Signals and Systems

ELCE 301

Time-domain Analysis of LTI Systems (p.2)
113
Analysis of LTI Systems using o inputs
Limitations of ODE-based analysis of LTI systems
1. Complex systems need complex differential equations (analytic
solution difficult or impossible).
2. For LTI systems, the solutions of homogeneous equations (which
are easier to find) have limited practicality.
Solutions of homogeneous equations define natural (zero-input)
responses of LTI systems.
3. Particular solutions of nonhomogeneous equations must be
individually found for each input signal x(t).
Solutions of nonhomogeneous equations define forced (zero-
state) responses of LTI systems.
114
Analysis of LTI Systems using o inputs
Sampling properties of o signals
117
( ) ( ) ( ) 0
}

=| o | dt t t
( ) ( ) ( )
}

= T dt T t t | o |
] [ ] [ ] [ ] [ ] [ ] [ ] 0 [ ] [ ] [
0 0 0 0
n n n n n n n n n n
n n n
| o | | o | | o | = + = =

+
=
+
=
+
=
For discrete signals (using Kronecker o):
( ) ( ) ( )
}

= + T dt t T t | o |
For continuous signals (using Dirac o):
Assume we know a response h[n] of a discrete LTI system H
to o [n] input, i.e.


The output signal h[n] is called impulse response of an LTI
system.
Then, we want to estimate the systems response y[n] to any
input signal x[n], using a sequence of mathematical
manipulations.
Analysis of LTI Systems using o inputs
LTI system outputs for o inputs (discrete variant)
118
( ) ] [ ] [ n n h o H =
Analysis of LTI Systems using o inputs
LTI system outputs for o inputs (discrete variant)
119
( ) ] [ ] [ ] [ ] [ |
.
|

\
|
= =

+
= k
k n k x n x n y o H H
( ) ] [ ] [ n x n y H =
This is what we want to find.
Sampling property of o function.
( ) ] [ ] [

+
=
=
k
k n k x o H Linearity of the system.
( ) ] [ ] [

+
=
=
k
k n k x o H Linearity of the system.
] [ ] [

+
=
=
k
k n h k x Time-invariance of the system.
Analysis of LTI Systems using o inputs
LTI system outputs for o inputs (discrete variant)
120
] [ ] [ ] [

+
=
=
k
k n h k x n y
CONCLUSION: Response of a discrete LTI system to any input signal
x[n] is a linear combination of delayed or advanced copies of its
impulse response h[n]. The coefficients of this linear combination are
the corresponding samples of x[n].
[ ] [ ] [ ] ... [ 1] [ 1] [0] [ ] [1] [ 1] ...
k
y n x k h n k x h n x h n x h n
+
=
= = + + + + +

Analysis of LTI Systems using o inputs

LTI system outputs for o inputs (discrete variant)
121
Response of a discrete LTI system to a signal x[n] is
a linear combination of delayed or advanced copies
of the systems impulse response h[n].
Assume we know a response h(t) of a CT LTI system H to o (t)
input, i.e.


The output signal h(t) is called impulse response of an LTI
system.
Then, we want to estimate the systems response y(t) to any
input signal x(t), using a similar sequence of mathematical
manipulations.
Analysis of LTI Systems using o inputs
LTI system outputs for o inputs (continuous variant)
122
| | ) ( ) ( t t h o H =
Analysis of LTI Systems using o inputs
LTI system outputs for o inputs (continuous variant)
123
| |
|
|
.
|

\
|
= =
}
+

t t o t d t x t x t y ) ( ) ( ) ( ) ( H H
| | ) ( ) ( t x t y H =
This is what we want to find.
Sampling property of o function.
( ) ) ( ) (
}
+

= t t o t d t x H Linearity of the system.
( ) t t o t d t x
}
+

= ) ( ) ( H Linearity of the system.
) ( ) ( t t t d t h x
}
+

= Time-invariance of the system.
Analysis of LTI Systems using o inputs
LTI system outputs for o inputs (continuous variant)
124
t t t d t h x t y
}
+

= ) ( ) ( ) (
CONCLUSION: Response of a continuous-time LTI system to any
input signal x(t) is an integral (summation) of delayed or advanced
copies of its impulse response. The coefficients of summation are the
corresponding values of the input signal x(t).
Analysis of LTI Systems using o inputs
Convolution
125
The presented results can be formally expressed using the concept
of convolution, i.e. by convolving the input signal with the impulse
response.
Convolution of discrete functions:



Convolution of continuous functions:
] [ ] [ ] [ ] [ ] [ ] [ ] [ ] [ ] [ n x n h k n x k h k n h k x n h n x n y
k k
- = = = - =

