CH 2

Time-domain representations of
linear time-invariant systems

Chapter 2
Outline
Introduction
The Convolution Sum
Convolution Sum Evaluation Procedure
The Convolution Integral
Convolution Integral Evaluation Procedure
Interconnection of LTI Systems
Relations between LTI Systems Properties and the
Impulse Response
Step Response
Broadband Mobile Wireless

Research Laboratory
Outline
Differential and Difference Equation Representations

of LTI Systems
Solving Differential and Difference Equations
Characteristics of Systems Described by Differential
and Difference Equations
Block Diagram Representations
State-Variable Descriptions of LTI Systems
Exploring Concepts with MATLAB
Summary

Research Laboratory
Introduction
Time-domain representation:
I/P and O/P signals are represented as functions

of time
Four representations
Impulse response
Linear constant-coefficient differential or difference
equation
Block diagram
State-variable description

Research Laboratory
Introduction
Impulse response:
The O/P of an LTI system due to a unit impulse signal I/P

applied at time t=0 or n=0
Completely characterize the behavior of any LTI system
Given the impulse response,
I/P signal: a weighted superposition of time-shifted impulses

O/P signal: a weighted superposition of time-shifted impulses
response by
Linearity and time invariance
Weighted superposition of time-shifted impulses
Continuous-time system: convolution integral

Discrete-time system: convolution sum

Research Laboratory
Impulse Response of LTI System
x[n]=d[n]
y[n]=h[n]
h[n]
k
k
x[n]
y[n]
h[n]
n
Introduction
Characterizing the I/P-O/P behavior of LTI

system
Continuous-time system:
Linear constant-coefficient differential equation
M
dk
dk
ak k y(t ) bk k x(t ), ak , bk are constant coefficients
dt
dt
k 0
k 0
N
Discrete-time system:
Linear constant-coefficient difference equation
a y[n k ] b x[n k ], a , b
k 0

Research Laboratory
k 0
are constant coefficients
Introduction
Block diagram
Represents the system as an interconnection of three

elementary operations
Scalar multiplication
Addition
Time shift (DT) or integration (CT)

Research Laboratory
Introduction
State variable description
State
A set of variables associated with energy storage or

memory devices in the system
A series of coupled first-order differential or

difference equations that represents the systems
state
An equation that relates that state to the O/P of
the system

Research Laboratory
The Convolution Sum

Express x[n] as weighted sum of time-shifted impulses:
x[n]d [n] x[0]d [n]
x[n]d [n k ] x[k ]d [n k ]
x[2]d [n 2] x[1]d [n 1] x[0]d [ n]
x[n]
x[1]d [n 1] x[2]d [n 2]
x[k ]d [n k ]
Figure 2.1 (p. 99)

Graphical example illustrating the
representation of a signal x[n] as a
weighted sum of time-shifted impulses.

Research Laboratory
10
The Convolution Sum

Let the operator H denote the system to which x[n] is applied
y[n] H {x[n]} H { x[k ]d [n k ]}

k
H {x[k ]d [n k ]} p.63 property of superposition
x[k ]H {d [n k ]} p.63 property of homogeneity
Let h[n] H{d [n]}

y[n]
time-invariant H {d [n k ]} h[n k ]
x[k ]h[n k ] x[n]* h[n],
where * denotes the convolution sum

Research Laboratory
11
x[n]=d[n]
y[n]=h[n]
h[n]
k
k

Research Laboratory
12
d[n]
h[n]
h[n]
x[0]
n
n
y[n]
x[n]
2h[n-1]
d[n-1]
x[1]
h[n]
n 2*
d[n-2]
x[2]
1*

Research Laboratory
-h[n-2]
h[n]
n
13
Figure 2.2a (p. 100)

Illustration of the convolution sum. (a) LTI system
with impulse response h[n] and input x[n].

Research Laboratory
14
Figure 2.2b (p. 101)

(b) The decomposition of the input
x[n] into a weighted sum of timeshifted impulses results in an output
y[n] given by a weighted sum of
time-shifted impulse responses.
h[n]
h[n]
h[n]
h[n]
x[n]
h[n]
y[n]
y[1]=Sx[k]h[1-k]
y[n]=Sx[k]h[n-k]
Research Laboratory
h[n]
15
The Convolution Sum
Ex. 2.1 Multipath Communication Channel

A discrete-time LTI system representing a two-path propagation, y[n] x[n]
Let x[n] d [n], we find the impulse response is
1, n 0
1
1
h[n] , n 1
i.e. h[n] d [n] d [n 1]
2
2
0, otherwise
Determine te output of the system in response to the input
2,
4,
x[n]
2,
0,
n0
n 1
n2
otherwise

Research Laboratory
16
1
x[n 1]
2
The Convolution Sum

Sol.
x[n] 2d [n] 4d [n 1] 2d [ n 2]; h[n] d [ n]
1
d [n 1]
2
y[n] 2h[n] 4h[ n 1] 2h[ n 2]

1
1
1
d [ n 1]} 4{d [ n 1] d [ n 2]} 2{d [ n 2] d [ n 3]}
2
2
2
2d [ n] d [ n 1] 4d [ n 1] 2d [ n 2] 2d [ n 2] d [ n 3]
2d [ n] 5d [n 1] d [n 3]
2{d [ n]
0,
2,
5,
0,
1,
0,
n0
n0
n 1
n2
n3
n4

Research Laboratory
17
y[n] is a weighted sum of each time-shifted impulse response
y[n]
x[k ]h[n k ] w [k ],
Eq. (2.6)
where wn [k ] x[k ]h[n k ] is the intermediate signal

Research Laboratory
18
k:, n: , h[n-k]?
Precedence Rule: time-shifting time-scaling
k
3
Ex. h[k ] u[k ]
4
v[ k ] h[ k n]
3

4
y[ k ] v[ k ] h[ n k ]
k n
0 1 2 3

Research Laboratory
u[ k n]
y[k]=h[n-k]
v[k]=h[k+n]
h[k]
3

4
u[ k n]
k n
-n 1-n 2-n 3-n
19
n-3 n-2 n-1 n
Q&A
Q: h[n-k]nk
y[n]
x[k ]h[n k ] w [k ],
Eq. (2.6)

Research Laboratory
20
Ex. 2.2 Convolution Sum by using an Intermediate Signal

3
Consider a system with impulse response h[n] ( ) n u[n]
4
Find the O/P of the system at times n=-5,5,10 when the I/P x[n] u[n]
Figure 2.3 (p. 103)

Evaluation of Eq. (2.6) in Example 2.2. (a) The input
signal x[k] above the reflected and time-shifted
impulse response h[n k], depicted as a function of k.

Research Laboratory
21
y[n]
x[k ]h[n k ] w [k ],
Eq. (2.6)
-5
=-5
Figure 2.3 (p. 103)
(b) The product signal w5[k] used to evaluate y [5].

Research Laboratory
22
y[n]
x[k ]h[n k ] w [k ],
Eq. (2.6)
-5
=5
Figure 2.3 (p. 103)
(c) The product signal w5[k] used to evaluate y[5].

Research Laboratory
23
y[n]
x[k ]h[n k ] w [k ],
Eq. (2.6)
-5
=10
Figure 2.3 (p. 103)
(d) The product signal w10[k] used to evaluate y[10].

