5/10/2015

Convolution Convolution
LTI System
• It is defined mathematically as

x(t ) h(t) y (t ) y (t )  x(t )  h(t )

Input Output
where  indicates the convolution process.
• Convolution is the interaction between:
• Input x(t)
• System impulse response h(t) • Or as

to produce: y (t )   x( )h(t   )d
• Output y(t) 

• In a Linear Time-Invariant (LTI) System. which is the Convolution Integral.

Convolution Convolution
Properties of Convolution: Properties of Convolution:
1. Associative: 2. Commutative:
x(t )  h1 (t )  h2 (t )  x(t )  h1 (t )  h2 (t )  x(t )  h2 (t )  h1 (t )  x(t )  h(t ) x(t )  h(t )  h(t )  x(t )

x(t) h1(t) h2(t)

x(t) h(t) = h(t) x(t)

x(t) h2(t) h1(t)

1
 5/10/2015

Convolution Convolution
Properties of Convolution: Example 1 – Continuous-time signals convolution.
3. Distributive: Sketch the signals below.
x(t )  h1 (t )  x(t )  h2 (t )  x(t )  h1 (t )  h2 (t )  x(t )  h(t )
x(t )  u (t )  u (t  2)
h(t )  u (t )  u (t  1)
x(t) h1(t)

Obtain the convolution of these two signals, and

+ = x(t) h2(t) + h1(t)
sketch the produced signal.
x(t) h2(t)

Convolution Convolution
Step 0: Sketch both signal completely. Step 1: Replace the time variable t with τ for both signals.
x(t)
Time Ranges
Signal
0 2
x(τ) h(τ)
u(t) 0 1 1
- u(t - 2) 0 0 -1 1

x(t) 0 1 0 t
0 1 2 3 4
1 1
h(t)
Time Ranges τ τ
Signal
0 1 0 1 2 3 4 0 1 2
u(t) 0 1 1
1
- u(t - 1) 0 0 -1
h(t) 0 1 0 t
0 1 2

2
 5/10/2015

Convolution Convolution
Step 2: Choose a simpler signal. Scale the signal with the Step 3: Shift the reversed signal with a time-shifting
time scaling factor α = -1 (i.e. time reversal operation). constant of t.

h(-τ) h(t - τ)

1 1

τ τ
-2 -1 0 t -1 t 0

Convolution Convolution
Step 4: Make sure the two signals required to apply the Step 5: Combine both signals in one plot.
convolution integral are obtained.

y (t )   x( )h(t   )d

x(τ) h(t - τ)
h(t - τ) x(τ)
1

τ
1 1 t-1 t 0 1 2 3 4

τ τ
0 1 2 3 4 t -1 t 0

3
 5/10/2015

Convolution Convolution
Step 6: Move the signal produced in step 3 toward the Step 6: continued…
positive direction of τ so that the two signals overlapped. t0
Solve the convolution integral within the overlapped limit.

h(t - τ) 1 x(τ)

τ
t-1 t 0 1 2 3 4

No overlap, therefore

y (t )  0

Convolution Convolution
Step 6: continued… Step 6: continued…
0  t 1
t
y (t )   x( )h(t   )d
0
t
h(t - τ) 1 x(τ)   (1)(1)d
0

t-1 0 t 1 2 3 4
τ  [ ]t0
t

Signals are overlapping from time 0 to t, therefore

4
 5/10/2015

Convolution Convolution
Step 6: continued… Step 6: continued…
1 t  2
t
y (t )   x( )h(t   )d
t 1
t
h(t - τ)
  (1)(1)d
1
x(τ) t 1

τ  [ ]tt 1
0 t-1 1 t 2 3 4
 t  (t  1)  1

Signals are overlapping from time t –1 to t, therefore

Convolution Convolution
Step 6: continued… Step 6: continued…
2t 3
2
y (t )   x( )h(t   )d
t 1
2
h(t - τ)
  (1)(1)d
1
x(τ) t 1

τ  [ ]t21
0 1 t-1 2 t 3 4
 2  (t  1)  t  3

Signals are overlapping from time t –1 to 2, therefore

5
 5/10/2015

Convolution Convolution
Step 6: continued… Step 7: Sketch the result completely.
3t Time Ranges, t
Signal
0 1 2 3
y(t) 0 t 1 -t+3 0

h(t - τ) y(t)
1
x(τ)
τ 3
0 1 2 t-1 3 t 4
2

1
No overlap, therefore

y (t )  0 0 1 2 3 t

Convolution – Discrete-time Signals Convolution – Discrete-time Signals

• It is defined mathematically as Example 2 – Discrete-time signals convolution.
Given two arbitrary sequences h[n] = [1,1] and x[n] = [1,1,2]
below:
y[n]  x[n]  h[n]
2
where  indicates the convolution process. h[n] x[n]
1 1 1 1

• Or as 0 1 -1 0 1
n n

y[n]   x[k ]h[n  k ]
k  

Obtain the convolution of these two sequences, and sketch

which is the Convolution Sum. the produced signal.

