1

Ship encountering the superposition of 3 waves.

2
Fourier Series Representation of
Periodic Signals
•Signals can be represented as linear
combinations of basic signals with
the following 2 properties.
•The set of basic signals can be used
to construct a broad and useful class
of signals.
•The response of an LTI system is a
combination of the responses to these
basic signals at the input. 3
 Fourier Analysis
• Both properties are provided for by the
complex exponential signals in the
continuous and discrete time
• I.e. Signal of the form:-
e st in continuous time.
z n in the discrete time.
where s & z are complex numbers.
4
The principle of superposition for
linear systems.
x(t) y(t)
LTI System
x[n] y[n]

if x(t)  a11 (t )  a 2 2 (t )  ......

 k (t )   k (t )
and systemis linear
Then : y(t)  a11 (t )  a 2 2 (t )  ...
Similarly for the discrete - time case.
5
Criteria for choosing a set of basic signals in
terms of which to decompose the input
if x(t)  a11 (t )  a 2 2 (t )  ......
 k (t )   k (t )
and systemis linear
Then : y(t)  a11 (t )  a 2 2 (t )  ...
Choose  k (t ) or  k [n] so that :
- a broad class of signals can be
constructed as a linear combinatio n
of  k ' s.
- responses to  k ' s are easy to compute 6
 Choice for basic signals that led to the
convolution integral and convolution sum for
LTI Systems
C - T :  k (t )   (t  k)
 k (t )  h(t  k)
 Convolution Integral
D - T :  k [ n]   [ n  k ]
 k [n]  h[n  k ]
 Convolution Sum
7
Complex exponentials as a set of basic signals

 k (t )  e sk t
s k complex
 k [ n]  z k n
z k complex
Fourier Analysis :
j k t
C - T : s k  j k  k (t )  e
D - T : | z k | 1  k [ n]  e j k n

s k complex  Laplace transforms

z k complex  z - transforms.
8
Eigenfunction (est , zn) & Eigenvalues (H(s) , H(z))

x(t) y(t)
LTI System
x[n] y[n]

continuous time : x(t)  e st  y (t )  H ( s )e st

discrete time : x[n]  z n  y[n]  H(z)zn
The importance of complex exponentials in the
study of LTI systemsstems from the fact that
the response of an LTI systemto a complex
exponential input is the same complex exponential
with only a change in amplitude. 9
 Showing complex exponential as
eigenfunction of system
From the convolution integral with input x(t)  est

y (t )   h( ) x(t   )d


  h( )e s ( t  )
d .
-


 s
e st
h( )e d .


Assuming that theintegral on RHS converge,

the reponse to e is of the form y(t)  H(s)e where
st st


H(s) is a complex constant   h( )e  s
d . 10

Showing complex exponential as
eigenfunction of system
Similarly for the D - Time case : -
From the convolution sum with input x[n]  z n ,

y[n]   h[k ]x[n  k ]


  h[k ]z n  k


z n


h[ k ] z k

Assuming that thesummation on RHS converges,

the reponse to z n is of the form y[n]  H(z)zn where

H(z) is a complex constant   h[k ]z  k 11

 Example x(t) to be linear combination of 3
complex exponentials
x(t)  a1e  a 2 e  a 3e
s1t s2t s3t

From the eigenfunct ion property,

the response to each separately is : -
a1e  a1 H ( s1 )e
s1t s1t

a 2e s2t
 a 2 H ( s 2 )e s2t

a 3e s 3t
 a 3 H ( s3 )e s 3t
12
 Example x(t) is linear combination of 3
complex exponentials
x(t)  a1e s1t  a 2 e s2t  a 3e s3t
From the superposition property,
the response to the sum is the sum of the responses : -
y(t)  a1 H ( s1 )e s1t  a 2 H ( s2 )e s2t  a 3 H ( s3 )e s3t
Generally, if input x(t)   a k e sk t ,
k

then the output will be y(t)   a k H ( sk )e sk t

k

similarly for D - time if input x[n]   a k z k ,

n

then the output will be y[n]   a k H ( z k ) z k .

n
13
k
 Fourier Series Representation of Continuous-
time Periodic Signals.
A signal is periodic if x(t)=x(t+T), for all t.
T is the fundamental period.