Research Laboratory
24

Sol.
3 nk
3 nk
( ) , k n
( ) , 0 k n
h[n k ] 4
, wn [k ] x[k ]h[n k ] h[ n k ]u[ k ] 4
0,
0,
otherwise
otherwise
For n 5, w5 [k ] 0, y[5]
[k ] 0
4 6
3 5 k
1
(
)
5
3 5 k 3 5 5 4 k 3 5
( ) , 0 k 5
3
For n 5, w5 [k ] 4
, y[5] ( ) ( ) ( ) ( )
4 k 0 3
4 1 ( 4)
k 0 4
0,
otherwise
3
4 11
3 10 k
1
(
)
10
3 10 k 3 10
( ) , 0 k 10
3
For n 10, w10 [k ] 4
, y[10] ( )
( )
4 1 ( 4)
k 0 4
0,
otherwise
3
Research Laboratory
25
y[n]
x[k ]h[n k ] w [k ],
Eq. (2.6)
In general, we may determine y[n] for all n without

evaluating Eq. (2.6) at an infinite number of distinct
shifts n
It is helpful to graph both x[k] and h[n-k] in
determining wn[k] and identifying the appropriate
intervals of shifts

Research Laboratory
26

Procedure 2.1: Reflect and Shift convolution Sum Evaluation
1. h[k ] reflect h[ k ] about k 0 h[k ] shift by -n h[n k ]
2. Begin with n large and negative, shift h[n k ] to the far left
3. Write wn [k ]
4. Move h[n k ] toward the right until the representation for wn [k ] changes
5. Let n be in the new interval, repeat 3 & 4
6. For each interval of time shifts, sum all wn [k ] to obtain y[n]

Research Laboratory
27
1
Ex. 2.3 Moving Average System: h[n] (u[n] u[n 4])
4
Determine the O/P of the system for x[n] u[n] u[n 10]
Figure 2.4 (p. 106)
Evaluation of the convolution sum for Ex.2.3

(a) The system impulse response h[n].
(b) The input signal x[n].
h[-k]
h[n-k]

Research Laboratory
n-3
28
n: -
Step 3: Shift h[n-k] from the far left
Step 4: Move h[n-k] toward right until the representation for wn[k] is changed
Research Laboratory
29
0
0n 3
Figure 2.4 (p. 106)

(c) The input above the reflected and time-shifted impulse
response h[n k], depicted as a function of k.
(d) The product signal wn[k] for the interval of shifts 0 n 3.

Research Laboratory
30
3<n 9
0
9 < n 12
Figure 2.4 (p. 106)

Research Laboratory
(e) The product signal wn[k] for the interval of shifts 3 < n 9.
(f) The product signal wn[k] for the interval of shifts 9 < n 12.
31
x[k]h[0-k]=w0[k]
n=0
x[k]h[1-k]= w1[k]
0n 3
w2[k]
w3[k]
0123
n=3
w5[k]
3<n 9
w6[k]
=Swn[k]
9 < n 12
Figure 2.4 (p. 106)

(g) The output y[n].

Research Laboratory
32
y[6]=Sw6[k]
=4*1/4=1
0n 3
Sol.
n 0, y[n] 0
1 n 1
4
k 0 4
3<n 9
0 n 3, y[n]
n
1 1
3 n 9, y[n] (n (n 3) 1) 1
4
k n 3 4
13 n
1 1
9 n 12, y[n] (9 (n 3) 1)
4
4
k n 3 4
n 12, y[n] 0
9
9 < n 12
Research Laboratory
33
Linear and Time-Invariance
d [n k ] H h[n k ]
Linear:
d [n] H h[n]
Impulse Response:
Time-Invariance:
Superposition
Homogeneity
a d [n k ]
k
H ak h[n k ]
k
LTI System
x[n]
x[k ]d [n k ]
H y[n]

Research Laboratory
x[k ]h[n k ] w [k ]
34
y[n] x[n] x[n 1], find h[n]?

x[n] H y[n]
d [n] H h[n]
Let x[n] d [n], then y[n] h[n]
h[n] d [n] d [n 1]

Research Laboratory
35
Ex. 2.4 First-Order Recursive System: Reflect-and-Shift Convolution

y[n] y[n 1] x[n]
Let x[n] b nu[n 4]; assume b and the system is causal. Find y[n]
Figure 2.5a&b (p. 109)

Evaluation of the convolution sum for
Example 2.4. (a) The input signal x[k]
depicted above the reflected and timeshifted impulse response h[n k].
(b) The product signal wn[k] for 4 n.

Research Laboratory
36
Sol.
Let x[n] d [n], we have h[n] h[n 1] d [ n] (2.7)
Since the system is causal, h[ n] 0 for n 0
From (2.7), h[0] 1, h[1] , h[2] 2 ,
b k , 4 k
x[k ]
0, otherwise
nk , k n
h[n k ]
0, otherwise
Research Laboratory
37
, or h[ n] nu[ n]
For n 4, wn [k ] 0, thus, y[n] 0

b k n k , 4 k n
For n 4, wn [k ] x[k ]h[n k ]
0, otherwise
n
n
n4
b k
b
k nk
n
n
and y[n] b ( ) ( ) m 4
k 4
k 4
m0
n 5 b n 5
( ) ( ) b (
)
m 0
b
n
n4
0,
y[n] 4 n 5 b n 5
b ( b ),

Research Laboratory
n 4
n 4
38
Figure 2.5c (p. 110)

(c) The output y[n] assuming
that p = 0.9 and b = 0.8.

Research Laboratory
39
Ex. 2.5 Investment Computation
Refer to the textbook by yourself
Prob. 2.1~2.2

Research Laboratory
40
Convolution Integral
O/P of a continuous-time LTI system is determined from
knowledge of the I/P, and

the systems impulse response
Convolution integral:
x(t )
x( )d (t )d , y(t ) H {x(t )} H { x( )d (t )d }
linear y (t )
x( ) H {d (t )}d
Let h(t ) H {d (t )},
time-invariant H {d (t )} h(t )
y (t )
x( )h(t )d x(t ) * h(t )

Research Laboratory
41
Figure 2.9 (p. 115)
(a) Impulse response of an LTI system H. (b)

The output of an LTI system to a time-shifted
and amplitude-scaled impulse is a time-shifted
and amplitude-scaled impulse response.
d(t)
x()h(t)
x() d(t)
x()

Research Laboratory
42
y (t )
x( )h(t )d w ( )d , where w ( ) x( )h(t )

t
Thus, the system O/P at any time t is the area under the signal wt ( )
Often, it is helpful to graph both x( ) and h(t ) in determining wt ( )

and identifying the appropriate set of shifts.

Research Laboratory
43

Procedure 2.2: Reflect-and-Shift Convolution Integral Evaluation
1. h( ) reflect h ( ) about 0 h( ) shift by -t h(t )
2. Begin with t large and negative, shift h(t ) to the far left
3. Write wt ( )
4. Move h(t ) toward the right until the representation for wt ( ) changes
5. Let t be in the new set, repeat 3 & 4
6. For each set of shifts t , integrate all wt ( ) from to to obtain y (t )

Research Laboratory
44
x( )
x( )

Research Laboratory
45
Ex. Reflect-and-shift convolution evaluation
x(t ) u (t 1) u (t 3), h(t ) u (t ) u (t 2).

Find y(t )
Figure 2.10 (p. 117)

Input signal and LTI system
impulse response for Example 2.6.