+
=
+
=
) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( t x t h d t x h d t h x t h t x t y - = = = - =
} }
+

+

t t t t t t
Analysis of LTI Systems using o inputs
Convolution
126
Commutative Property:

x(t)*y(t)=y(t)*x(t)

x[n]*y[n]=y[n]*x[n]

Distributive Property:
x(t)*(y
1
(t) + y
2
(t))=x(t)*y
1
(t) + x(t)*y
2
(t)

x[n]*(y
1
[n] + y
2
[n])=x[n]*y
1
[n] + x[n]*y
2
[n]

Associative Property:
x(t)*(y
1
(t)*y
2
(t))=(x(t)*y
1
(t))*y
2
(t)

x[n]*(y
1
[n]*y
2
[n])=(x[n]*y
1
[n])*y
2
[n]

Analysis of LTI Systems using o inputs
Main steps in analysis of LTI systems
127
1. Check if the system is LTI.
2. If yes, connect o(t) signal to the input and
calculate/estimate the impulse response h(t) (e.g. by
solving a differential equation, using measuring tools, etc.).
3. The impulse response h(t) provides complete information
about an LTI system. With it, you can predict the output
signal y(t) for any input signal x(t) by computing the
convolution, i.e. y(t) = x(t)*h(t).
Analysis of LTI Systems using o inputs
A note about causal systems
128
What is wrong with this system? Output exists before input!
Impulse response of a causal system MUST BE zero for negative times.
Analysis of LTI Systems using o inputs
A note about causal systems
129
Impulse response of a causal system MUST BE zero for negative times.
0 for 0 ) ( < = t t h
CONCLUSION: For causal LTI systems any integration
regarding the impulse response should be always done over a
limited domain. For example, the output y(t) for any input signal
x(t) would be computed by the following convolution
0
( ) ( ) ( ) ( ) ( ) ( ) ( )
t
y t x t h t x h t d x t h d t t t t t t
+

= - = =
} }
130
Analysis of LTI Systems using o inputs
Example (the same as before)

) (
1
) (
1
) ( ' or ) (
1
) (
1 ) (
t
RC
t h
RC
t h t
RC
t h
RC dt
t dh
o o = + = +
Complete solution:
) ( ) ( ) (
1 1
t h t h C t h
p
+ =
Already known from the related homogeneous ODE.
RC
t
e t h

= ) (
1
131
Analysis of LTI Systems using o inputs
Example
Particular solution of the nonhomogeneous equation:
) (
1
) (
1
) (
) (
) ( ) (
1
1
t u e
RC
d e
RC
e dt
t h
t b
t h t h
RC
t t
RC RC
t
p

= = =
} }
t t o
t
) (
1
) ( ) ( ) (
1 1 1
t u e
RC
e C t h t h C t h
RC
t
RC
t
p

+ = + =
Because the system must be causal, then
0 0 ) 0 (
1 1
= = = <

C e C t h
RC
t
) (
1
) ( t u e
RC
t h
RC
t

=
132
Analysis of LTI Systems using o inputs
Example
Now, we can calculate the response of the RC circuit to any
input. For example, assume that x(t)= u(t).
0 0
( ) ( ) ( ) ( ) ( )
1 1 1
( ) ( ) ( ) ( )
0 0
( ) ( 1) (1 ) ( )
1 0
t
t t t
t t t
RC RC RC RC
t t t
RC RC RC
t
RC
y t u t h t u h t d
u e u t d u t e d u t e e d
RC RC RC
t
u t e e e u t
e t
t t t
t t t
t t t t t

= - =
= = =
s

= = =

>

}
} } }
The same result as before!
OBSERVATION: Convolutions are computationally complex and
tedious operations. Simplified/approximate methods are welcome.
133
Analysis of LTI Systems using o inputs
Graphical evaluation of convolution integrals
Four basic steps to evaluate c(t)= a(t)*b(t):
1. Plot diagrams of the functions a(t) and b(t).
2. Create a mirror reflection of the function b, i.e. b(t) => b(-t).
3. Slide the diagram of function b(-t) along the horizontal axis,
i.e. b(-t) => b(t-t).
4. Estimate the integral over the intersection area of diagrams
a(t) and b(t-t).