2
0  , is the fundamental radian frequency.
T
We have studied before 2 basic periodic
signals, sinusoidal x(t)  cos( 0t )
j 0 t
and complex exponential x(t)  e ,
14
Associated with this basic complex exponential signal is
the set of harmonical ly related complex exponentials,
2
jk t
 k (t )  e jk t  e
0 T
, k  0,1,2,...
Each of these signals has a fundamental frequency that
is a multiple of  0, and each is periodic with T ( for k  2,
the fundamental period is a fraction of T).
Thus, a linear combinatio n of harmonical ly related complex
  2

 k a e
jk 0 t
jk 0t
exponential x(t)  a e  k
T
, is the Fourier
k -  k -

series and is also periodic with period T. k  1 is the first harmonic

component, k  2 is the second harmonics components.

15
Example 3.2
3
x(t )   k ,
a e jk 2t

k  3

1 1 1
where a0  1, a1  a1  , a2  a 2  , a3  a3  .
4 2 3
1 j
use Euler relationship, cos  (e  e - j ),
2
to get the cosine form below : -
1 2
x(t )  1  cos 2t  cos 4t  cos 6t.
2 3
Figure 3.4 pg188 OWN illustrates how these cosine waveforms
can be added up to construct a peroidic signal x(t) above.
16
17
18
19
 Trigonometric forms of Fourier Series.

x(t )   k ........eqn.3.25 (Synthesis Equation)
a e
k  
jk0t

is the Complex Exponential Form of

the Fourier Series .
Taking conjugatio n of both side of eqn. 3.25, we have : -

x (t) 
*
a e
k  
*  jk0t
k , For real x(t), x * (t)  x(t).

 x(t)   k e , replacing k by - k,
a *  jk0t

k  

x(t)   k and comparing with eqn. 3.25,
a *
e
k  
jk0t

ak  a-*k , or ak*  a k .

From Eqn 3.25 : - x(t )  a0   [ak e jk0t  a k e  jk0t ]
k 1 20
 Trigonometric forms of Fourier Series.
 
x(t )  a0   [ak e jk0t  a *e  jk0t ]  a0   2 Re{a k e jk0t }.
k 1 k 1
j k
Expressing a k  Ak e , (polar form)

x(t )  a0   2 Re{Ak e j ( k0t  k ) }.
k 1

 x(t)  a 0  2 A k cos(k0t   k ) ,
k 1

the other form by letting a k  Bk  jC k ,


x(t )  a0  2 [ Bk cos k0t Ck sin k0t ].
k 1
21
Solving for Fourier Series Coefficients(Analysis)
M ultiplying both side of the Fourier Series Synthesis eqn by e  jn 0t : -

x(t )e  jn 0t
  k e .
a e jk 0t  jn 0t

k  

2
Integrating both sides of equation from 0 to T  , we have
0
T 
 k e dt
T
 dt  
 jn 0t jk 0t  jn 0t
x(t )e a e
0 0
k  



T T
 a k [  e jk 0t e  jn 0t dt ]
 jn 0t
x(t )e dt 
0 0
k  



T T
 a k [  e j ( k  n ) 0t dt ] .........equation 3.34
 jn 0t
x(t )e dt 
0 0
k  
T T T
However  e j ( k  n ) 0 t
dt   cos(k  n) 0 tdt  j  sin( k  n) 0 tdt
0 0 0
22
 Solving for Fourier Series Coefficients(Analysis)

