Signals and Systems

Ch2: Continuous-Time Signals

and Systems
Lecturer: Nguyen-Son Vo (Ph.D), email: [email protected]

1/55
Ch2-Continuous-Time Signals and Systems
2.1 Representation of signals in terms of impulses
2.2 Continuous-time systems (CTS)
2.3 Convolution integral
2.4 Properties of linear time-invariant (LTI) CTS
2.5 Systems described by differential equations
2.6 Block-diagrams of LTI systems by differential equations
2.7 Singularity functions
2.8 State-space of continuous-time LTI systems
2.9 Matlab exercises
2.10 Problems
2/55
2.1 Representation of signals in terms of impulses

 Consider pulse or staircase approximation xˆ (t ) of x(t )

 Since (t) = 1

3/55
 2.2 Continuous-time systems (CTS)

 Linear systems
• x1(t) and x2(t) are the inputs of a system
• y1(t) and y2(t) are the outputs, respectively

• Homogeneity

• Superposition

• Linearity

4/55
 2.2 Continuous-time systems (CTS)

 Time-invariant system
• The behavior and characteristics of the system are fixed over time
• A time shift in the input  an identical time shift in the output

Example 2.1: Check for linearity and time-invariance of the following system

Solution:

 linear
x(t – t0)  y(t) = tx(t – t0)  time variant
5/55
 2.2 Continuous-time systems (CTS)

 Time-invariant system
Example 2.2: Check for linearity and time-invariance of the following system

Solution:

 Nonlinear
x(t – t0)  y(t) = sin(x(t – t0)) = y(t – t0)  time invariant
6/55
 2.2 Continuous-time systems (CTS)

 Time-invariant system
Example 2.3: Consider an LTI system with the response y(t) to the input signal
x(t) as shown in the figure below. Determine and sketch the response of the
system to the following inputs
a) x1(t) = x(t) – x(t – 4)
b) x2(t) = x(t) + x(t + 2)
Solution:
a) Since it is an LTI system, the response y1(t) to x1(t) is y1(t) = y(t) – y(t – 4)
b) Similarly, y2(t) = y(t) + y(t + 2)

7/55
 2.2 Continuous-time systems (CTS)

 Causal system
• The output depending upon present and past inputs only
• Non-anticipatory
 Stable system
• Bounded output for bounded input
• If the signal is bounded, then its magnitude will always be finite
 Memory and memoryless system
• Memory: output at any specified time depends on the inputs at that
specified time and others, with memory or energy storage elements
• Memoryless/Static: output depends upon the present input only

8/55
 2.2 Continuous-time systems (CTS)

 Invertible system
• Input can be recovered from its output
• Otherwise, noninvertible
 Step and impulse response
• If the input to the system is unit impulse input δ(t), the system output is
called the impulse response and denoted by h(t)

• If the input to the system is a unit step input u(t), then the system output
is called the step response s(t)

9/55
 2.3 Convolution integral

 Output of a system for an input expressed as weighted

superposition is given by

 Applying the linearity property, we have

 Applying time-invariant property to [(t – )], we have

10/55
 2.3 Convolution integral

 Properties of convolution integral

• Commutative property

Proof

Let v = t –    = t – v and d = –dv

11/55
 2.3 Convolution integral

 Properties of convolution integral

• Distributive property

Proof: refer to [2]

• Associative property

Proof: refer to [2]

12/55
 2.3 Convolution integral

 Properties of convolution integral

• Convolution with an impulse

Proof: refer to [2]

• Convolution with delayed input and delayed impulse response

Proof: refer to [2]

13/55
 2.3 Convolution integral
 Properties of convolution integral
Example 2.4: Determine the continuous-time convolution of x(t) and h(t) for the
following
a) x(t) = u(t), h(t) = u(t)
b) x(t) = u(t – a), h(t) = u(t – b)
c) x(t) = u(t + 1) – u(t – 1), h(t) = u(t + 1) – u(t – 1)
d) x(t) = e-(t – 2)u(t – 2), h(t) = u(t + 2)
e) x(t) = u(t), h(t) = e-tu(t)
f) x(t) = 2(u(t) – u(t – 2)), h(t) = e-t/2u(t)
g) x(t) = ((t – 2)/2), h(t) = (t – 1)