Research Laboratory
46
h(-)
-2
-1
h(t-)
t-2

Research Laboratory
t-1
47
X()
X()
X()
h(t-)
h(t-)
h(t-)
wt()
wt()
t 1, wt ( ) 0
t 5, wt ( ) 0
Figure 2.11 (p. 118)

Evaluation of the convolution integral for
Example 2.6. (a) The input x() depicted
above the reflected and time-shifted
impulse response. (b) The product signal
wt() for 1 t < 3. (c) The product signal
Wireless
wtBroadband
() for 3 t < 5.Mobile
(d) The system
output
y(t). Research Laboratory
1 t 3,
3 t 5,
1, 1 t
wt ( )
0, otherwise
1, t 2 3
wt ( )
0, otherwise
48
t 1, and t 5, wt ( ) 0 y(t)=0
t
1 t 3, y(t)= wt ( )d t 1
1
3
3 t 5, y(t)= wt ( ) d 3 (t 2) 5 t
t -2
t 1
0,
t 1, 1 t 3
y (t )
5 t , 3 t 5
0,
t 5
Research Laboratory
49
Ex. 2.7 RC circuit output

Find y(t ) if RC 1 and x(t ) u(t ) u(t 2)
Figure 2.12 (p. 119)
RC circuit system with the voltage source
x(t) as input and the voltage measured
across the capacitor y(t), as output.
Sol.
From Ex. 1.21, we have h(t ) et u (t ),
Research Laboratory
50
( )
( )
Figure 2.13 (p. 120)
( )
( )
0 t 2,
2 t,
e (t ) , 0 t
wt ( )
0, otherwise
e (t ) , 0 2
wt ( )
0, otherwise
y (t ) e
0
( t )

Example 2.7. (a) The input x()
superimposed over the reflected and
time-shifted impulse response h(t ),
depicted as a function of . (b) The
product signal wt() for 0 t < 2. (c) The
product signal wt() for t 2. (d) The
system output y(t).
d 1 e

Research Laboratory
y (t ) e (t ) d (e 2 1)e t
0
51
t0
wt ( ) 0, y(t ) 0
Ex. 2.8 Another reflect-and-shift convolution evaluation
x(t ) (t 1)[u (t 1) u(t 3)], h(t ) u (t 1) 2u(t 2).

Find y(t )
h()
-1
-2
2
h(t)
-1

Research Laboratory
h(t)
52
( )
( )
( )
( )
t 0,
wt ( ) 0
y (t ) 0

53
Research Laboratory
( )
0 t 2,
1, 1 t 1
wt ( )
o.w.
0,
t 1
t2
y (t ) ( 1)d
1
2
2 t 3,
1, 1 3
wt ( )
o.w.
0,
3
y (t ) ( 1)d 2
1
( )
( )
( )
( )
3 t 5,
5 t,
( 1), 1 t 2
wt ( ) 1,
t -2 3
0,
o.w.
( 1), 1 3
wt ( )
o.w.
0,
y (t )
t 2
y (t ) ( 1)d 2
1
( 1)d ( 1)d
Broadband Mobile
Wireless
t 2 6t 7
Research Laboratory
t 2
54
x()
x()
h(t-)
h(t-)
wt()
x()
h(t-)
h(t-)
x()
h(t-)
x()
wt()
wt()
wt()
Figure 2.14 (p. 121)

Example 2.8. (a) The input x() superimposed on
the reflected and time-shifted impulse response
h(t ), depicted as a function of . (b) The
product signal wt() for 0 t < 2. (c) The
product signal wt() for 2 t < 3. (d) The
product signal wt() for 3 t < 5. (e) The
Broadband
product
signal wt()Mobile
for t 5. Wireless
The system
output yResearch
(t).
Laboratory
55
Ex. 2.9 Radar range measurement: Propagation model
sin( wc t ), 0 t T0
x(t )
otherwise
0,
Suppose we transmit an impulse from the radar. The impulse is delayed
in time and attnuated in amplitude, which results in the impulse response of
h(t ) ad (t ).
Use the convolution of x(t ) with h(t ) to verify this result.
Figure 2.16 (p. 124)
Radar range measurement. (a)
Transmitted RF pulse. (b) The
received echo is an attenuated
and delayed version of the
transmitted pulse.

Research Laboratory
56
x(t)
r(t)
h(t)
d (t ) d (t )
Sol.
h( ) ad ( ), h( ) ad ( ) ad ( )
h(t ) ad ( t ) ad ( (t )),
r (t ) x( )h(t )d ax(t ),
which is an attnuated and delayed version of the transmitted signal
= =
h()
a
a
0

Research Laboratory
57
h(-)
h(t-)
Ex. 2.10 Radar range measurement: The matched filter
The impulse response of the matched filter is a reflected, or time-reversed,

version of the transmitted signal x(t ). That is, hm (t ) x(t ), so
sin( wc t ), - T0 t 0
hm (t )
otherwise
0,
Find y (t ) r (t ) * hm (t )
Figure 2.17a (p. 125)
(a) Impulse response of the matched filter
for processing the received signal.

Research Laboratory
58
Reference: Textbook p.126~127

Sol.
hm (t ) is a reflected version of x(t ), hm (t ) x( t )
First, we form wt ( ) r ( )hm (t ) ax( ) x( t )
For t T0 , t T0 , wt ( ) 0 y (t ) 0
a sin(c ( )) sin(c ( t )), t T0
For T0 t , wt ( ) r ( )hm (t )
otherwise
0,
T0 a
a
a
y (t ) [ cos(c (t ))+ cos(c (2 t ))]d [t ( T0 )]cos(c (t ))
2
2
2
a sin(c ( )) sin(c ( t )), t T0
For t T0 , wt ( )
otherwise
0,
y (t )
T0
a
a
a
[ cos(c (t ))+ cos(c (2 t ))]d [ t T0 )]cos(c (t ))
2
2
2

Research Laboratory
59
( )
(t )
Figure 2.17b (p. 126)
(b) The received signal r() superimposed on the reflected and

time-shifted matched filter impulse response hm(t ), depicted
as functions of . (c) Matched filter output x(t).

Research Laboratory
60
Interconnections of LTI Systems
Parallel connection of LTI systems

y (t ) y1 (t ) y2 (t ) x(t ) * h1 (t ) x(t ) * h2 (t )
x( )h1 (t )d x( )h2 (t )d
x( )[h1 (t ) h2 (t )]d x( )h(t )d

x(t ) * h(t ), where h(t ) h1 (t ) h2 (t )
Figure 2.18 (p. 128)

Research Laboratory
61
Interconnection of two LTI systems.

(a) Parallel connection of two
systems. (b) Equivalent system.
Property of Convolution
Convolution possesses the distributive property:

x(t ) * h1 (t ) x(t ) * h2 (t ) x(t ) *{h1 (t ) h2 (t )}
For discrtet case,
x[n]* h1[n] x[n]* h2 [n] x[n]*{h1[n] h2 [n]}

Research Laboratory
62
Cascade connection of LTI systems
y (t ) z (t ) * h2 (t ) z ( )h2 (t )d
[ x( ) * h1 ( )]h2 (t ) d [ x( ) h1 ( ) d ]h2 (t ) d
x( )[ h1 ( ) h2 (t )d ]d
Let , y (t ) x( )[ h1 ( )h2 (t )d ]d
x( )h(t ) d
x(t ) * h(t ),
where h(t ) h1 ( )h2 (t )d
h(t ) h1 ( )h2 (t )d h1 (t ) * h2 (t )

Research Laboratory
63
Cascade connection of LTI systems
h(t ) h1 (t )* h2 (t )
h(t ) h1 ( )h2 (t )d
h2 ( )h1 (t )d h2 (t )* h1 (t )
t -
Figure 2.19 (p. 128)

Interconnection of two LTI systems. (a)
Cascade connection of two systems.
(b) Equivalent system. (c) Equivalent
system: Interchange system order.

Research Laboratory
64
Property of Convolution
For continuous case,

Associative property:
{x(t ) * h1 (t )}* h2 (t ) x(t ) *{h1 (t ) * h2 (t )}
Commutative property:
h1 (t ) * h2 (t ) h2 (t ) * h1 (t )
For discrtet case,

Associative property:
{x[n]* h1[n]}* h2 [n] x[ n]*{h1[n]* h2 [n]}
Commutative property:
h1[n]* h2 [n] h2 [n]* h1[n]}

Research Laboratory
65
Ex. 2.21 Equivalent system to four interconnected systems
h1[n] u[n], h2 [n] u[n 2] u[n]

h3[n] d [n 2], h4 [n] nu[n]
Find the impulse response h[n] of the overall system.
Figure 2.20 (p. 131)

Interconnection of systems for Example 2.11.