For k  n, cos(k  n)0 t and sin(k  n)0 t are periodic sinusoids

T
with fundamental period ( ).
|k-n |
The integratio ns over whole integer numbers of fundamental periods
for sine and cosine term are zeros.
For k  n, cos(k  n) t  1and sin(k  n) t  0
0 0
T j(k  n) t
 0 e 0 dt   T 1dt  j T 0dt  T
0 0
Overall we have : -
T e j(k  n)0 t dt  T, k  n
0
 0. k  n
From Equ 3.34 taking k  n : -
.
1 T  jn  t
 a  0 x(t)e 0 dt
n T 23
 Summarizing for Fourier Series Pair
Representation
Fourier Series Synthesis:- eqn 3.38
  2

a e a e
jk t
jk 0t
x(t )  k  k
T

k   k  
Fourier Series Analysis or F.S. Coefficients
0r Spectral Coefficents:- eqn 3.39
2
1  jk  t 1  jk t
a   x(t)e 0 dt   x(t)e T dt
k T TT
T
24
Example 3.4

x(t )  1  sin  0t  2 cos  0t  cos(2 0t  )
4
1 j
use Euler relationship for cos  (e  e - j )
2
1 j
and sin  (e  e - j ) : -
2j
1 j 0 t  j 0 t j 0 t  j 0 t 1 j 2 0t  / 4  j 2 0t  / 4
x(t)  1  [e  e ]  [e  e ]  [e e ]
2j 2
1 j 0 t 1  j 0t 1 j ( / 4 ) j 2 0t 1  j ( / 4 )  j 2 0t
x(t )  1  (1  )e  (1  )e  e )e  e )e
2j 2j 2 2
 Fourier series coefficien ts are : -
1 1
a0  1, a1  (1  ), a1  (1  ),
2j 2j
1 1
a2  e j ( / 4 ) , a 2  e  j ( / 4 ) , ak  0 for | k | 2.
2 2
25
 Magnitude of Fourier Coefficients.
1.2

0.8

| ak | 0.6

0.4

0.2

0
k= -3 -2 -1 0 1 2 3 4

26
 Phase of Fourier Coefficients.
1
0.8
0.6
0.4
angle ak 0.2
0
-0.2 k= -3 -2 -1 0 1 2 3 4
-0.4
-0.6
-0.8
-1

27
 Example 3.5

-T/2 -T1 T1 T/2

This periodic signal x(t) repeats every T seconds.
x(t)=1, for |t|<T1 , and x(t)=0, for T1 <|t|< T/2

Fundamental period= T,
Fundamental frequency 0 2/T.
Choosing the period of integration to be between
-T/2 and +T/2. Use eqn 3.39 to get at Fourier
Series Coefficients. 28
 Example 3.5 continued
2
1  jk  t 1  jk t
a   x(t)e 0 dt   x(t)e T dt
k T TT
T

Let us get the dc, constant term or average value

over a period, first, i.e. k  0
1 T1 2T1
a   dt 
0 T T1 T
29
 Example 3.5 continued
2
1  jk  t 1  jk t
a   x(t)e 0 dt   x(t)e T dt
k T TT
T

For fundamental first order and higher order harmonics : -

we have k  0.
1 T1  jk  0 t 1
a   e dt   e  jk 0t T1
|T1 ,
k T T1 jk 0T
jk 0T1  jk 0T1
2 e e 2 sin( k 0T1 ) sin( k 0T1 )
a  [ ] 
k k 0T 2j k 0T k
30
 Example 3.5 Continued
• In this example the coefficients are real
values.
• Generally the coefficients are complex.
• In this case we can represent a single plot of
magnitude of coefficient against k.
• Generally we will have the magnitude plot
and the phase plot of the coefficients.

31
 Example 3.5 Continued
2 
For T  4T1 ,  0T1  T1  ,
T 2
sin( k / 2)
From eqn 3.44 : - a k  , k  0,
k
1 1 1
 a1  a1  , a3  a3   , a5  a5  ,
 3 5
For even k' s the a' s are all zeros.
1
and from eqn 3.42 a 0  .
2
32
Plot of the Coefficients with T1
Fixed and T varied.
1
x
2

a) T=4T1

b) T=8T1

c) T=16T1

33