Unit 2 signal &system

ELG 3120 Signals and Systems Chapter 2
1/2 Yao
Chapter 2 Linear Time-Invariant Systems
2.0 Introduction
• Many physical systems can be modeled as linear time-invariant (LTI) systems
• Very general signals can be represented as linear combinations of delayed impulses.
• By the principle of superposition, the response ]
[n
y of a discrete-time LTI system is the sum
of the responses to the individual shifted impulses making up the input signal ]
[n
x .
2.1 Discrete-Time LTI Systems: The Convolution Sum
2.1.1 Representation of Discrete-Time Signals in Terms of Impulses
A discrete-time signal can be decomposed into a sequence of individual impulses.
Example:
n
x[n]
2
1
-1
-1
-2
-3
-4 1 2
3
4
Fig. 2.1 Decomposition of a discrete-time signal into a weighted sum of shifted impulses.
The signal in Fig. 2.1 can be expressed as a sum of the shifted impulses:
...
]
2
[
]
2
[
]
1
[
]
1
[
]
[
]
0
[
]
1
[
]
1
[
]
2
[
]
2
[
]
3
[
]
3
[
...
]
[ +
−
+
−
+
+
+
−
+
+
−
+
+
−
+
= n
x
n
x
n
x
n
x
n
x
n
x
n
x δ
δ
δ
δ
δ
δ
(2.1)
or in a more compact form
∑
∞
−∞
=
−
=
k
k
n
k
x
n
x ]
[
]
[
]
[ δ . (2.2)
This corresponds to the representation of an arbitrary sequence as a linear combination of shifted
unit impulse ]
[ k
n −
δ , where the weights in the linear combination are ]
[k
x . Eq. (2.2) is called
the sifting property of the discrete-time unit impulse.

2/2 Yao
2.1.2 Discrete-Time Unit Impulse Response and the Convolution – Sum Representation of LTI
Systems
Let ]
[n
hk be the response of the LTI system to the shifted unit impulse ]
[ k
n −
δ , then from the
superposition property for a linear system, the response of the linear system to the input ]
[n
x in
Eq. (2.2) is simply the weighted linear combination of these basic responses:
∑
∞
−∞
=
=
k
k n
h
k
x
n
y ]
[
]
[
]
[ . (2.3)
If the linear system is time invariant, then the responses to time-shifted unit impulses are all
time-shifted versions of the same impulse responses:
]
[
]
[ 0 k
n
h
n
hk −
= . (2.4)
Therefore the impulse response ]
[
]
[ 0 n
h
n
h = of an LTI system characterizes the system
completely. This is not the case for a linear time-varying system: one has to specify all the
impulse responses ]
[n
hk (an infinite number) to characterize the system.
For the LTI system, Eq. (2.3) becomes
∑
∞
−∞
=
−
=
k
k
n
h
k
x
n
y ]
[
]
[
]
[ . (2.5)
This result is referred to as the convolution sum or superposition sum and the operation on the
right-hand side of the equation is known as the convolution of the sequences of ]
[n
x and ]
[n
h .
The convolution operation is usually represented symbolically as
]
[
]
[
]
[ n
h
k
x
n
y ∗
= . (2.6)
2.1.3 Calculation of Convolution Sum
• One way to visualize the convolution sum of Eq. (2.5) is to draw the weighted and shifted
impulse responses one above the other and to add them up.

3/2 Yao
Example: Consider the LTI system with impulse response ]
[n
h and input ]
[n
x , as illustrated in
Fig. 2. 2.
n
h[n]
1
1 2
0
1
1
n
x[n]
0.5
1
0
2
(a)
The output response based on Eq. (2.5) can be expressed
]
1
[
2
]
[
5
.
0
]
1
[
]
1
[
]
0
[
]
0
[
]
[
]
[
]
[
1
0
−
+
=
−
+
−
=
−
= ∑
=
n
h
n
h
n
h
x
n
h
x
k
n
h
k
x
n
y
k
.
2
n
0.5
1
0
x[0]h[n]=0.5h[n]
3
n
2
2
1
x[1]h[n-1]=2h[n-1]
0
(b)
3
n
2.5
2
1
y[n]
0
2
0.5
2.5
(c)
Fig. 2.2 (a) The impulse response ]
[n
h of an LTI system and an input ]
[n
x to the system; (b) the
responses to the nonzero values of the input; (c) the overall responses.