y (t )   x( )h(t   )d


 Applying time reversal to h() and then applying time shifting by t

14/55
 2.3 Convolution integral

 Properties of convolution integral

Example 2.4: Determine the continuous-time convolution of x(t) and h(t) for
the following
a) x(t) = u(t), h(t) = u(t)
Solution:

y (t )   x( )h(t   )d

t
  1d  t , t  0
0

t , t  0
y (t )    tu (t )
0, t  0

15/55
 2.3 Convolution integral

 Properties of convolution integral

Example 2.4: Determine the continuous-time convolution of x(t) and h(t) for
the following
b) x(t) = u(t – a), h(t) = u(t – b)
Solution:

y (t )   x( )h(t   )d

t
  1d  t  a  b, t  a  b
a b

t  a  b, t  a  b
y (t )  
0, t  a  b
= (t  a  b)u (t  a  b)
16/55
 2.3 Convolution integral

 Properties of convolution integral

Example 2.4: Determine the continuous-time convolution of x(t) and h(t) for
the following
c) x(t) = u(t + 1) – u(t – 1), h(t) = u(t + 1) – u(t – 1)
Solution:
x(t)*h(t) = (u(t + 1) – u(t – 1))*(u(t + 1) – u(t – 1))
= u(t + 1)* u(t + 1) – 2u(t + 1)*u(t – 1)
+ u(t – 1)*u(t – 1)
t
y (t )   1d  (t  2), t  2
2
 (t  2)u (t  2)
17/55
 2.3 Convolution integral

 Properties of convolution integral

Example 2.4: Determine the continuous-time convolution of x(t) and h(t) for
the following
c) x(t) = u(t + 1) – u(t – 1), h(t) = u(t + 1) – u(t – 1)
Solution:
x(t)*h(t) = (u(t + 1) – u(t – 1))*(u(t + 1) – u(t – 1))
= u(t + 1)* u(t + 1) – 2u(t + 1)*u(t – 1)
+ u(t – 1)*u(t – 1)
t
y (t )   2d  2t , t  0
0
 2tu (t )

18/55
 2.3 Convolution integral

 Properties of convolution integral

Example 2.4: Determine the continuous-time convolution of x(t) and h(t) for
the following
c) x(t) = u(t + 1) – u(t – 1), h(t) = u(t + 1) – u(t – 1)
Solution:
x(t)*h(t) = (u(t + 1) – u(t – 1))*(u(t + 1) – u(t – 1))
= u(t + 1)* u(t + 1) – 2u(t + 1)*u(t – 1)
+ u(t – 1)*u(t – 1)
t
y (t )   d  t  2, t  2
2
 (t  2)u (t  2)

19/55
 2.3 Convolution integral

 Properties of convolution integral

Example 2.4: Determine the continuous-time convolution of x(t) and h(t) for
the following
c) x(t) = u(t + 1) – u(t – 1), h(t) = u(t + 1) – u(t – 1)
Solution:
x(t)*h(t) = (u(t + 1) – u(t – 1))*(u(t + 1) – u(t – 1))
= u(t + 1)* u(t + 1) (1)
– 2u(t + 1)*u(t – 1) (2)
+ u(t – 1)*u(t – 1) (3)

20/55
 2.3 Convolution integral

 Properties of convolution integral

Example 2.4: Determine the continuous-time convolution of x(t) and h(t) for
the following
d) x(t) = e-(t – 2)u(t – 2), h(t) = u(t + 2)
t t
 
Solution: y (t )   e d  (e )
0
0

 1  e t , t  0
t 2
  e ( 2) d
2
 t 2 y(t) = 0, t < 0
 e (e 2
)
2
t  2
 e 2 (e  e 2 )
 1  e t , t  0 21/55
 2.3 Convolution integral

 Properties of convolution integral

Example 2.4: Determine the continuous-time convolution of x(t) and h(t) for
the following
e) x(t) = u(t), h(t) = e-tu(t)
Solution:
t t
 
y (t )   e d  (e )
0
0

 1  e t , t  0
= (1  et )u (t )

