704 - CONTROL
SYSTEMS II
ANITH KRISHNAN 21
Cross-correlation
Relation between two random processes as a function
of time lag applied to one of them.
( ) ( ) ( ) | |
( ) y x y x f xy
t Y t X E t t R
XY
XY
} }
=
=
d d ,
, ,
2 1 2 1
EE 704 - CONTROL
SYSTEMS II
ANITH KRISHNAN 22
WSS
A random process X(t) is WSS if
( ) ( ) | |
( ) ( ) | | ( )
X X
X
R t X t X E t t R
t X E t
= + = +
= =
) , (
constant
In other words, a random process X(t) is WSS if its two
statistics, its mean and auto-correlation, do not vary with
a shift in time origin.
SSS is WSS
12
EE 704 - CONTROL
SYSTEMS II
ANITH KRISHNAN 23
Power Spectral Density (PSD)
( ) ( ) | | ( )
( ) ( ) | | ( )
}
}
= =
= =
d
2
1
IFT
d FT
j
X X X
j
X X X
e S S R
e R R S
For a given random process X(t), the PSD of X(t) is the
Fourier transform of its auto-correlation.
Hz in is where
2 rad/sec in is
f
f =
EE 704 - CONTROL
SYSTEMS II
ANITH KRISHNAN 24
Properties of PSD
13
EE 704 - CONTROL
SYSTEMS II
ANITH KRISHNAN 25
Ergodic Process
Time average = Ensemble average
For a process to be ergodic, it has to be stationary.