Research Laboratory
66
Figure 2.21 (p. 131)

(a) Reduction of parallel combination
of LTI systems in upper branch of Fig.
2.20. (b) Reduction of cascade of
systems in upper branch of Fig. 2.21(a)
(c) Reduction of parallel combination
of systems in Fig. 2.21(b) to obtain an
equivalent system for Fig. 2.20.

Research Laboratory
67
h1[n] u[n], h2 [n] u[n 2] u[n]

h3[n] d [n 2], h4 [n] nu[n]
Find the impulse response h[n] of the overall system.
Sol.
h[n] (h1[n] h2 [n]) * h3[ n] h4 [ n]
(u[n] u[n 2] u[n]) * d [n 2] nu[n]
u[n 2]* d [n 2] nu[n]
u[n] nu[n] (1 n )u[n]
II :
x[n] u[n 2], h[n] d [n 2],
y[n] x[n]* h[n]
x[k ]h[n k ]
u[k 2]d [(n k ) 2] u[(n 2) 2] u[n]

Research Laboratory
68
Prob. 2.8
h(t ) [h1 (t )* h4 (t ) h2 (t ) h3 (t )]* h5 (t )
Figure 2.22 (p. 132)

Interconnection of LTI systems for Problem 2.8.

Research Laboratory
69
Often, it is easy to break a complex system

into simpler subsystems, analyze each
subsystem, and then study the entire system
as an interconnection of subsystems.
This is an example of divide-and-conquer
approach to problem solving and is possible
due to the assumptions of linearity and time
invariance

Research Laboratory
70

Research Laboratory
71

Research Laboratory
72
Relations between LTI system properties and

the impulse response
The impulse response completely characterizes

the I/P-O/P behavior of an LTI system
Properties of the system, such as memory,
causality, and stability, are related to the
systems impulse response
Memoryless LTI system

Stable LTI system
Invertible system and impulse response

Research Laboratory
73
Memoryless LTI systems
O/P of a memoryless LTI system depends only on

the current I/P
y[n] h[n]* x[n]
h[k ]x[n k ]
h[2]x[n 2] h[1]x[n 1]
h[0] x[ n] h[1] x[ n 1] h[2] x[ n 2]

For memoryless system, y[n] must depend only on x[n] h[k ] 0 for k 0
A discrete-time LTI system is memoryless if and only if

h[ n] cd [ n], where c is an arbitrary constant
A continuous-time LTI system is memoryless if and only if

h(t ) cd (t ), where c is an arbitrary constant
All memoryless LTI system simply perform scalar multiplication on the I/P

Research Laboratory
74
Causal LTI systems
O/P of a causal LTI system depends only on past or present

values of the I/P
y[n] h[n]* x[n]
h[k ]x[n k ]
h[2] x[ n 2] h[1] x[ n 1] h[0] x[ n] h[1] x[ n 1]

For causal system, y[n] must depend only on past or present value of x[n]
h[k ] 0 for k 0.
Hence, the convolution sum takes the new form: y[n] h[k ]x[n - k ]
k 0
A discrete-time LTI system is causal if h[n] 0 for n 0

A continuous-time LTI system is causal if h(t ) 0 for t 0.
And the convolution integral is y (t ) h( )x(t ) d

0
Causal systems are nonanticipatory, no O/P is generated before the I/P is applied

Research Laboratory
75
Stable LTI systems
A system is BIBO stable if the O/P is guaranteed to be bounded

for every bounded I/P
y[n] h[n]* x[n]
h[k ]x[n k ]
h[ k ] x[ n k ] M x
h[k ]x[n k ]
h[ k ]
For stable system, y[n] <, hence,

A discrete-time LTI system is stable iff
h[k ]
A continuous-time LTI system is stable iff

Research Laboratory
76
h( ) d
Ex. 2.12 Properties of the first-order recursive system
y[n] y[n 1] x[n], h[n] nu[n]

Is the system causal, memoryless, and BIBO stable?
Prob. 2.10

Research Laboratory
77
Sol.
The system is causal, since h[n] 0 for n 0
The system is not memoryless because h[n] 0 for n 0
k 0
k 0
h[n] k iff 1 Stable!

Research Laboratory
78
Stable LTI systems
A system can be unstable even through the impulse

response has a finite value
Ex.
t
y (t ) x( )d
For x(t ) d (t ) h(t ) u (t )

h(t ) is never greater than unity, but is not abaulutely integrable.
Thus, the system is unstable.
An ideal accumulator: y[n]
x[k ]
For x[n] d [n] h[n] u[n], which is not absolutably summable.

It's not stable.
Research Laboratory
79
Invertible systems and deconvolution
A system is invertible if the I/P to the system can be recovered from

the O/P except for a constant scale factor. It implies the existence of
an inverse system, denoted as hinv(t).
The process of recovering x(t) from h(t)*x(t) is termed deconvolution
An inverse system performs deconvolution
A common problem is that of reversing or equalizing the distortion

introduced by a non-ideal system
Ex. Telephone line
Equalizer reverses the distortion and permits much higher data rate

Research Laboratory
80
Invertible systems and deconvolution
The relationship between the impulse response of an LTI system,

h(t), and that of the corresponding inverse system, hinv(t), is easily
derived.
The continuous-time LTI inverse system, hinv (t ), must satisfy

x(t )*(h(t )* hinv (t )) x(t ) h(t )* hinv (t ) d (t )
(2.30)
The discrete-time LTI inverse system, hinv [n], must satisfy

h[n]* hinv [ n] d [n]
(2.31)
Figure 2.24 (p. 137)

Cascade of LTI system with
impulse response h(t) and
inverse system with impulse
response h-1(t).

Research Laboratory
81
Ex. 2.13 Multipath communication channels

A discrete-time model for a two-path communication
channel is y[n]=x[n]+ax[n-1]
a. Find a causal inverse system that recovers x[n]
from y[n]
b. Check whether the system is stable

Research Laboratory
82
1, n 0
Sol. h[n] d [n] ad [n 1], h[n] a, n 1

0, otherwise
The inverse system hinv [n] must satisfy h[n]* hinv [n] d [n],
y[n]=x[n] + ax[n-1]
d [n] hinv [n] ahinv [n 1]

*hinv [ n ]
For causal system, hinv [n] 0 for n 0

For n 0, hinv [0] ahinv [1] 1 hinv [0] 1
For n 0, hinv [n] ahinv [n 1] 0 hinv [n] ahinv [ n 1]
hinv [1] a, hinv [2] a 2 , hinv [3] a 3 ,
For stability, we have to check
hinv [n] ( a) n u[ n]
h [n] a k if a 1
inv
k 0
inverse system is stable if the multipath component ax[n 1] is weaker than x[n]
Research Laboratory
83

h[n] d [n] ad [n 1]
hinv [n] (a) n u[n]
h[n]* hinv [n]
inv
[k ]h[n k ]
(a ) k u[k ](d [n k ] ad [n k 1])
(a) k d [n k ] - (a) k 1d [n k 1])

k 0
(a 0d [n] a1d [n 1] a 2d [n 2]
d [ n]
Research Laboratory
84
) - (a1d [n 1] a 2d [n 2]
In general, it is difficult to obtain the inverse system

by directly solving (2.30) an (2.31)
Not every LTI system has a stable and causal inverse

Research Laboratory
85

Research Laboratory
86
Step response
Step I/P signal are often used to characterize the response of an

LTI system to sudden changes in the I/P
Step response: O/P due to a unit step I/P signal
h[n] : impulse response

s[n] : step response
s[n] h[n]* u[n]
h[k ]u[n k ] h[k ]
0, for k n
u[n k ]
1, for k n
t
For continuous-time system: s(t ) h( ) d

Research Laboratory
87
Step response
Express the impulse response in terms of the step

response
For discrete-time system:
s[n]
h[k ], h[n] s[n] s[n 1]
For continuous-time system:

t
d
s (t ) h( )d , h(t ) s(t )
dt

Research Laboratory
88
Ex. RC circuit: step response

1 RCt
h(t )
e u (t ), find the step response of the system
RC
Figure 2.12 (p. 119)
RC circuit system with the voltage source
x(t) as input and the voltage measured
across the capacitor y(t), as output.
Prob. 2.12, 2.13, 2.50

Research Laboratory
89
Sol.
t 0 0,
0,
t 0
t
1 RC
s (t )
e u ( )d 1 t
RC
RC
e d , t 0 1 e RC , t 0
0
RC
Figure 2.25 (p. 140)

RC circuit step response for RC = 1 s.

