Fourier Series

Content
Periodic Functions
Fourier Series
Complex Form of the Fourier Series
Impulse Train
Analysis of Periodic Waveforms
Half-Range Expansion
Least Mean-Square Error Approximation
Fourier Series

Periodic Functions
The Mathematic Formulation
Any function that satisfies

f (t ) f (t T ) for all t

where T is a constant and is called the period

of the function.
 Example:
t t
f (t ) cos cos Find its period.
3 4
t t 1 1
f (t ) f (t T ) cos cos cos (t T ) cos (t T )
3 4 3 4

Fact: cos cos( 2m)

T
2m T 6m
3 T 24 smallest T
T
2 n T 8n
4
 Example:
f (t ) cos 1t cos 2t Find its period.

f (t ) f (t T ) cos 1t cos 2t cos 1 (t T ) cos 2 (t T )

1T 2m 1 m 1 must be a

2 n 2 rational number
2T 2n
Example:
f (t ) cos 10t cos(10 )t
Is this function a periodic one?

1 10 not a rational

2 10 number
Sinusoidal Steady State Response
(Response to a complex exponential sequence)

x(t) = est LTI system y(t) = ?

with h(t)

H(s) => gain given by system

If a complex exponential function is given as input to a LTI system wit
h impulse response h(t), then the output is scaled version of input.

est H(s)LTI
est h(t)
i.e the response of an LTI system to a complex exponential input is th
e same complex exponential with only a change in amplitude

Also suppose x(t) = aes1t + bes2t a,b being constants;

aes1t + bes2t aH(s1) es1t + bH(s2) es2t
LTI h(t)

If input x(t) is given as sum of complex exponentials, then output is al

so written as weighted sum of same complex exponentials.
Fourier Representation of signals
Study of signals using sinusoidal representations is termed Fourier analy
sis.
Convolution integral/Convolution sum
Input was given as weighted sum of time shifted impulses.
Output was given by weighted superposition of timeshifted impulse responses in time do
main.
Next, consider representing a signal as a weighted sum of complex sinus
oids. Then output of a LTI system is weighted sum of system response to
each complex sinusoid.
If the input to an LTI system is expressed as a linear combination of peri
odic complex exponentials or sinusoids, the output can also be expressed in t
his form.
Signals can be represented using complex exponentials continuous-ti
me and discrete-time Fourier series and transform.
Fourier Representation of signals

For a CT LTI system, consider the complex exponential

x(t) = e jt (where s=
then, y(t) = H(j) e jt
where frequency response H(j) =

For a DT LTI system, If x[n] = ejn

then y[n] = H(ej) ejn
where frequency response,
H(j) =
Fourier Representation of signals

4distinct Fourier representations based on peri

odicity of signals
Time Property Periodic NonPerodic
Continuous Continuous Time Fourier
Fourier Transform(CTFT)
Series(CTFS)
Discrete Discrete Time Discrete Time
Fourier Fourier
Series(DTFS) Transform(DTFT)

Discrete Fourier
Transform (DFT)
Continuous Time F
ourier Series
Fourier Series
Analysis of CT
periodic Signals
Introduction
Decompose a periodic input signal into pri
mitive periodic components.

AAperiodic
periodicsequence
sequence
f(t)

t
T 2T 3T
Fourier Series representation of Contin
uous-Time Periodic Signals

A signal that repeats itself at regular intervals o

f time i.e periodically over -
3 important classes of Fourier Series
Trigonometric form
Cosine form
Exponential form
 We introduced two basic periodic signals

the sinusoidal signal

and the periodic complex exponential

Both these signals are periodic with fundamental frequency and fundame
ntal period T = 2 /
Fourier showed that these signals can be represented by a
sum of scaled sines and cosines at multiples of the fundamental frequenc
y. The series can also be expressed as sums of scaled complex exponent
ials at multiples of the fundamental frequency. A sinusoid at frequency n
0 is called an nth harmonic.
Convergence of the Fourier Series

For the Fourier series to exist for a periodic signal:

The conditions under which a periodic signal can be represented
by a Fourier series are known as Dirichlets conditions.
In each period,
1. The function x(t) must be a single valued function.
2. The function x(t) has only a finite number of maxima and minima.
3. In any finite interval of time, the signal x(t) has a finite number of disc
ontinuities.
4. x(t) must be absolutely integrable over one period, that is
Trigonometric form of Fourier Series
Let the signal x(t) be, a sum of sine and cosine functions whose
frequencies are integral multiples of fundamental frequency 0
+ cos 0t + cos 20t +. + ak cos k0t
+ sin 0t + sin 20t +.+ sin k0t
=+
where a0,a1ak and b1,b2bk are constants.
It satisfies the periodicity condition x(t) = x(t + T)
If k in x(t), we obtain the Fourier series representation of any periodic
signal x(t).
Synthesis

2nt 2nt
x(t ) a 0 a n cos
DC Part Even Part bn sin Odd Part
n 1 T n 1 T
T is a period of all the above signals

Let 0=2/T.