4/2 Yao
• Another way to visualize the convolution sum is to draw the signals ]
[k
x and ]
[ k
n
h − as
functions of k (for a fixed n), multiply them to form the signal ]
[k
g , and then sum all
values of ]
[k
g .
Example: Calculate the convolution of ]
[k
x and ]
[n
h shown in Fig. 2.2 (a).
k
x[k]
0.5
1
0
h[n-k], n<0
1
1 2
0
k
h[0-k], n=0
1
1 2
0
-1
-2
k
h[1-k], n=1
1
1
0
-1
k
h[2-k], n=2
1
1
0 2
k
h[3-k], n=3
1
1 2 3
k
h[n-k], n>3 1
0
0
2
Fig. 2.3 Interpretation of Eq. (2.5) for the signals ]
[k
x and ]
[n
h .

5/2 Yao
For 0
<
n , 0
]
[ =
n
y
For 0
=
n , ∑
∞
−∞
=
=
−
=
k
k
h
k
x
y 5
.
0
]
0
[
]
[
]
0
[
For 1
=
n , ∑
∞
−∞
=
=
+
=
−
=
k
k
h
k
x
y 5
.
2
2
5
.
0
]
1
[
]
[
]
1
[
For 1
=
n , ∑
∞
−∞
=
=
+
=
−
=
k
k
h
k
x
y 5
.
2
2
5
.
0
]
2
[
]
[
]
2
[
For 1
=
n , ∑
∞
−∞
=
=
−
=
k
k
h
k
x
y 2
]
2
[
]
[
]
3
[
For 3
>
n , 0
]
[ =
n
y
The resulting output values agree with those obtained in the preceding example.
Example: Compute the response of an LTI system described by its impulse response


 ≤
≤
=
otherwise
n
n
h
n
,
0
6
0
,
]
[
α
to the input signal


 ≤
≤
=
otherwise
n
n
x
,
0
4
0
,
1
]
[ .
x[n]
n
1
1 2
0 3 4
h[n],
n
1 2
0 3 4
1
>
α
To do the analysis, it is convenient to consider five separate intervals:
For 0
<
n , there is no overlap between the nonzero portions of ]
[n
x and ]
[ k
n
h − , and
consequently, .
0
]
[ =
n
y
For 4
0 ≤
≤ n ,


 ≤
≤
=
−
−
otherwise
n
k
k
n
h
k
x
k
n
,
0
0
,
]
[
]
[
α
,

6/2 Yao
Thus, in this interval
α
α
α
α
α
α
α
α
−
−
=








−
−
=
=
=
+
−
−
−
=
−
=
−
∑
∑ 1
1
1
1
]
[
1
1
1
0
0
n
n
n
n
k
k
n
n
k
k
n
n
y
For 6
4 ≤
< n ,


 ≤
≤
=
−
−
otherwise
k
k
n
h
k
x
k
n
,
0
4
0
,
]
[
]
[
α
( ) ( )
α
α
α
α
α
α
α
α
α
−
−
=
−
−
=
=
=
+
−
−
−
=
−
=
−
∑
∑ 1
1
1
]
[
1
4
1
5
1
4
0
1
4
0
n
n
n
k
k
n
k
k
n
n
y .
For 10
6 ≤
< n ,


 ≤
≤
−
=
−
−
otherwise
k
n
k
n
h
k
x
k
n
,
0
4
)
6
(
,
]
[
]
[
α
∑
−
=
−
=
4
6
]
[
n
k
k
n
n
y α .
Let 6
+
−
= n
k
r , ( ) α
α
α
α
α
α
α
α
α
−
−
=
−
−
=
=
=
−
−
−
−
=
−
−
=
−
∑
∑ 1
1
1
]
[
7
4
1
11
6
10
0
1
6
10
0
6
n
n
n
r
r
n
r
r
n
y .
For 4
6 >
−
n , or 10
>
n , there is no overlap between the nonzero portions of ]
[k
x and ]
[ k
n
h − ,
and hence, 0
]
[ =
n
y .
The output is illustrated in the figure below.
y[n]
n
1 2
0 3 4 5 6 7 8 9 10

7/2 Yao
2.2 Continuous-Time LTI systems: the Convolution Integral
The response of a continuous-time LTI system can be computed by convolution of the impulse
response of the system with the input signal, using a convolution integral, rather than a sum.
2.2.1 Representation of Continuous-Time Signals in Terms of Impulses
A continuous-time signal can be viewed as a linear combination of continuous impulses:
∫
∞
∞
−
−
= τ
τ
δ
τ d
t
x
t
x )
(
)
(
)
( . (2.7)
The result is obtained by chopping up the signal )
(t
x in sections of width ∆, and taking sum
∆
∆
− ∆
2 ∆
3
0
t
)
(t
x
Recall the definition of the unit pulse )
(t
∆
δ ; we can define a signal )
(
ˆ t
x as a linear combination
of delayed pulses of height ∆
k
x( )
∑
∞
−∞
=
∆ ∆
∆
−
∆
=
k
k
t
k
x
t
x )
(
)
(
)
(
ˆ δ (2.8)
Taking the limit as 0
→
∆ , we obtain the integral of Eq. (2.7), in which when 0
→
∆
(1) The summation approaches to an integral
(2) τ
→
∆
k and )
(
)
( τ
x
k
x →
∆
(3) τ
d
→
∆
(4) )
(
)
( τ
δ
δ −
→
∆
−
∆ t
k
t
Eq. (2.7) can also be obtained by using the sampling property of the impulse function. If we
consider t is fixed and τ is time variable, then we have )
(
)
( τ
δ
τ −
t
x
)
(
)
(
))
(
(
)
( t
t
x
t
x −
=
−
−
= τ
δ
τ
δ
τ . Hence