6
 5/10/2015

Convolution – Discrete-time Signals Convolution – Discrete-time Signals

Perform the convolution sum of the two sequences using Step 8: Sketch the result completely.
table:
k −2 −1 0 1 2
1. Replace variable n with k for both discrete signals. x[k] 1 1 2
2. Fill the values of amplitude of x[k] and h[k] in the table. h[k] 1 1
3. Scale h[k] with the time-scaling factor of α = -1 to obtain h[−k] 1 1
h[−k]. h[n−k] n y[n]
h[−1−k] 1 1 −1 1
4. Slide h[−k] to the left with shifting factor of n to obtain h[0−k] 1 1 0 2
h[n−k]. h[1−k] 1 1 1 3
5. Slide h[n−k] to the right for the value of n that results in h[2−k] 1 1 2 2
the overlapping of h[n−k] and x[k]. 3
6. Obtain y[n] using the Convolution Sum. y[n]
2 2
7. Repeat step 5 and 6 until there is no more overlap
between h[n−k] and x[k].
8. Sketch the result completely. -1 0 1 2 n

Systems
LTI System Type of System
• There are eight typical systems:
x(t ) h(t) y (t )
Input Output – Linear and nonlinear systems.
– Time-invariant and time-varying systems.
• Systems are used to process signals to allow:
– Modification of signals. – Dynamic and instantaneous systems.
– Extraction of additional information from the signals.
– Causal and noncausal systems.
• Systems can be realized through: – Stable and unstable systems.
– Hardware – physical components.
– Software – algorithm. – Continuous-time and discrete-time systems.
• Systems are studied in three major areas:
– Analog and digital systems.
– Mathematical modeling – rules governing the system.
– Analysis – find outputs based on given inputs and system rules.
– Design – construct system that produces desired outputs from given
inputs.

7
 5/10/2015

Linear and Nonlinear Systems Linear and Nonlinear Systems

• Linear system Nonlinear system is otherwise Properties of Linear System:
– Output is proportional to the input. 1. Additivity:
Output
x1 (t )  h(t )  x2 (t )  h(t )  x1 (t )  x2 (t ) h(t )  y1 (t )  y2 (t )

x1(t) h(t) x1(t)

Input
+ = + h(t)

– Satisfy the following properties:

x2(t) h(t) x2(t)
• Additivity
• Homogeneity (scaling)
• Superposition.

Linear and Nonlinear Systems Linear and Nonlinear Systems

Properties of Linear System: Properties of Linear System:
2. Homogeneity/Scaling: 3. Superposition:
x(t )  h(t )  y (t )  x(t )  h(t )  y (t ) x1 (t )  h(t )   x2 (t )  h(t )  x1 (t )  x2 (t ) h(t )  y1 (t )   y2 (t )

αx1(t) h(t) αx1(t)

x(t) h(t) y(t) → αx(t) h(t) αy(t) + = + h(t)

βx2(t) h(t) βx2(t)

8
 5/10/2015

Time-invariant and Time Varying Systems Dynamic and Instantaneous Systems

• Time-invariant system • Dynamic system
– Output does not depend explicitly on time. – System with memory.
Time varying system is otherwise – Output depends on both present and past inputs.
– Contain energy-storage elements.
x(t )  h(t )  y (t )  x(t  t0 )  h(t )  y (t  t0 ) – Characterized by differential equations.
i(t)
1 t
x(t) h(t) y(t) → x(t – t0) h(t) y(t – t0) v(t )  i (t )dt
+ C 
C v(t)
dv(t )
− i (t )  C
dt

Dynamic and Instantaneous Systems Causal and Noncausal Systems

• Instantaneous system • Causal system Noncausal system is otherwise
– Memory less or static system. – Output does not depend on the future input.
– Output depends on instantaneous input value.
y (t0 )  x (t )  h(t ), t  t0
– Contain resistive elements.
– Operate in steady state. – The output is produced after the input is
triggered.
i(t)

+
R v(t) v(t )  i (t ) R

9
 5/10/2015

Stable and Unstable Systems

• Stable system Unstable system is otherwise
– Bounded output is produced by every bounded
input.
– It is known as BIBO
(bounded-input/bounded-output)

10