Signals and Systems

Spring 2008
Lecture #4
(2/14/2008)
• Covers O & W pp. 90-102, 127-136
• Representation of CT Signals in terms of shifted
unit impulses
• Introduction of the unit impulse δ(t)
• Convolution integral representation of CT LTI
systems
• Properties and Examples
“Figures and images used in these lecture notes by permission,
copyright 1997 by Alan V. Oppenheim and Alan S. Willsky”
1

Representation of CT Signals

• Approximate any input x(t) as a sum of shifted, scaled square pulses

(in fact, that is how we do integration)

xˆ (t) = x(k")
constant in
k" < t (k + 1)"
changes discontinuously at the next step
2
 Representation of CT Signals (cont.)

δ∆(t) has a unit area

1
424
3
=1, located at k"

!
+%
xˆ(t) = & x(k")# " (t $ k")"
k = $%
!
' limit as " ( 0
%
x(t) = *$% x() )# (t $ ) )d)

The Unit-impulse function δ(t)

Paul Dirac first introduced this function in 1920’s to
describe the charge density of an electron (diameter < 10-13
cm), extremely useful, but made many mathematicians
uncomfortable. Definition of δ(t):

" (t) = 0 for t # 0%

& (1)
" (t) = $ for t = 0 '
0+
) 0(
" (t)dt = 1 (2)

–– an infinitesimally sharp pulse with an unity area.

! 4
 Construction of the Unit-impulse function δ(t)
One of the simplest way –– rectangular pulse, taking the limit
∆ → 0.

But this is by no means the only way. One can construct a δ(t)
function out of many other functions, Eg. Gaussian pulses,
triangular pulses, sinc functions, etc., as long as the pulses are
short enough –– much shorter than the characteristic time
scale of the system.
5

Demo: Response of an RC circuit to short pulses

Response from initial rest to pulses

of different shapes and durations,
but with unit area. As the duration
decreases, the response becomes
similar for different pulse shapes.

6
 Response of a CT LTI System
x(t) h(t) y(t)

• Now suppose the system is LTI, and define the unit

impulse response h(t):
" (t) #
#$ h(t)
%
From Time-Invariance:

" (t & ' ) #

#$ h(t & ' )
From Linearity:
+( +(
x(t) = ) &(
x(' )" (t & ' )d' #
#$ y(t) = ) x(4
144
' )h(t & ' )d' = x(t) * h(t)
&(2444 3
Convolution Integration

Operation of CT Convolution

+&
y(t) = x(t) " h(t) # ' %&
x($ )h(t % $ )d$

h($ ) (Flip
( ( () h(t % $ ) (Multiply
() h(%$ ) (Slide (( ()
Integrate +&
x($ )h(t % $ ) ( ( (() ' %&
x($ )h(t % $ )d$

!
8
 Demo: CT convolution

Properties and Examples

1) Commutativity: x(t)! h(t) = h(t) ! x(t)
2) x(t)! " (t # t o ) = x(t # to ) (Sifting property: x(t) ! " (t) = x(t))
3) An integrator: t
y(t) = $ x(! )d!
"#

%
t
h(t) = $"#& (! )d! = u(t )
That is
t
x(t)' u(t) = $"# x(! )d!

4) Step response
t
s(t) = u(t)! h(t) = h(t) !u(t) = %#$ h(" )d"

10
Distributivity

11

Associativity

12
 More properties
Causality: CT LTI system is causal ↔ h(t) = 0, at t < 0
As a result:
+$ t
y(t) = % #$
x(" )h(t # " )d" = % #$
x(" )h(t # " )d"
y(t) only depends on x(τ < t).
+%
BIBO Stability: CT LTI system is stable " &$% h(# ) d# < %
!
Sufficient condition: For | x(t) | ' x max < %.
+% +%
y(t) = & $%
x(# )h(t $ # )d# ' x max & $%
h(t $ # )d# < %.
+$
Necessary condition: Suppose % #$
h(" ) d" = $

! Let x(t) = h * (#t) / h * (#t), then x(t) & 1 bounded

+$ +$ h * (#" )h(#" ) +$
But y(0) = % x(" )h(#" )d" = % d" = % h(#" ) d" = $
#$ #$ h(#" ) #$

13

More About δ(t):

The Operation Definition of the Unit Impulse

δ(t) — idealization of a unit-area pulse that is so short that, for

any physical systems of interest to us, the system responds
only to the area of the pulse and is insensitive to its duration

Operationally: The unit impulse is the signal which when

applied to any LTI system results in an output equal to the
impulse response of the system. That is,

! (t) " h(t) = h(t) for all h(t)

— δ(t) is defined by what it does under convolution.

14
 The Unit Doublet — Differentiator

d dx(t)
x(t) y(t) =
dt dt

Impulse response = unit doublet

d! (t )
u1 (t ) =
dt

The operational definitions of the unit doublets:

dx(t)
x(t)! u1 (t ) =
dt

15

Triplets and beyond!

n>0
un (t) = u1 (t) !L!u1 (t)
1442443
n times

Operational definitions:

d n x(t)
x(t)! un (t) = (n > 0)
dt n

16
 Integrators
t
x(t) ! y(t) = $"# x(! )d!

Impulse response: u!1(t) " u(t)

t
Operational definition: x(t)! u"1(t) = %"$ x(# )d#

Cascade of n integrators:

u!n (t) = u1 (t)4

!14 "L"u
4244 4
!1 (t)
3 (n > 0)
n times

17

Integrators (continued)

u!1(t)
! (t) ! ! u!2(t)

t t
u!2(t) = $!# u!1 (" )d" = $!# u(" )d"
t
= u(t)$0 d"
= t % u(t) the unit ramp

More generally, for n > 0

t (n!1)
u!n (t) = u(t)
(n ! 1)!

18
 Notation

Define u0 (t) = ! (t)

Then un (t)! um (t) = un+ m (t)

n and m can be ±

E.g. u1 (t )! u"1 (t) = u0 (t)

||
# d u(t)% = ' (t)
$ dt &

Note: The order of integration and differentiation

may not be exchangeable (e.g. x(t) = f(t) + A).
19

Example: Calculating x(t)∗h(t)

dx(t)
= ! (t + 1) " 2! (t " 1) + ! (t " 2)
dt
20
 Example (continued)

dx(t)
! h(t) = h(t + 1) " 2h(t " 1) + h(t " 2)
dt

x(t) " h(t) = u#1 (t) "{[u1 (t) " x(t)]" h(t)]}

t % dx($ ) (
= , #+ '
& d$
" h( $ )*) d$

! 21

Next lecture covers:

O & W pp. 177-201

22