8/2 Yao
)
(
)
(
)
(
)
(
)
(
)
(
)
( t
x
d
t
t
x
d
t
x
d
t
x =
−
=
−
=
− ∫
∫
∫
∞
∞
−
∞
∞
−
∞
∞
−
τ
τ
δ
τ
τ
δ
τ
τ
τ
δ
τ . (2.9)
As in discrete time, this is the sifting property of continuous-time impulse.
2.2.2 Continuous-Time Unit Impulse Response and the Convolution Integral Representation
of an LTI system
The linearity property of an LTI system allows us to calculate the system response to an input
signal )
(
ˆ t
x using Superposition Principle. Let )
(
ˆ t
hk∆ be the pulse response of the linear-varying
system to the unit pulses )
( ∆
−
∆ k
t
δ for +∞
<
<
∞
− k . The response of the system to )
(
ˆ t
x is
∑
∞
−∞
=
∆ ∆
∆
−
∆
=
k
k k
t
h
k
x
t
y )
(
)
(
)
(
ˆ . (2.10)
Note that the response )
(
ˆ t
hk∆ tends to the impulse response )
(t
hτ as 0
→
∆ . Then at the limit,
we obtain the response of the system to the input signal )
(
ˆ
lim
)
(
0
t
x
t
x
→
∆
= :
∫
+∞
∞
−
→
∆
=
= τ
τ τ d
t
h
x
t
y
t
y )
(
)
(
)
(
ˆ
lim
)
(
0
. (2.11)
For an LTI system, the impulse responses )
(t
hτ are the same as )
(
0 t
h , except they are shifted by
τ , that is, )
(
)
( 0 k
t
h
t
h −
=
τ . Then we may define the unit impulse response of the LTI system
)
(
)
( 0 t
h
t
h = , (2.12)
and an LTI system is completely determined by its impulse response.
So the response to the input signal )
(t
x can be written as a convolution integral:
∫
+∞
∞
−
−
= τ
τ
τ d
t
h
x
t
y )
(
)
(
)
( , (2.13)
or it can be expressed symbolically
)
(
)
(
)
( t
h
t
x
t
y ∗
= . (2.14)

9/2 Yao
2.2.3 Calculation of convolution integral
The output )
(t
y is a weighted integral of the input, where the weight on )
(τ
x is )
( τ
−
t
h . To
evaluate this integral for a specific value of t ,
• First obtain the signal )
( τ
−
t
h (regarded as a function of τ with t fixed) from )
(τ
h by a
reflection about the origin and a shift to the right by t if t >0 or a shift to the left by t is t <0.
• Then multiply together the signals )
(τ
x and )
( τ
−
t
h .
• )
(t
y is obtained by integrating the resulting product from −∞
=
τ to +∞
=
τ .
Example: Let )
(t
x be the input to an LTI system with unit impulse response )
(t
h , where
)
(
)
( t
u
e
t
x at
−
= , 0
>
a and )
(
)
( t
u
t
h = .
Step1: The functions )
(τ
h , )
(τ
x and )
( τ
−
t
h are depicted
τ
)
(τ
h
0
1
τ
)
(τ
x
0
1
τ
)
( τ
−
t
h
0
1 0
<
t
t

10/2 Yao
t
τ
)
( τ
−
t
h
0
1
0
>
t
Step 2: From the figure we can see that for 0
<
t , the product of the product )
(τ
x and )
( τ
−
t
h is
zero, and consequently, )
(t
y is zero. For 0
>
t


 <
<
=
−
−
otherwise
t
e
t
h
x
at
,
0
0
,
)
(
)
(
τ
τ
τ
Step 3: Compute )
(t
y by integrating the product for 0
>
t :
)
1
(
1
1
)
( 0
0
at
t
a
t
a
e
a
e
a
d
e
t
y −
−
−
−
=
−
=
= ∫
τ
τ
τ .
The output of )
(t
y for all t is
)
(
)
1
(
1
)
( t
u
e
a
t
y at
−
−
= , and is shown in figure below.
t
)
(t
y
0
a
1
Example: Compute the convolution of the two signals below:


 <
<
=
otherwise
T
t
t
x
,
0
0
,
1
)
( and


 <
<
=
otherwise
T
t
t
t
h
,
0
2
0
,
)
(
For this example, it is convenient to calculate the convolution in separate intervals. )
(τ
x is
sketched and )
( τ
−
t
h is sketched in each of the intervals:
t > 0

11/2 Yao
For 0
<
t , and T
t 3
> , 0
)
(
)
( =
−τ
τ t
h
x for all the values of τ , and consequently )
(t
y =0.
For other intervals, the product )
(
)
( τ
τ −
t
h
x can be found in the figure on the next page. Thus for
these three intervals, the integration can be calculated with the result shown below:

12/2 Yao
τ
)
( τ
−
t
h
0
0
<
t
τ
)
(τ
x
0
1
T
2
T
t 2
− t
)
( τ
−
t
h
0
T
t <
<
0
T
2
T
t 2
− t
τ
)
( τ
−
t
h
0
T
t
T 2
<
<
T
2
T
t 2
− t
τ
)
( τ
−
t
h
0
T
t
T 3
2 <
<
T
2
T
t 2
− t
τ
)
( τ
−
t
h
0
T
t 3
>
T
2
T
t 2
− t
T
τ
)
(
)
( τ
τ −
t
h
x
0
T
t <
<
0
T
2
t
t
)
(
)
( τ
τ −
t
h
x
0
T
2
T
t T
t
T 2
<
<
)
(
)
( τ
τ −
t
h
x
0
T
2
T
T
t −
T
t
T 3
2 <
<

13/2 Yao











>
<
<
+
+
−
<
<
−
<
<
<
=
T
t
T
t
T
T
Tt
t
T
t
T
T
Tt
T
t
t
t
t
y
3
,
0
3
2
,
2
3
2
1
2
,
2
1
0
,
2
1
0
,
0
)
(
2
2
2
2
2.3 Properties of Linear Time-Invariant Systems
LTI systems can be characterized completely by their impulse response. The properties can also
be characterized by their impulse response.
2.3.1 The Commutative Property of LTI Systems
A property of convolution in both continuous and discrete time is a Commutative Operation.
That is
∑
∑
∞
−∞
=
∞
−∞
=
−
=
−
=
∗
=
∗
k
k
k
n
k
k
h
k
n
h
k
x
n
x
n
h
n
h
n
x ]
[
]
[
]
[
]
[
]
[
]
[
]
[
]
[ , (2.15)
τ
τ
τ
τ
τ
τ d
t
x
h
d
t
h
x
t
x
t
h
t
h
t
x ∫
∫
∞
∞
−
∞
∞
−
−
=
−
=
∗
=
∗ )
(
)
(
)
(
)
(
)
(
)
(
)
(
)
( . (2.16)
h
x y x
h y
2.3.2 The Distributive Property of LTI Systems
2
1
2
1 )
( h
x
h
x
h
h
x ∗
+
∗
=
+
∗ (2.17)
for both discrete-time and continuous-time systems. The property means that summing the
outputs of two systems is equivalent to a system with an impulse response equal to the sum of
the impulse response of the two individual systems, as shown in the figure below.
t
)
(t
y
0 T T
2 T
3

14/2 Yao
h1
x
h2
y
+
x y
h1+h2
The distributive property of convolution can be exploited to break a complicated convolution
into several simpler ones.
For example, an LTI system has an impulse response ]
[
]
[ n
u
n
h = , with an input
]
[
2
]
[
2
1
]
[ n
u
n
u
n
x n
n
−
+






= . Since the sequence ]
[n
x is nonzero along the entire time axis. Direct
evaluation of such a convolution is somewhat tedious. Instead, we may use the distributive
property to express ]
[n
y as the sum of the results of two simpler convolution problems. That is,
]
[
2
1
]
[
1 n
u
n
x
n






= , ]
[
2
]
[
2 n
u
n
x n
−
= , using the distributive property we have
( ) ]
[
]
[
)
(
)
(
)
(
)
(
)
(
)
(
)
(
]
[ 2
1
2
1
2
1 n
y
n
y
t
h
t
x
t
h
t
x
t
h
t
x
t
x
n
y +
=
∗
+
∗
=
∗
+
=
2.3.3 The Associative Property
( ) ( ) 2
1
2
1 h
h
x
h
h
x ∗
∗
=
∗
∗ . (2.18)
for both discrete-time and continuous-time systems.
h1 h2
x y
h1*h2
x y
• For LTI systems, the change of order of the cascaded systems will not affect the response.