Research Laboratory
90
The homogeneous solution

The homogeneous solution of a differential/difference equation is
obtained by setting all terms involving the I/P to zero.
N
d k (h)
(h)
For a continuous-time system: ak k y (t ) 0 y (t ) ci e rit ,
dt
k 0
i 1
N
where ri are the N roots of the characteristic equation
k
a
r
k 0
k 0
For a discrete-time system: ak y [n k ] 0 y [n] ci ri n ,

(h)
(h)
k 0
i 1
where ri are the N roots of the characteristic equation
N k
a
r
k 0
k 0
ci are determined such that the complete solution satisfy initial conditions
Research Laboratory
102
Ex. 2.17 RC circuit: homogeneous solution

Determine the homogeneous solution of the system
d
y(t ) RC y (t ) x(t )
dt
Figure 2.30 (p. 148)
RC circuit.

Research Laboratory
103
Sol.
The homogeneous equation is
d
y (t ) 0, y ( h ) (t ) c1e r1t (V ),
dt
where r1 is the root of the characteristic equation
y (t ) RC
t
1
(h)
RC
1 RCr1 0. Hence, r1
, y (t ) c1e
(V )
RC

Research Laboratory
104
Ex. 2.18 First-order recursive system:

Homogeneous Solution
Determine the homogeneous solution of the system
y[n] y[n 1] x[n]
Prob. 2.16, 2.17

Ex. 2.21

Research Laboratory
105
Sol.
The homogeneous equation is
y[n] y[n 1] 0, y ( h ) [n] c1r1n ,
where r1 is the root of the characteristic equation
r1 0. Hence, r1 , y ( h ) [n] c1 n
Ex. 2.21

Research Laboratory
106
Prob. 2.16
(a ) r 2 5r 6 0
(b) r 2 3r 2 0
9
0
16
1
0
4
( c ) N 2, r 2
( d ) N 2, r 2

Research Laboratory
107
(a ) r 2 5r 6 0
(r+3)(r+2)=0, r=-3, -2
(b) r 2 3r 2 0
( c ) N 2, r 2
( d ) N 2, r 2

Research Laboratory
9
0
16
1
0
4
108
The particular solution
The particular solution y(p) represents any

solution of the differential/difference equation
for the given I/P
y(p) is not unique

Usually obtained by assuming an O/P of the same
general form as the I/P
x[n]=n, y(p)[n]=c n
x[n]=Acos(Wn+f), y(p)[n]= c1cos(Wn)+ c2sin(Wn)
The above mentioned method is modified if the I/P

is of the same form of the homogeneous solution
y(p) must be indep. of all terms of y(h)

y(p)=t y(h) or y(p)=n y(h)

Research Laboratory
109

Research Laboratory
110

Particular Solution
Determine the particular solution of the system
1 n
y[n] y[n 1] x[n], if the I/P is x[n] ( )
2
Prob. 2.18
Ex. 2.21

Research Laboratory
111
1
Sol. We assume a particular solution of the form y ( p ) [n] c p ( ) n ,
2
1
1
1
c p ( ) n c p ( ) n 1 ( ) n c p (1 2 ) 1
2
2
2
1
1
1
y ( p ) [ n]
( ) n
1 2 2
2
1
Since y ( h ) [ n] c1 n , y ( p ) [ n] has the same form of y ( h ) [n] if .
2
In this case, c p is not existed
1
Hence, we must assume y ( p ) [ n] c p n( ) n ,
2
1
1
1
c p n( ) n c p ( n 1)( ) n 1 ( ) n c p n(1 2 ) 2 c p 1
2
2
2
1
1
Replace by , we have c p 1 y ( p ) [n] n( ) n ,
2
2
Research Laboratory
101
Ex. 2.21
Ex. 2.20 RC circuit: particular solution

Determine the particular solution of the system
d
y(t ) RC y(t ) x(t ) with an I/P x(t ) cos(0t )
dt
Figure 2.30 (p. 148)

RC circuit.

Research Laboratory
102
Sol.
We assume a particular solution of the form
y ( p ) (t ) c1 cos(0t ) c2 sin(0t ),
then we have
[c1 cos(0t ) c2 sin(0t )] RC[0 c1 sin(0t ) 0 c2 cos(0t )] cos(0t )
c1 RC0 c2 1, and RC0 c1 c2 0
c1
RC0
1
,
an
d
c
2
1 ( RC0 ) 2
1 ( RC0 ) 2
y ( p ) (t )
RC0
1
cos(
t
)
sin(0t ) (V)
0
2
2
1 ( RC0 )
1 ( RC0 )

Research Laboratory
114
(a) y
( p)
1
(t ) c p e , c p e 5c p e 6c p e 2e e , 2c p 1, c p
2
t
(b) y ( p ) (t ) c1 cos(2t ) c2 sin(2t ),
d2
d
d
c
cos(2
t
)
c
sin(2
t
)
3
c
cos(2
t
)
c
sin(2
t
)
2
c
cos(2
t
)
c
sin(2
t
)
cos(2
t
)
cos(2t )
1
2
1
2
1
2
dt 2
dt
dt
4c1 cos(2t ) 4c2 sin(2t ) 3 2c1 sin(2t ) 2c2 cos(2t ) 2 c1 cos(2t ) c2 sin(2t ) cos(2t ) 2sin(2t )
1
1
cos(2t ) : 4c1 6c2 2c1 1, sin(2t ) : 4c2 6c1 2c2 1, c1 , c2
4
4

Research Laboratory
115
9
7
32
c 2,
c 2, c
16
16
7
1 n 1 n 1 1 n2
1 n
1 n2
( p)
(d ) y [n] c( ) , c( ) c( ) ( ) 2( ) ,
2
2
4 2
2
2
1 n
1 n
9
(c c)( ) (1 8)( ) , c
2
2
2
(c) y ( p ) [n] c,

Research Laboratory
116
The complete solution

Procedure 2.3: Solving a Differential or Difference Equation
1. Find a homogeneous sloution y ( h ) from roots of
characteristic equation
2. Find a particular solution y ( p ) by assuming that it has the
same from as the I/P, yet is independent of all terms of y ( h )
3. Determine the coefficient in the homogeneous solution so
that the complete solution y y ( h ) y ( p ) satisfying the
initial conditions
Research Laboratory
117
Assume I/P applied at t=0 or n=0
Particular solution applied only to times t,n0
Discrete-time system
y[-N],,y[-1] must be translated to y[0],,y[N-1]
Continuous-time system
Initial conditions at t=0- must be translated to t=0+

y(k)(t=0-) = y (k)(t=0+), a necessary and sufficient condition
for this is Sbkx(k)(t) contains no impulses or derivatives of
impulse

Research Laboratory
118

Complete solution
Determine the complete solution of the system
1
1 n
y[n] y[n 1] x[n], if the I/P is x[n] ( )
4
2
and initial condition y[1] 8

Research Laboratory
119
1
1
( ) n (2.47)
1 2 2
Ex. 2.18
Ex. 2.19
1
1 n
1 n
since , y[n] c1 ( ) 2( ) , n 0
4
4
2
we first translate the initial condition to time n 0
1
1
y[0] x[0] y[1] 1 8 3
4
4
1
1
From (2.47), we have 3=c1 ( )0 +2( )0 c1 1
4
2
1
1
y[n] ( ) n 2( ) n , n 0
4
2
Sol. y[n] y ( h ) [n] y ( p ) [n] c1 n

Research Laboratory
120
Summary
Ex. 2.22 RC circuit: complete solution

Determine the complete solution of the system
d
y (t ) x(t ) with an I/P x(t ) cos(t )u (t ) (V )
dt
R 1W, C 1F , y (0 ) 2 (V )
y (t ) RC
Figure 2.30 (p. 148)

RC circuit.