22/55
 2.3 Convolution integral

 Properties of convolution integral

Example 2.4: Determine the continuous-time convolution of x(t) and h(t) for
the following
f) x(t) = 2(u(t) – u(t – 2)), h(t) = e-t/2u(t)
t
 /2  /2 t
Solution: y (t )  2  e d  4(e )
0
0
t
 /2  /2 t
y (t )  2  e d  4(e )
t 2
t 2
t /2
= 4e (e  1), t  2

23/55
 2.3 Convolution integral

 Properties of convolution integral

Example 2.4: Determine the continuous-time convolution of x(t) and h(t) for
the following
g) x(t) = ((t – 2)/2), h(t) = (t – 1)
Solution: 
  2 
y (t )    2   (t  1   )d



  2 
=   2    (  t  1) d



  2 
   2   (  (t  1)) d


 t 1 2   t 3
     
 2   2  24/55
 2.3 Convolution integral

 Graphical convolution
• Step 1: make x() fixed
• Step 2: invert h() to obtain h(-)
• Step 3: shift h(-) along  axis by t, i.e., h(-( - t))
• Step 4: compute the area under the product of x() and h(-( - t))
with respect to different positive and negative values of t, i.e.,
different lower and upper integral limits, to have y(t)

Note: we can fix h() and shift x(-) by t if it is more comfortable


y (t )   x( )h(t   )d

25/55
 2.3 Convolution integral

 Graphical convolution
1, 0  t  4
Example 2.5: find y(t) = x(t)*h(t), x(t )  h(t )  
 t  0, otherwise
  1d  t, 0  t  4
0
4

y(t )   1d  8  t , 4  t  8

t  4


0, otherwise


26/55
 2.3 Convolution integral

 Graphical convolution
Example 2.6: find y(t) = x(t)*h(t)
t
  1d  t, 0  t  2
0
2 t

  1d   1d  2t  6, 2  t  4
y(t )  t 2 2
4

  1d  t  6, 4  t  6
t  2
0, otherwise


27/55
 2.3 Convolution integral

 Graphical convolution
Example 2.7: find y(t) = x(t)*h(t)
Y (s) 1/ sC a 1 1
H(s)    ,a  
X ( s ) R  1/ sC s  a T RC
 a t  a 1
  e d  (1  e at ), 0  t  T0
 T0 0 T0
  at
 a t  a e
y(t )  
    1), t  T0
aT0
e d (e
 T0 t T0 T0


0, otherwise


28/55
 2.3 Convolution integral

 Graphical convolution
Example 2.7: find y(t) = x(t)*h(t)

29/55
 2.3 Convolution integral

 Graphical convolution
Example 2.7: find y(t) = x(t)*h(t)

30/55
 2.3 Convolution integral

 Graphical convolution
Example 2.8: Consider an LTI system with the impulse response
h(t) = e-tu(t). Find the system response to the input x(t) = sin(2t)u(t).

t t
y (t )   sin(2 ) e t d  sin(2 )e t |t0 2  cos(2 )e t d
0 0
  t t
t
 t

= sin(2t )  2 cos(2 ) e |0 2  sin(2 ) e d 
 0 
= sin(2t )  2 cos(2t )  2e t  4 y (t )

31/55
 2.3 Convolution integral

 Graphical convolution
Example 2.8: Consider an LTI system with the impulse response
h(t) = e-tu(t). Find the system response to the input x(t) = sin(2t)u(t).
y (t )  (sin(2t )  2 cos(2t )  2e t ) / 5
y (t )  [a(2 e-at  2 cos(2t )  a sin(2t ))] / ( a 2  4)

32/55
 2.3 Convolution integral

 Graphical convolution
Example 2.9: Show y(t) = x(t)*h(t), T = 2, where x(t) and h(t) are given in the
following figures
Solution
y (t)  h(t ) * x(t )
n 
 h(t ) *   (t  nT )
n 
n 
=  h(t ) *  (t  nT )
n 
n 
=  h(t  nT )
n 

33/55
 2.3 Convolution integral
 Computation of Convolution Integral Using Matlab
function[y,ty]=convint(x,tx,h,th)
%Inputs:
%x is the input signal vector
%tx is the times of the samples in x
%h is the impulse response vector
%th is times of the samples in h
%outputs:
%y is the output signal vector,
%length(y)=length(x)+length(h)-1
%ty is the time of the samples in y
dt=tx(2)-tx(1);
y=conv(x,h)*dt;
ty=(tx(1)+th(1))+[0:(length(y)-1)]*dt;