15/2 Yao
• For nonlinear systems, the order of cascaded systems in general cannot be changed. For
example, a two memoryless systems, one being multiplication by 2 and the other squaring the
input, the outputs are different if the order is changed, as shown in the figure below.
2
x y=4x2
w2
w=2x
x2
x y=2x2
2
w=x2
2.3.4 LTI system with and without memory
A system is memoryless if its output at any time depends only on the value of its input at the
same time. This is true for a discrete-time system, if 0
]
[ =
n
h for 0
≠
n . In this case, the impulse
response has the form
]
[
]
[ n
K
n
h δ
= , (2.19)
where ]
0
[
h
K = is a constant and the convolution sum reduces to the relation
]
[
]
[ n
Kx
n
y = . (2.20)
Otherwise the LTI system has memory.
For continuous-time systems, we have the similar results if it is memoryless:
)
(
)
( t
K
t
h δ
= , (2.21)
)
(
)
( t
Kx
t
y = . (2.22)
Note that if 1
=
K in Eqs. (2.19) and (2.21), the systems become identity systems, with output
equal to the input.
2.3.5 Invertibility of LTI systems
We have seen that a system S is invertible if and only if there exists an inverse system S-1
such
that S-1
S is an identity system.
h
x y=x
h1

16/2 Yao
Since the overall impulse response in the figure above is 1
h
h ∗ , 1
h must satisfy for it to be the
impulse response of the inverse system, namely δ
=
∗ 1
h
h .
identity
system
x y=x
Applications - channel equalization: for transmission of a signal over a communication channel
such as telephone line, radio link and fiber, the signal at the receiving end is often processed
through a filter whose impulse response is designed to be the inverse of the impulse response of
the communication channel.
Example: Consider a system with a pure time shifted output )
(
)
( 0
t
t
x
t
y −
= .
The impulse response of this system is )
(
)
( 0
t
t
t
h −
= δ , since )
(
)
(
)
( 0
0 t
t
t
x
t
t
x −
∗
=
− δ , that is,
convolution of a signal with a shifted impulse simply shifts the signal
To recover the signal from the output, that is, to invert the system, all that is required is to shift
the output back. So the inverse system should have a impulse response of )
( 0
t
t +
δ , then
)
(
)
(
)
( 0
0 t
t
t
t
t δ
δ
δ =
+
∗
−
Example: Consider the LTI system with impulse response ]
[
]
[ n
u
n
h = .
The response of this system to an arbitrary input is
∑
+∞
∞
−
−
= ]
[
]
[
]
[ k
n
u
k
x
n
y .
Considering that ]
[ k
n
u − is 0 for 0
<
− k
n and 1 for 0
≥
− k
n , so we have
∑
∞
−
=
n
k
x
n
y ]
[
]
[ .
This is a system that calculates the running sum of all the values of the input up to the present
time, and is called a summer or accumulator. This system is invertible, and its inverse is given as
]
1
[
]
[
]
[ −
−
= n
x
n
x
n
y ,
It is a first difference operation. The impulse response of this inverse system is
]
1
[
]
[
]
[
1 −
−
= n
n
n
h δ
δ ,

17/2 Yao
We may check that the two systems are really inverses to each other:
{ } ]
[
]
1
[
]
[
]
1
[
]
[
*
]
[
]
[
*
]
[ 1 n
n
u
n
u
n
n
n
u
n
h
n
h δ
δ
δ =
−
−
=
−
−
=
2.3.6 Causality for LTI systems
A system is causal if its output depends only on the past and present values of the input signal.
Specifically, for a discrete-time LTI system, this requirement is ]
[n
y should not depend on ]
[k
x
for n
k > . Based on the convolution sum equation, all the coefficients ]
[ k
n
h − that multiply
values of ]
[k
x for n
k > must be zero, which means that the impulse response of a causal
discrete-time LTI system should satisfy the condition
0
]
[ =
n
h , for 0
<
n (2.23)
A causal system is causal if its impulse response is zero for negative time; this makes sense as
the system should not have a response before impulse is applied.
A similar conclusion can be arrived for continuous-time LTI systems, namely
0
)
( =
t
h , for 0
<
t (2.24)
Examples: The accumulator ]
[
]
[ n
u
n
h = , and its inverse ]
1
[
]
[
]
[ −
−
= n
n
n
h δ
δ are causal. The
pure time shift with impulse response )
(
)
( 0
t
t
x
t
y −
= for 0
0 >
t is causal, but is not causal
for 0
0 <
t .
2.3.7 Stability for LTI Systems
Recall that a system is stable if every bounded input produces a bounded output.
For LTI system, if the input ]
[n
x is bounded in magnitude
B
n
x ≤
]
[ , for all n
If this input signal is applied to an LTI system with unit impulse response ]
[n
h , the magnitude of
the output
∑
∑
∑
+∞
−∞
=
+∞
−∞
=
+∞
−∞
=
≤
−
≤
−
=
k
k
k
k
h
B
k
n
x
k
h
k
n
x
k
h
n
y ]
[
]
[
]
[
]
[
]
[
]
[ (2.25)
]
[n
y is bounded in magnitude, and hence is stable if