134
Analysis of LTI Systems using o inputs
Graphical evaluation of convolution integrals (example)
Find the convolution y(t) = h(t)*x(t), where
135
Analysis of LTI Systems using o inputs
Graphical evaluation of convolution integrals - Example
Steps 1 and 2 (note that is used instead of t)
Steps 3 and 4 (several cases shown)
136
Analysis of LTI Systems using o inputs
Graphical evaluation of convolution integrals - Example
Steps 3 and 4 (cont.)
137
Analysis of LTI Systems using o inputs
Graphical evaluation of convolution integrals - Example
Steps 3 and 4 (cont.)
Final result
138
Analysis of LTI Systems using o inputs
Evaluation of convolution integrals Delta method
The graphical evaluation is tedious.
Example:
0
0
t
t
x(t) h(t)
-2
1
2
1
2
1 2 1
3
-1
2
-1 -2
x(2- )
1
3 2
h()
x(- )
139
Analysis of LTI Systems using o inputs









t
x(t - )
t
h()
Evaluation of convolution integrals Delta method
The graphical evaluation is tedious.
140
Analysis of LTI Systems using o inputs









Evaluation of convolution integrals Delta method
The graphical evaluation is tedious.
Selected case of integration (for 1< t s 2)
} } } }

+ + + + =
1
0
1
1 1
3
) 5 . 0 5 . 0 1 ( 2 2 1 2 ) (
t
t
t
t
d t d d d t y t t t t t
3 2 1 -1 -2
h(t)
1 2
dx/dt
141
Analysis of LTI Systems using o inputs









Evaluation of convolution integrals Delta method
Delta-representation of polynomial functions.
) 3 ( 2
1


t o
) 1 (
1

t o
) 1 (
1

t o
) (
1
t

o
) 2 (
2
1
2
+

t o
) (
2
1
2
t

o
-2
t
t
t
t
dh/dt
0 0
) 2 ( 2
1


t o
x(t)
) 3 ( 2 ) 1 ( ) ( ) (
1 1 1
+ =

t t t t h o o o
) 2 ( 2 ) 1 (
) (
2
1
) 2 (
2
1
) (
1 1
2 2
+
+ =


t t
t t t x
o o
o o
142
Analysis of LTI Systems using o inputs
Evaluation of convolution integrals Delta method
Table for Delta-method convolution.
x(t)
(
(
(

2
2
2 / 1
(
(
(

0
2
2 / 1
(
(
(

1
1
1
(
(
(

2
1
2
(
(
(

0
1
1
(
(
(

1
1
1
(
(
(

3
1
2
(
(
(

2
3
2 / 1
(
(
(

1
3
2 / 1
(
(
(

1
3
1
(
(
(

0
3
2 / 1
(
(
(

1
3
2 / 1
(
(
(

3
3
1
(
(
(

1
2
1
(
(
(

2
2
2
(
(
(

3
2
2
(
(
(

2
2
1
(
(
(

4
2
2
(
(
(

5
2
4
h(t)
143
Analysis of LTI Systems using o inputs
Evaluation of convolution integrals Delta method
Table for Delta-method convolution.
(
(
(

2
3
2 / 1
(
(
(

3
2
2
) 2 (
2
1
) (
3
+ =

t t y o
2
( ) 2 ( 3) y t t o

=
) 5 ( 4 ) 4 ( 2 ) 3 (
) 2 ( ) 1 (
2
3
) 1 (
2
1
) 3 ( 4 ) 1 ( ) (
2
1
) 2 (
2
1
) 5 ( 4 ) 4 ( 2 ) 3 ( ) 1 ( ) 3 ( 2
) 2 ( ) 1 (
2
1
) 1 (
2
1
) 3 ( 2 ) 1 ( ) (
2
1
) 2 (
2
1
) (
2 2 3
2 3 3 2 2 3 3
2 2 3 3 2
2 3 3 2 2 3 3
+ +
+ + + + +
= + +
+ + + + + =




t t t
t t t t t t t
t t t t t
t t t t t t t t y
o o o
o o o o o o o
o o o o o
o o o o o o o
144
Analysis of LTI Systems using o inputs
Extras (other operations similar to convolution)
Cross-correlation
The cross-correlation (symbol is used) of functions x(t) and h(t) is defined
by the convolution of functions or x(t) and h( t)], i.e.
Cross-correlation indicates how similar two signals are when one of them
is delayed/advanced by various amounts of time. The maximum value of
the cross-correlation corresponds to the delay when both signals are the
most similar.
145
Analysis of LTI Systems using o inputs
Extras (other operations similar to convolution)
Cross-correlation
Properties and computational schemes for cross-correlation can be easily
obtained from the corresponding properties/schemes of convolution.
Auto-correlation
Auto-correlation is the cross-correlation of a signal with itself.


It represents similarity between the original signal and its
delayed/advanced copies. Auto-correlation is a tool for detecting repeating
patterns in the analyzed signals (including detection of periodic signals).