34/55
 2.3 Convolution integral

 Computation of Convolution Integral Using Matlab

Example 2.10: Consider the RC circuit with a = 3. Find y(t) for x(t) = u(t-3) – u(t-5).
clc; clear all; close all;
tx=0:0.01:5; th=0:0.01:5;
x=rectangularPulse(3,5,tx);
h=3*exp(-3*th);
[y ty]=convint(x,tx,h,th);
figure;
plot(ty,y,'r-');
xlabel('Time')
ylabel('Amplitude');
legend('y(t)')
grid;
axis([0 10 -0.5 1.5]);
35/55
 2.3 Convolution integral

 Computation of Convolution Integral Using Matlab

Example 2.10: Consider the RC circuit with a = 3. Find y(t) for x(t) = u(t-3) –
u(t-5).
Solution
 t 3( t )
  3e d  e3( t ) |t3  1  e3(3 t) ,3  t  5
3
5

y(t)   3e3( t ) d  e3( t ) |5  e3(3t ) (e6  1), t  5
3
 3



0, otherwise


36/55
 2.3 Convolution integral

 Interconnected systems
• Cascade (associative property)
y (t)  x(t ) * h1 (t ) * h2 (t )
 
h(t )  h1 (t ) * h2 (t )   h1( )h2 (t   )d   h2 ( )h1(t   )d
 

• Parallel (distributive property)

y (t)  x(t ) *[h1 (t )  h2 (t )]  x(t ) * h1 (t )  x(t ) * h2 (t )

h(t )  h1 (t )  h2 (t )
37/55
 2.3 Convolution integral

 Interconnected systems
Example 2.11: An LTI system consists of two subsystems in cascade. The
impulse responses of the subsystems are given by h1(t) = e-3tu(t), h2(t) = e-tu(t).
Find the impulse response of the overall system.
Solution 
h(t )  h1 (t ) * h2 (t )   h1( )h2 (t   )d

t t
  e3 e(  t)  et  e2 d
0 0 Max?

et 2 t
= e |0
2
e t
 (1  e2t ), t  0
2 38/55
 2.3 Convolution integral

 Periodic convolution
• If the signals x1(t) and x2(t) are periodic with common period T, it can be
easily shown that the convolution of x1(t) and x2(t) does not converge
• In such a case, the periodic convolution of x1(t) and x2(t) is defined as

Example 2.12 Let y(t) be the periodic convolution of x1(t) and x2(t). Show that
y(t) = y(t + T).
Solution
T T
y (t  T )   x1 ( ) x2 (t  T   )d   x1 ( ) x2 (t   )d  y (t )
0 0

39/55
 2.3 Convolution integral

 Applications
• If the signal x(t) is transmitted to hit a target and we receive an echo
signal x(t – ), i.e., a delay version of x(t).
• However, we do not know  because we do not know the distance from
the transmitter to the target.
• How to determine  to get the distance?

40/55
 2.3 Convolution integral

 Applications
• A microphone recording made in a room is the convolution of
the sound signal produced by the voice with the impulse
response of the surrounding space.
• The impulse response h(t) is known for such a system. How to
remove the effect of surrounding space (deconvolution)?

41/55
 2.4 Properties of LTI-CTS

 LTI systems with and without memory

• The relationship between the y(t) and x(t) for a memoryles
system is y(t) = cx(t)

while y (t )   h( ) x(t   )d

• So, we have h(t) = c(t)
 Causality for LTISystem
• We have y (t )   h( ) x(t   )d

since h() = 0 for  < 0 for a causal CTS.
• So, the output of a causal system is

y (t )   h( ) x(t   )d
0 42/55
 2.4 Properties of LTI-CTS

 Stability for LTI system

• A continuous-time system is BIBO stable if and only if the
impulse response is absolutely integrable, that is


 | h( ) | d  

Example 2.13: Check stability of continuous-time system having the following
impulse responses
a) h(t) = e-tu(t), Y b) h(t) = e-tcos(2t)u(t), Y c) h(t) is periodic and nonzero, N
Example 2.14: Determine if each of the following system is causal (C) or stable (S)
a) h(t) = e-tu(t-1)  C, S b) h(t) = e-tu(-t+1)  no C nor S
c) h(t) = e-2tu(t+10)  no C, but S d) h(t) = te-tu(t)  C, S
e) h(t) = etu(-t-1)  no C, but S f) h(t) = e-2|t|  no C, but S
43/55
 2.4 Properties of LTI-CTS