18/2 Yao
∞
<
∑
+∞
−∞
=
k
k
h ]
[ . (2.26)
So discrete-time LTI system is stable is Eq. (2.26) is satisfied.
The similar analysis applies to continuous-time LTI systems, for which the stability is equivalent
to
∞
<
∫
+∞
∞
−
τ
τ d
h )
( . (2.27)
Example: consider a system that is pure time shift in either continuous time or discrete time.
In discrete time, 1
[
]
[ 0 =
−
= ∑
∑
+∞
−∞
=
+∞
−∞
= k
k
n
n
k
h δ ,
while in continuous time, 1
)
(
)
( 0 =
−
= ∫
∫
+∞
∞
−
+∞
∞
−
τ
δ
τ
τ d
t
t
d
h ,
and we conclude that both of these systems are stable.
Example: The accumulator ]
[
]
[ n
u
n
h = is unstable because ∞
=
= ∑
∑
+∞
=
+∞
−∞
= 0
]
[
]
[
k
k
n
u
k
h .
2.3.8 The Unit Step Response of an LTI System
The step response of an LTI system is simply the response of the system to a unit step. It conveys
a lot of information about the system. For a discrete-time system with impulse response ]
[n
h , the
step response is ]
[
]
[
]
[ n
h
n
u
n
s ∗
= . However, based on the commutative property of convolution,
]
[
]
[
]
[ n
u
n
h
n
s ∗
= , and therefore, ]
[n
s can be viewed as the response to input ]
[n
h of a discrete-
time LTI system with unit impulse response. We know that ]
[n
u is the unit impulse response of
the accumulator. Therefore,
∑
−∞
=
=
n
k
k
h
n
s ]
[
]
[ . (2.28)
From this equation, ]
[n
h can be recovered from ]
[n
s using the relation
]
1
[
]
[
]
[ −
−
= n
s
n
s
n
h . (2.29)
It can be seen the step response of a discrete-time LTI system is the running sum of its impulse
response. Conversely, the impulse response of a discrete-time LTI system is the first difference
of its step response.

19/2 Yao
Similarly, in continuous time, the step response of an LTI system is the running integral of its
impulse response,
∫ ∞
−
=
t
d
h
t
s τ
τ )
(
)
( , (2.30)
and the unit impulse response is the first derivative of the unit step response,
)
(
'
)
(
)
( t
s
dt
t
ds
t
h =
= . (2.31)
Therefore, in both continuous and discrete time, the unit step response can also be used to
characterize an LTI system.
2.4 Causal LTI Systems Described by Differential and Difference Equations
This is a class of systems for which the input and output are related through
• A linear constant-coefficient differential equation in continuous time, or
• A linear constant-coefficient difference equation in discrete-time.
2.4.1 Linear Constant-Coefficient Differential Equations
In a causal LTI difference system, the discrete-time input and output signals are related
implicitly through a linear constant-coefficient differential equation.
Let us consider a first-order differential equation,
)
(
)
(
2
)
(
t
x
t
y
dt
t
dy
=
+ , (2.32)
where )
(t
y denotes the output of the system and )
(t
x is the input.
This equation can be explained as the velocity of a car )
(t
y subjected to friction force
proportional to its speed, in which )
(t
x would be the force applied to the car.
In general, an Nth
-order linear constant coefficient differential equation has the form
∑
∑
=
=
=
M
k
k
k
k
N
k
k
k
k
dt
t
x
d
b
dt
t
y
d
a
0
0
)
(
)
(
, (2.33)