Research Laboratory
121
Sol.
From Ex. 2.17, y ( h ) (t ) ce
From Ex. 2.20, y ( p ) (t )
t
RC
(V )
1
RC
cos(
t
)
sin(t ) (V )
2
2
1 ( RC )
1 ( RC )
t
RC
1
1
cos(t ) sin(t ) (V ), t 0
2
2
The I/P does not introduce impulses into the right-hand side of the differential equation.
y (t ) ce
Thus, the coefficient c is determined from the initial condition y (0 ) y (0 ) 2
0
RC
1
1
1
3
2 ce
cos(0 ) sin(0 ) c c
2
2
2
2
3 RCt 1
1
y (t ) e
cos(t ) sin(t ) (V ), t 0
2
2
2
Research Laboratory
122
Ex. 2.23 Financial computation: Loan repayment

Find the payment required to pay off a $20000 loan in
ten periods, assume equal payment and the interest
rate per period =10%.
Let x[n]=b is the payment each period, first payment is made

when n=0
Let y[n] is the total balance, y[-1]=20000; ten periods, y[9]=0
= x +
= 1 + []
Prob. 2.19, 2.20

Research Laboratory
123
y[n] y[n 1] x[n]

Sol.
1 r 1.1, y[1] 20000,
Assume x[n] b is the payment/period, and the first payment is made when n 0,
Thus, the last payment is made when n 9, that is, y[9] 0.
Homogeneous solution: y ( h ) [n] ch (1.1)n ,
Particular solution: y ( p ) [n] c p , since the I/P is a constant
Solving for c p by substituting y ( p ) [n] c p , we have
c p 1.1c p b, c p 10b
Complete solution: y[n] ch (1.1) n 10b, n 0
y[0] 1.1y[1] x[0], ch (1.1) 0 10b 22000 b ch 22000 11b
y[n] (22000 11b)(1.1) n 10b, n 0, substitute y[9] 0, we have
22000(1.1)9
3254.91
0=(22000 11b)(1.1) 10b, b
9
11(1.1) 10
9

Research Laboratory
124
Figure 2.31 (p. 155)

Balance on a $20,000 loan for
Example 2.23. Assuming 10%
interest per period, the loan is paid
off with 10 payments of $3,254.91.

Research Laboratory
125
Prob. 2.19
Homogeneous solution:
a. Set all items involved to I/P to 0
x(t)=0
b. Find characteristic equation
r2+5r+6=0 (r+2)(r+3)=0
yh (t ) c1e2t c2e3t

Research Laboratory
126
Particular solution:
a. Assume the O/P has the same general form as I/P
t
y
(
t
)
c
e
p
p
b. Replacing y(t) by yp(t)

d2
d
t
t
t
t
(
c
e
)
5
(
c
e
)
6
c
e
e
p
p
p
dt 2
dt
1
c p e 5c p e 6c p e e , 2c p 1, c p
2
1 t
y p (t ) e , t 0
2
t

Research Laboratory
127
Complete solution:
1 t
y (t ) yh (t ) y p (t ) e c1e 2t c2e 3t , t 0
2
1
1 1
y (0) , c1 c2
2
2 2
d
d 1
1
y (t ) ( e t c1e 2t c2e 3t ) e t 2c1e 2t 3c2e 3t
dt
dt 2
2
d
1 1
1
y (t ) t 0 , 2c1 3c2
dt
2 2
2
1 t
c1 2, c2 1, y (t ) ( e 2e 2t e 3t )u (t )
2
Research Laboratory
128
Prob. 2.19
Homogeneous solution:
a. Set all items involved to I/P to 0
x[n]=x[n-1]=0
b. Find characteristic equation
r2-1/4=0 (r-1/2)(r+1/2)=0
1 n
1 n
yh [n] c1 ( ) c2 ( )
2
2

Research Laboratory
129
Particular solution:
a. Assume the O/P has the same general form as I/P
y p [n] c p
b. Replacing y[n] by yp[n]

1
cp cp 2 1
4
c p 4, y p [n] 4, n 0

Research Laboratory
130
Complete solution:
1
1
y[n] c1 ( ) n c2 ( ) n 4, n 0, y[2] 8, y[1] 0
2
2
1
1
y[n] y[n 2] 2 x[n] x[n 1], y[n] 2 x[n] x[n 1] y[n 2]
4
4
1
1
Let n=0, y[0] 2u[0] u[1] y[2] 2 8 4
4
4
1
1
Let n=1, y[1] 2u[1] u[0] y[1] 2 1 0 3
4
4
1
1
y[0] 4, c1 ( )0 c2 ( )0 4 4, c1 c2 0
2
2
11
11
y[1] 3, c1 ( ) c2 ( ) 4 3, c1 c2 2
2
2
1
1
c1 1, c2 1, y[n] (( ) n ( ) n 4)u[n]
2
2
Research Laboratory
131
Characteristics of systems described by

differential and difference equations
Express the O/P of a system described by a

differential or difference equation as the sum
of two components:
One associated only with the initial condition
The other due only to the I/P signal
Natural response, y(n)

Forced response, y(f)
The complete solution is y= y(n)+y(f)

Research Laboratory
132
The natural response
The natural response
Is the system O/P for zero I/P

Describe the manner that system dissipates any
store energy or memory of the past represented
by non-zero initial conditions
It is obtained from homogeneous solution by
choosing ci such that the initial conditions are
satisfied.

Research Laboratory
133
Ex. 2.24 RC circuit: Natural response

d
y (t ) x(t ), assume that y (0) 2(V ), R 1(W), C 1( F )
dt
Find the natural response of the system.
y (t ) RC
Sol.
The homogeneous solution, derived in Ex. 2.17, is
y ( h ) (t ) c1e t (V ).
From the initial condition,
y ( n ) (0) 2 c1 2 y ( n ) (t ) 2e t (V ), for t 0.

Research Laboratory
134

Natural response
1
y[n 1] x[n], assume that y[1] 8
4
Find the natural response of the system.
y[n]
Sol.
1
The homogeneous solution, derived in Ex. 2.21, is y ( h ) [n] c1 ( ) n .
4
From the initial condition, y ( n ) [1] 8, we have
1
1
8 c1 ( ) 1 c1 2 y ( n ) [n] 2( ) n , for n 1.
4
4
Prob. 2.21

Research Laboratory
Summary
135
The forced response
The forced response
Is the system O/P due to zero initial conditions

Describe the system behavior that is forced by
the I/P when the system is at rest
at rest: no energy or memory stored in the system
It is the same form as the complete solution

The forced response depends on the particular
solution, which is only valid for times t>0 or n0
The at-rest conditions must be translated forward to t=0+

or n0

Research Laboratory
136

Forced response
y[n]
1
y[n 1] x[n],
4
1
Find the forced response of the system if the I/P is x[n] ( ) n u[n].
2
Sol.
1
1
Form Ex. 2.21, the complete solution is of the form y[n] 2( ) n c1 ( ) n , n 0.
2
4
To obtain c1 , we have to translate the at-rest condition y[1] 0 to n 0
1
1
1
y[1] ( ) 0 0 1, then we can use y[0] 1 to find c1
4
2
4
1
1
1
1
1 2( )0 c1 ( )0 c1 1 y ( f ) [n] 2( ) n ( ) n , n 0
2
4
2
4
y[0] x[0]
Summary