 Invertible LTI System

• A system is invertible if its input x(t) can be recovered from its
output y(t) = x(t)*h(t)
• The process of recovering x(t) from x(t)*h(t) is called
deconvolution aforementioned in the applications of conv.

x(t ) * h(t ) * h 1 (t )  x(t )  h(t ) * h 1 (t )   (t )

Example 2.15: It is obvious that the inverse system for a continuous-time

integrator is a differentiator.

44/55
 2.4 Properties of LTI-CTS

 Invertible LTI System

Example 2.16: Check y(t) = x(2t) for causality and invertibility.
Solution
We can see that y(1) = x(2)  not causal
And x(t) = y(t/2)  invertible
Example 2.17: Check the following systems are invertible or not
a) y(t) = x(t-4) b) y(t) = x2(t)
Solution
a) x(t) = y(t+4)  invertible
b) We can see that there are two values of x(t), i.e., - and +, that yield the same y(t)
 Not invertible

45/55
Ch2-Continuous-Time Signals and Systems
2.1 Representation of signals in terms of impulses
2.2 Continuous-time systems (CTS)
2.3 Convolution integral
2.4 Properties of linear time-invariant (LTI) CTS
2.5 Systems described by differential equations
2.6 Block-diagrams of LTI systems by differential equations
2.7 Singularity functions
2.8 State-space of continuous-time LTI systems
2.9 Matlab exercises
2.10 Problems
46/55
 2.10 Problems

Prob. 2.1 If y(t) = x(t)*h(t), show that y(2t) = 2x(2t)*h(2t).

Prob. 2.2 If x(t) and h(t) are odd signals, show that y(t) = x(t)*h(t) is an even
signal.
Prob. 2.3 If the response of an LTI system to input x(t) is the output y(t), show
that the response of the system to dx/dt is dy/dt. Using this result to determine
the impulse response of an LTI system having the response y(t) = sin(2t) for an
input x(t) = e-4tu(t).
Prob. 2.4 Consider a cont.-time LTI system with the unit step response s(t).
a) deduce that the response y(t) of the system to the input x(t) is

dx(t   ) dx(t )
y (t)   s ( ) d  * s(t )

dt dt

b) and also show that dx(t   ) dx(t )
x(t )   u ( )
dt
d 
dt
* u (t )
 47/55
 2.10 Problems

Prob. 2.1 If y(t) = x(t)*h(t), show that y(2t) = 2x(2t)*h(2t).

Prob. 2.2 If x(t) and h(t) are odd signals, show that y(t) = x(t)*h(t) is an even
signal.
Prob. 2.3 If the response of an LTI system to input x(t) is the output y(t), show
that the response of the system to dx/dt is dy/dt. Using this result to determine
the impulse response of an LTI system having the response y(t) = sin(2t) for an
input x(t) = e-4tu(t).
Prob. 2.4 Consider a cont.-time LTI system with the unit step response s(t).
a) deduce that the response y(t) of the system to the input x(t) is

dx(t   ) dx(t )
y (t)   s ( ) d  * s(t )

dt dt

b) and also show that dx(t   ) dx(t )
x(t )   u ( )
dt
d 
dt
* u (t )
 48/55
 2.10 Problems

Prob. 2.5 Using the result of 2.4 to determine y(t) given s(t) = (e-2t – e-t +1)u(t)
and x(t) = etu(t).
Prob. 2.6 For each of the following impulse responses, determine wheter the
corresponding system is memoryless, causal, and stable
a) h(t) = u(t + 1) – u(t – 1) b) h(t) = u(t) – 2u(t – 1) c) h(t) = e-2|t|
Prob. 2.7 Compute the following convolution integrals
a) y(t) = (u(t) – u(t – 2))*u(t)
b) y(t) = e-3tu(t)*u(t + 3)
c) y(t) = cos(t)(u(t + 1) – u(t – 1))*u(t)
d) y(t) = (u(t + 3) – u(t – 1))*u(–t + 4)
e) y(t) = cos(2t)(u(t + 1) – u(t – 1))*(u(t + 1) – u(t – 1))
f) y(t) = e-3tu(t)*(u(t +2) – u(t))
49/55
 2.10 Problems