20/2 Yao
The solution of the differential equation can be obtained if we have the N initial conditions (or
auxiliary conditions) on the output variable and its derivatives.
Recall that the solution to the differential equation is the sum of the homogeneous solution of the
differential equation 0
)
(
0
=
∑
=
N
k
k
k
k
dt
t
y
d
a (a solution with input set to zero) and of a particular
solution (a function that satisfy the differential equation).
Forced response of the system = particular solution (usually has the form of the input signal)
Natural response of the system = homogeneous solution (depends on the initial conditions and
forced response).
Example: Solve the system described by )
(
)
(
2
)
(
t
x
t
y
dt
t
dy
=
+ . Given the input is )
(
)
( 3
t
u
Ke
t
x t
= ,
where K is a real number.
As mentioned above, the solution consists of the homogeneous response and the particular
solution:
)
(
)
(
)
( t
y
t
y
t
y p
h +
= , (2.34)
where the particular solution )
(t
yp satisfies )
(
)
(
2
)
(
t
x
t
y
dt
t
dy
=
+ and homogenous solution )
(t
yh
satisfies
0
)
(
2
)
(
=
+ t
y
dt
t
dy
. (2.35)
For the particular solution for 0
>
t , )
(t
yp is a signal that has the same form as )
(t
x for 0
>
t ,
that is
t
p Ye
t
y 3
)
( = . (2.36)
Substituting )
(
)
( 3
t
u
Ke
t
x t
= and t
p Ye
t
y 3
)
( = into )
(
)
(
2
)
(
t
x
t
y
dt
t
dy
=
+ , we get
t
t
t
Ke
Ye
Ye 3
3
3
2
3 =
+ , (2.37)
Canceling the factor t
e3
on both sides, we obtain 5
/
K
Y = , so that
t
p e
K
t
y 3
5
)
( = , 0
>
t (2.38)

21/2 Yao
To determine the natural response )
(t
yh of the system, we hypothesize a solution of the form of
an exponential,
st
h Ae
t
y =
)
( . (2.39)
Substituting Eq. (3.38) into Eq. (3.35), we get
0
2 =
+ st
st
Ase
Ase , (2.40)
which holds for 2
−
=
s . With this value of s, t
Ae 2
−
is a solution to the homogeneous equation
0
)
(
2
)
(
=
+ t
y
dt
t
dy
for any choice of A.
Combining the natural response and the forced response, we get the solution to the differential
equation )
(
)
(
2
)
(
t
x
t
y
dt
t
dy
=
+ :
t
t
p
h e
K
Ae
t
y
t
y
t
y 3
2
5
)
(
)
(
)
( +
=
+
= −
, 0
>
t (2.41)
Because the initial condition on )
(t
y is not specified, so the response is not completely
determined, as the value of A is not known.
For causal LTI systems defined by linear constant coefficient differential equations, the initial
conditions are always 0
)
0
(
...
)
0
(
)
0
( 1
1
=
=
=
= −
−
N
N
dt
dy
dt
dy
y , which is called initial rest.
For this example, the initial rest implies that 0
)
0
( =
y , so that
5
0
5
)
0
(
K
A
K
A
y −
=
⇒
=
+
= , the
solution is
)
(
5
)
( 2
3 t
t
e
e
K
t
y −
−
= , 0
>
t (2.42)
For 0
<
t , the condition of initial rest and causality of the system implies that 0
)
( =
t
y , 0
<
t ,
since 0
)
( =
t
x , 0
<
t .
2.4.2 Linear Constant-Coefficient Difference Equations
In a causal LTI difference system, the discrete-time input and output signals are related
implicitly through a linear constant-coefficient difference equation.

22/2 Yao
In general, an Nth
-order linear constant coefficient difference equation has the form
∑
∑ =
=
−
=
−
M
k
k
N
k
k k
n
x
b
k
n
y
a
0
0
]
[
]
[ , (2.43)
The solution of the differential equation can be obtained when we have the N initial conditions
(or auxiliary conditions) on the output variable.
The solution to the difference equation is the sum of the homogeneous solution
0
]
[
0
=
−
∑
=
N
k
k k
n
y
a (a solution with input set to zero, or natural response) and of a particular
solution (a function that satisfy the difference equation).
]
[
]
[
]
[ n
y
n
y
n
y p
h +
= , (2.44)
The concept of initial rest of the LTI causal system described by difference equation means that
0
]
[ =
n
x , 0
n
n < implies 0
]
[ =
n
y , 0
n
n < .
Example: consider the difference equation
]
[
]
1
[
2
1
]
[ n
x
n
y
n
y =
−
− , (2.45)
The equation can be rewritten as
]
[
]
1
[
2
1
]
[ n
x
n
y
n
y +
−
= , (2.46)
It can be seen from Eq. (2.46) that we need the previous value of the output, ]
1
[ −
n
y , to
calculate the current value.
Suppose that we impose the condition of initial rest and consider the input
]
[
]
[ n
K
n
x δ
= . (2.47)
Since 0
]
[ =
n
x for 1
−
≤
n , the condition of initial rest implies that 0
]
[ =
n
y , for 1
−
≤
n , so that
we have as an initial condition: 0
]
1
[ =
−
y . Starting from this initial condition, we can solve for
successive values of ]
[n
y for 0
≥
n :
K
x
y
y =
+
−
= ]
0
[
]
1
[
2
1
]
0
[ ,