Research Laboratory
137
Summary
y[n]
1
1
y[n 1] x[n], I/P is x[ n] ( ) n u[n], y[-1]=8.
4
2
Sol.
I . Complete solution: y[n] y ( h ) [ n] y ( p ) [ n]
1
1
Ex. 2.21, y ( h ) [n] c1 ( ) n , y ( p ) [n] 2( ) n , n 0
4
2
Ex. 2.21
1
1
with initial condition, c1 =1, y[ n] 2( ) n ( ) n , n 0.
2
4
y[1] 8 y[0] 3
II . Natrue response + Force response
1
Ex. 2.25, y ( n ) [n] 2( )n , for n 1.
4
Ex. 2.25
1 n 1 n
(f)
Ex. 2.26, y [n] 2( ) ( ) , n 0
2
4
Ex. 2.26

Research Laboratory
138
1 1
y[n] 2 , n 0
2 4
y[n]
1
1
y[n 1] x[n], if x[n] u[n], y[1] 8
2
2
1
1
1
Homogeneoue Solution: r 0, r , y ( h ) [n] c1
2
2
2
n
1
1 1
1
Particular Solution: Let y [n] c2 n , c2 n c2 (n 1)
2
2 2
2
( p)
1
c2 n c2 (n 1) 1, c2 1, y [n] n
2
Complete Solution:
n 1
1
,
2
( p)
1
1
1
1
1
y[1] 8, y[0] y[1] x[0] 5, y[n] c1 n , y[0] c1 5, y[n] 5 n
2
2
2
2
2
Natural Response: obtained from homogeneous solution with initial condition y[1] 8
n
1
1
1
y [n] c1 , y[1] 8, c1 8, c1 4, y ( n ) [n] 4 , n 1
2
2
2
Force Response: obtain from complete solution with zero initial condition
(n)
1
y[1] 0, y[0] y[1] x[0] 1,
2
Research Laboratory
1
1
1
1
y [n] c1 n , y[0] c1 1, y ( f ) [n] n , n 0
2
2
2
2
(f)
139
Ex. 2.27 RC circuit: Forced response

d
y (t ) RC y (t ) x(t ), assume that x (t ) cos(t )u (t )(V ), R 1(W), C 1( F )
dt
Find the forced response of the system.
Sol. The complete solution, derived in Ex. 2.22, is of the form
1
1
y (t ) ce t cos t sin t , t 0
2
2
The forced response is obtained by choosing c under the assumption that
1
y (0 ) y (0 ) 0, thus, we obtain c . The forced response is
2
1
1
1
y ( f ) (t ) e t cos t sin t (V ), t 0
2
2
2
Prob. 2.22

Research Laboratory
140
Impulse response
Step response
By definition, it assumes that the system is at rest
Represents the response of the system to a step

I/P with zero initial conditions
For given s(t) or s[n], the impulse response can

be obtained by
d
s(t )
dt
h[n] s[n] s[n 1]
h(t )

Research Laboratory
141
Impulse response
Impulse-response v.s. differential/difference equation
Impulse response description:
No provision for initial conditions
Applies only to systems that are initially at rest and when the
I/P is known for all time
Differential/difference equation description:
More flexible
Apply to systems either at rest or with nonzero initial

conditions

Research Laboratory
142
Linearity
The forced response of a LTI system is linear with

respect with the I/P
x1 y1( f ) , x2 y2( f ) x1 x2 y1( f ) y2( f )
The natural response of a LTI system is linear with

respect with the initial conditions
I1 y1( n) , I 2 y2( n) I1 I 2 y1( n) y2( n)

Research Laboratory
143
Time-invariance
Force response is time-invariant
Natural response, in general, is not time-invariant
A time shift in the I/P results in a time shift in the O/P

since the system is initially at rest
Initial conditions will result in an O/P term that does not
shift with a time shift of the I/P
Causality
Force response is causal
Since the system is initially at rest, the O/P does not

begin prior to the I/P is applied

Research Laboratory
144
Roots of the characteristic equation
Roots of the characteristic equation afford considerable

information about LTI system behavior
Force response
Natural response
depends on both the I/P and the roots of the characteristic

equation since it involves both the homogeneous solution
and a particular solution
Dependent entirely on the roots of the characteristic equation
Impulse response
Dependent entirely on the roots of the characteristic equation

Research Laboratory
145
Roots of the characteristic equation indicates
when an LTI system is unstable (in later chapters)
A discrete-time causal LTI system is stable iff all roots of
the characteristic equation have magnitude less than unity
A continuous-time causal LTI system is stable iff the real

part of all roots of the characteristic equation are negative
The stability conditions imply that the natural response of

an LTI system is a decaying exponential

Research Laboratory
146
Roots of the characteristic equation indicates
The response time of the system
Once the natural response has decayed to zero, the system

behavior is governed by the particular solution, which is of
the same form as the I/P
Natural-response component describes the transient

behavior of the system
The time it takes an LTI system to respond to a transient is

determined by the time it takes for the natural response to
decay to zero

Research Laboratory
147
Block diagram representations
A block diagram is
an interconnection of elementary operators

that act on the I/P signal
A more detailed representation of a system since
it describes how the systems internal
computations or operations are ordered
Impulse response and differential/difference
equations descriptions represent only the I/P-O/P
behavior of a system

Research Laboratory
148
Block diagram representations consist of an

interconnection of three elementary operations on
signals:
Scalar multiplication:
Addition
y(t)=cx(t) or y[n]=cx[n]
y(t)=x(t)+w(t) or y[n]=x[n]+w[n]
Integration/Time shift
y(t ) x( )d ; y[n] x[n 1]

Research Laboratory
149
Figure 2.32 (p. 162)

Symbols for elementary operations in block
diagram descriptions of systems. (a) Scalar
multiplication. (b) Addition. (c) Integration for
continuous-time systems and time shifting for
discrete-time systems.

Research Laboratory
150
Figure 2.33 (p. 162)

Block diagram representation of a
discrete-time LTI system described
by a second-order difference
equation.
w[n] b0 x[n] b1 x[n 1] b2 x[n 2]

+)
y[n] w[n] a1 y[n 1] a2 y[n 2]

y[n] a1 y[n 1] a2 y[n 2] b0 x[n] b1 x[n 1] b2 x[n 2]

Research Laboratory
151
Prob. 2.23
1
1
(a) y[n] y[n 1] y[n 3] x[n] 2 x[n 2]
2
3
1
Figure 2.34a (p. 163)

Block diagram representation for
Problem 2.33 (2.34b in next slide).

Research Laboratory
152
Prob. 2.23
1
1
(b) y[n] y[n 1] y[n 2] x[n 1]
2
4
1
Figure 2.34b (p. 163)

Block diagram representation for
Problem 2.33 (2.34b in next slide).

Research Laboratory
153
Block diagram representation of a system is not unique

Ex.
Direct form I:
w[n] b0 x[n] b1 x[n 1] b2 x[n 2]
y[n] w[n] a1 y[n 1] a2 y[n 2]
y[n] a1 y[n 1] a2 y[n 2] b0 x[n] b1 x[n 1] b2 x[n 2]
Direct form II:

f [n] a1 f [n 1] a2 f [n 2] x[n]
y[n] b0 f [n] b1 f [n 1] b2 f [n 2]

Research Laboratory
154
Figure 2.35 (p. 164)

Direct form II representation of
an LTI system described by a
second-order difference equation.

155
Research Laboratory
f [n] a1 f [n 1] a2 f [n 2] x[n]
y[n] b0 f [n] b1 f [n 1] b2 f [n 2]
Prob. 2.24 y[n]
1
1
y[n 1] y[n 2] x[n] x[n 1]
4
8
1
Figure 2.36 (p. 165)

Solution to Problem 2.24. (a)
Direct form 1. (b) Direct form 2.