Prob. 2.8 For each of the following impulse responses, determine wheter the
corresponding system is memoryless, causal, and stable
a) h(t) = eatu(t) b) h(t) = cos(t)
c) h(t) = e-2tu(t – 1) d) h(t) = u(t + 1)
e) h(t) = 3(t) f) h(t) = cos(t)u(t)
Prob. 2.9 Let x(t) = u(t – 3) – u(t-5) and h(t) = e-3tu(t)
a) Compute y(t) = x(t)*h(t)
b) Compute g(t) = (dx(t)/dt)*h(t)
c) Check by using Matlab

50/55
 2.11 Solution

Prob. 2.1 If y(t) = x(t)*h(t), show that y(2t) = 2x(2t)*h(2t).

Solution

y (t )   x( )h(t   )d

 
y (2t )   x( )h(2t   )d   x( )h[2(t   / 2)]d
 

Let  = /2, d = 2d


y (2t )   2 x(2 )h[2(t   )]d  2 x(2t ) * h(2t )


51/55
 2.11 Solution
Prob. 2.2 If x(t) & h(t) are odd, show that y(t) = x(t)*h(t) is an even signal.
Solution
y (t )  x(t ) * h(t )  [ x(t )]*[h(t )]  x(t ) * h(t )  y (t )
Prob. 2.3 If the response of an LTI system to x(t) is y(t), show that the response
to dx/dt is dy/dt. Using this result to determine the impulse response of an LTI
system having the response y(t) = sin(2t) for an input x(t) = e-4tu(t).
 
Solution dy dx(t   ) dx
y (t )   h( ) x(t   ) d     h( ) d  * h(t )

dt  dt dt
h (t )
x1 (t )  e-4t u (t ) 
 y1 (t )  sin(2t )
dx(t )
x2 (t )   4e4t u (t )  e 4t (t )
dt
4t dy (t )
u (t )   (t ) 
h (t )
 4e  y2 (t )   2 cos(2t )
dt
4 x1 (t )  x2 (t )   (t ) 
h (t )
 4sin(2 t)  2 cos(2 t) 52/55
 2.11 Solution

Prob. 2.4 Consider a cont.-time LTI system with the unit step response s(t).
a) deduce that the response y(t) of the system to the input x(t) is

dx(t   ) dx(t )
y (t)   s( )
dt
d 
dt
* s(t )
Solution 
s (t )  h(t ) * u (t )  s '(t )  h(t ) * u '(t )  h(t ) *  (t )  h(t )
 y (t )  h(t ) * x(t )  s '(t ) * x(t )  x '(t ) * s(t )
b) and also show that 
dx(t   ) dx(t )
x(t )   u ( )
dt
d 
dt
* u (t )


du (t ) dx(t )
x(t )  x(t ) *  (t )  x(t ) *  * u (t )
dt dt

53/55
 2.11 Solution

Prob. 2.5 Using the result of 2.4 to determine y(t) given s(t) = (e-2t – e-t +1)u(t)
and x(t) = etu(t).
Solution
We have shown that y(t) = x’(t)*s(t), and thus

 
y (t)  s (t ) * et u (t )   (t )  s (t ) * x(t )  s (t )
t
=  (e2  e  1) e (  t) d  s(t )
0

Use Matlab to compute the above integral… we have

1 t 1 2t 5 t
y (t )  e  e  e  1  s (t )
2 3 6

54/55
 2.11 Solution

Prob. 2.6 For each of the following impulse responses, determine whether the
corresponding system is memoryless, causal, and stable
a) h(t) = u(t + 1) – u(t – 1) b) h(t) = u(t) – 2u(t – 1)
c) h(t) = e-2|t| d) h(t) = eatu(t)
Solution
a) not mles, not cau, stable b) not mles, cau, not sta
c) not mles, not cau, sta

55/55