23/2 Yao
K
x
y
y
2
1
]
1
[
]
0
[
2
1
]
1
[ =
+
= ,
K
x
y
y
2
2
1
]
2
[
]
1
[
2
1
]
2
[ 





=
+
= ,
K
x
y
y
3
2
1
]
3
[
]
2
[
2
1
]
3
[ 





=
+
= ,
…
K
n
x
n
y
n
y
n






=
+
−
=
2
1
]
[
]
1
[
2
1
]
[ .
Since for an LTI system, the input-output behavior is completely characterized by its impulse
response. Setting 1
=
K , , ]
[
]
[ n
n
x δ
= we see that the impulse response for the system is
]
[
2
1
]
[ n
u
n
h
n






= . (2.48)
Note that the causal system in the above example has an impulse response of infinite duration. In
fact, if 1
≥
N in Eq. (2.43), the difference equation is recursive, it is usually the case that the LTI
system corresponding to this equation together with the condition of initial rest will have an
impulse response of infinite duration. Such systems are referred to as infinite impulse response
(IIR) systems.
2.4.3 Block Diagram Representations of 1st
-order Systems Described by Differential and
Difference Equations
Block diagram interconnection is very simple and nature way to represent the systems described
by linear constant-coefficient difference and differential equations.
For example, the causal system described by the first-order difference equation is
]
[
]
1
[
]
[ n
bx
n
ay
n
y =
−
+ . (2.49)
It can be rewritten as
]
[
]
1
[
]
[ n
bx
n
ay
n
y +
−
−
=
The block diagram representation for this discrete-time system is show:

24/2 Yao
+
D
]
[n
x
]
1
[ −
n
y
]
[n
y
b
a
−
Three elementary operations are required in the block diagram representation: addition,
multiplication by a coefficient, and delay:
]
[
1 n
x +
]
[
2 n
x
]
[
]
[ 2
1 n
x
n
x +
adder
multiplication by
a coefficient
a
]
[n
ax
]
[n
x
]
[n
x D ]
1
[ −
n
x
a unit delay
Consider the block diagram representation for continuous-time systems described by a first-order
differential equation:
)
(
)
(
)
(
t
bx
t
ay
dt
t
dy
=
+ . (2.48)
Eq. (2.48) can be rewritten as
)
(
)
(
1
)
( t
bx
a
b
dt
t
dy
a
t
y +
−
= .
Similarly, the right-hand side involves three basic operations: addition, multiplication by a
coefficient, and differentiation:

25/2 Yao
+
D
dt
t
dy )
(
)
(t
y
a
b /
a
/
1
−
)
(t
x
+
)
(
2 t
x
)
(
)
( 2
1 t
x
t
x +
)
(
1 t
x
adder
multiplication by
a coefficient
a
)]
(t
ax
)
(t
x
dt
t
dx )
(
)
(t
x D
differentiator
However, the above representation is not frequently used or the representation does not lead to
practical implementation, since differentiators are both difficult to implemented and extremely
sensitive to errors and noise.
An alternative implementation is to used integrators rather than the differentiators. Eq. (2.48) can
be rewritten as
)
(
)
(
)
(
t
ay
t
bx
dt
t
dy
−
= , (2.49)
integrating from ∞
− to t , and assuming 0
)
( =
−∞
y , then we obtain
[ ] τ
τ
τ d
ay
bx
t
y
t
∫ ∞
−
−
= )
(
)
(
)
( . (2.50)
In this form, the system can be implemented using the adder and coefficient multiplier, together
with an integrator, as shown in the figure below.

26/2 Yao
τ
τ d
x
t
∫ ∞
−
)
(
)
(t
x
integrator
∫
)
(t
x +
a
−
b
∫ )
(t
y
The integrator can be readily implemented using operational amplifiers, the above
representations lead directly to analog implementations. This is the basis for both early analog
computers and modern analog computation systems.
Eq. (2.50) can also express in the form
[ ] τ
τ
τ d
ay
bx
t
y
t
y
t
t
∫ −
+
=
0
)
(
)
(
)
(
)
( 0 , (2.51)
where we consider integrating Eq. (2.50) from a finite point in time 0
t . It makes clear the fact
that the specification of )
(t
y requires an initial condition, namely )
( 0
t
y .
Any higher-order systems can be developed using the block diagram for the simplest first-order
differential and difference equations.

Unit 2 signal &system

More Related Content

What's hot (20)

Similar to Unit 2 signal &system (20)

Recently uploaded (20)