Research Laboratory
156
There are many different implementations of an LTI

system whose I/P-O/P behavior is described by a
difference equation
While these different systems are equivalent from an
I/P-O/P perspective, they generally differ with respect
to other criteria, such as memory requirement (
), the number of computations required per
O/P value (), and numerical accuracy ().

Research Laboratory
157
M
N
M
dk
dk
N
( N k )
( N k )
a
y
(
t
)
b
x
(
t
)
a
y
(
t
)
b
x
(t )
k
k
k
k
k
k
dt
dt
k 0
k 0
k 0
k 0
N
d
d2
d
d2
a0 y(t ) a1 y (t ) a2 2 y (t ) b0 x(t ) b1 x(t ) b2 2 x(t )
dt
dt
dt
dt
If a2=1
y (t ) a1 y (1) (t ) a0 y (2) (t ) b2 x(t ) b1 x (1) (t ) b0 x (2) (t ),

where v ( n ) (t )
( n 1)
( n 1)
(n)
v
(
)
d
v
(
)
d
v
(0), n 1, 2,3,
Figure 2.37 (p. 166)

of a continuous-time LTI
system described by a secondorder integral equation. (a)
Direct form I. (b) Direct form II.

158
Research Laboratory

Research Laboratory
159
Prob. 2.25
Prob. 2.25
d2
d
d2
Sol. 2 y (t ) 3 y (t ) x(t ) 2 2 x(t )
dt
dt
dt
y (t ) 3 y (2) (t ) x (1) (t ) 2 x(t )
What is b0, b1, b2, a0, a1?
Figure 2.38 (p. 166)

Block diagram representation for Problem 2.25.

Research Laboratory
160
State-variable description of LTI system
The state-variable description of an LTI

system consists of
a series of coupled first-order differential or

difference equations that describes how the state
of the system evolves, and
an equation that relates the O/P of the system to
the current state variables and I/P
These equations are written in matrix form

Research Laboratory
161
State of a system
A minimal set of signals that represent the

systems entire memory of the past
That is, given only the value of the state at an
initial point in time ni (ti), and the I/P for time n ni
(tti), we can determine the O/P for all times n ni
(tti).

Research Laboratory
162
Figure 2.39 (p. 167)

Direct form II representation of a
second-order discrete-time LTI
system depicting state variables
q1[n] and q2[n].

Research Laboratory
163
We may choose q1[n] and q2 [n] as the state variable

The next value of the state,q1[n 1] and q2 [n 1], must correspond to
the variables at the I/P to the time-shift operations
q1[n 1] a1q1[n] a2 q2 [n] x[n]
q2 [n 1] q1[n]
The I/P-O/P relation of the system
y[n] x[n] a1q1[n] a2 q2 [n] b1q1[n] b2 q2 [n]
(2.57)
(b1 a1 ) q1[ n] (b2 a2 ) q2 [ n] x[ n]
(2.59)

Research Laboratory
164
(2.58)

q1[n 1] a1
(2.57), (2.58)
q2 [n 1] 1
a2 q1[n] 1
x[ n]
0 q2 [n] 0
q1[n]
y[ n] (b1 a1 ) (b2 a2 )
(2.59)
1 x[ n]
]
n
[
q
2
q1[n]
Define q[n]
]
n
[
q
2
(2.60) q[n 1] Aq[n] bx[n]
(2.61) y[n] cq[n] Dx[n]
a1
A
1

Research Laboratory
a2
1
, b , c = (b1 a1 ) (b2 a2 ) , D 1
0
0
165
(2.60)
(2.61)
Exploring concepts with MATLAB
Convolution: y=conv(x,h)
Step-response
Simulating difference equations
State-variable descriptions

Research Laboratory
166
Summary
We have examines four different description

of an LTI system
Impulse response
Differential/difference equations
Block diagram
All four description are equivalent in the I/PO/P sense
For a given I/P, each description will produce the

same O/P

Research Laboratory
167
Summary
Impulse response
is the O/P of a system when the I/P is an impulse

System properties, such as causality and stability,
are directly related to the impulse response
I/P must be known for all time in order to
determine the O/P of a system

Research Laboratory
168
Summary
Differential/Difference equations
Order of a differential equation reflects the maximum

number of energy storage devices in the system
Order of a difference equation represents the
maximum memory of past O/P
O/P of a system from a given point in time forward
can be determined without knowledge of all past I/Ps,
provided that the initial conditions are known
Solution
Natural response: system behavior due to initial conditions

Forced response: system behavior in response to I/P alone

Research Laboratory
169
Summary
Block diagram
Represents a system as an interconnection of

elementary operations on signals
The interconnection define the internal structure of
the system
May have different block diagram representations
with identical I/P-O/P characteristics

Research Laboratory
170
Summary
Consists of a set of coupled first-order differential

or difference equations representations
representing the systems behavior
Written in the matrix form, the description consists
of two equations
One describes how the system evolves

The other relates the state to the O/P

Research Laboratory
171

CH 2

Încărcat de

Informații document

Titlu original

Drepturi de autor

Formate disponibile

Partajați acest document

Partajați sau inserați document

Opțiuni de partajare

Vi se pare util acest document?

Este necorespunzător acest conținut?

Drepturi de autor:

Formate disponibile

CH 2

Încărcat de

Drepturi de autor:

Formate disponibile

Time-domain representations of

linear time-invariant systems

Broadband Mobile Wireless

Differential and Difference Equation Representations

Broadband Mobile Wireless

I/P and O/P signals are represented as functions

Broadband Mobile Wireless

The O/P of an LTI system due to a unit impulse signal I/P

I/P signal: a weighted superposition of time-shifted impulses

Linearity and time invariance

Weighted superposition of time-shifted impulses

Continuous-time system: convolution integral

Broadband Mobile Wireless

Impulse Response of LTI System

Characterizing the I/P-O/P behavior of LTI

Broadband Mobile Wireless

are constant coefficients

Represents the system as an interconnection of three

Time shift (DT) or integration (CT)

Broadband Mobile Wireless

State variable description

A set of variables associated with energy storage or

A series of coupled first-order differential or

Broadband Mobile Wireless

The Convolution Sum

Figure 2.1 (p. 99)

Broadband Mobile Wireless

The Convolution Sum

y[n] H {x[n]} H { x[k ]d [n k ]}

H {x[k ]d [n k ]} p.63 property of superposition

x[k ]H {d [n k ]} p.63 property of homogeneity

Let h[n] H{d [n]}

x[k ]h[n k ] x[n]* h[n],

where * denotes the convolution sum

Broadband Mobile Wireless

Broadband Mobile Wireless

Figure 2.2a (p. 100)

Broadband Mobile Wireless

Figure 2.2b (p. 101)

The Convolution Sum

Ex. 2.1 Multipath Communication Channel

Broadband Mobile Wireless

The Convolution Sum

y[n] 2h[n] 4h[ n 1] 2h[ n 2]

Broadband Mobile Wireless

Convolution Sum Evaluation Procedure

y[n] is a weighted sum of each time-shifted impulse response

where wn [k ] x[k ]h[n k ] is the intermediate signal

Broadband Mobile Wireless

Broadband Mobile Wireless

-n 1-n 2-n 3-n

n-3 n-2 n-1 n

where wn [k ] x[k ]h[n k ] is the intermediate signal

Broadband Mobile Wireless

Convolution Sum Evaluation Procedure

Ex. 2.2 Convolution Sum by using an Intermediate Signal

Figure 2.3 (p. 103)

Broadband Mobile Wireless

where wn [k ] x[k ]h[n k ] is the intermediate signal

Figure 2.3 (p. 103)

(b) The product signal w5[k] used to evaluate y [5].

Broadband Mobile Wireless

where wn [k ] x[k ]h[n k ] is the intermediate signal