x(t ) a 0 a n cos(n 0 t ) bn sin( n 0 t )
n 1 n 1
Evaluation of Fourier coefficients
The constants a0,a1,,an,b1,b2,bn are called Fourier coefficients.
To evaluate a0, integrate both sides of Eq.(1) over one
period (t0, t0 +T ) of x(t)
Orthogonal Functions
a set of functions {k} orthogonal
Call
on an interval a < t < b if it satisfies

b 0 mn
a m (t ) n (t )dt
rn mn
 Orthogonal set of Sinusoidal Functi
ons

Define 0=2/T.
T /2 T /2

T / 2
cos( m0t )dt 0, m0
T / 2
sin( m0t )dt 0, m0

T /2 0 mn
T / 2 cos(m0t ) cos(n0t )dt T / 2 m n We now prove this one
T /2 0 mn
T / 2 sin( m0t ) sin( n0t )dt T / 2 m n
T /2

T / 2
sin( m0t ) cos( n0t )dt 0, for all m and n
Decomposition

x(t ) a 0 a n cos(n 0 t ) bn sin( n 0 t )
n 1 n 1

1 t 0 T
a0
T
t0
x(t )dt

2 t 0 T
a n x(t ) cos n 0 tdt n 1,2,
T t0
2 t 0 T
bn x(t ) sin n 0 tdt n 1,2,
T t0
 Proof
Use the following facts:
T /2 T /2

T / 2
cos( m0t )dt 0, m0
T / 2
sin( m0t )dt 0, m0

T /2 0 mn
T / 2 cos(m0t ) cos(n0t )dt T / 2 m n
T /2 0 mn
T / 2 sin( m0t ) sin( n0t )dt T / 2 m n
T /2

T / 2
sin( m0t ) cos( n0t )dt 0, for all m and n
 Example (Square Wave)
f(t)
1

-6 -5 -4 -3 -2 - 2 3 4 5

2
a0
2 0
1dt 1

2 1
an
2 0
cos ntdt
n
sin nt 0
0 n 1,2,

2 1 1 2 / n n 1,3,5,
bn
2 0
sin ntdt
n
cos nt 0
n
(cos n 1)
0 n 2,4,6,
 1 2 1 1
f (t ) sin t sin 3t sin 5t
2 3 5
Example (Square Wave)
f(t)
1

-6 -5 -4 -3 -2 - 2 3 4 5

2
a0
2 0
1dt 1

2 1
an
2 0
cos ntdt
n
sin nt 0
0 n 1,2,

1 1 1 2 / n n 1,3,5,
bn
2 0
sin ntdt cos nt 0
n n
(cos n 1)
0 n 2,4,6,
 1 2 1 1
f (t ) sin t sin 3t sin 5t
2 3 5
Example (Square Wave)
f(t)
1

-6 -5 -4 -3 -2 - 2 3 4 5

2
a0
1.5 2
0
1dt 1

1
2 1
an 0.5 cos ntdt sin nt 0 0 n 1,2,
2 0 n
0
1 1 1 2 / n n 1,3,5,
5 sin ntdt cos nt 0
bn -0. (cos n 1)
2 0 n n 0 n 2,4,6,
Harmonics

2nt 2nt
x(t ) a 0 a n cos bn sin
n 1 T n 1 T

x(t ) a 0 a n cos(n 0 t ) bn sin( n 0 t )
DC Part Even Part Odd Part
n 1 n 1

T is a period of all the above signals

 Harmonics
a0
f (t ) an cos n t bn sin n t
2 n 1 n 1

a0
(an cos nt bn sin n t )
2 n 1
a0 an bn
2
an bn
2
cos nt sin nt
2 n 1 an2 bn2 an2 bn2

a0
an2 bn2 cos n cos n t sin n sin n t
2 n 1

C0 Cn cos( nt n )
n 1
Amplitudes and Phase Angles

f (t ) C0 Cn cos(nt n )
n 1

harmonic amplitude phase angle

a0
C0
2 bn
n tan 1

Cn an2 bn2 an
Fourier Series

Complex Form of the

Fourier Series
Complex Exponentials
jn 0t
e cos n 0 t j sin n 0 t
jn 0t
e cos n 0 t j sin n 0 t

1 jn0t
cos n 0 t e
2

e jn0t
sin n 0 t
1 jn0t
2j
e e
jn 0t
e
2

j jn0t

e jn0t
 Complex Form of the Fourier Series

x(t ) a 0 a n cos n 0 t bn sin n 0 t
n 1 n 1

a0 1
an e
2 2 n 1
jn0t
e jn0t j

bn e jn0t e jn0t
2 n 1

a0 1 1
(an jbn )e jn0 t
(an jbn )e jn0t
2 n 1 2 2 a0
c0
2

c0 cn e jn0t c n e jn0t
1
n 1 cn (an jbn )
2
1
c n (an jbn )
2
 Complex Form of the Fourier Series

x(t ) c0 c n e jn0t c n e jn0t
n 1

1
c0 cn e jn 0t
c n e jn0t
n 1 n
a0
c0
n
c e
n
jn 0t 2
1
cn (an jbn )
2
1
c n (an jbn )
2
Complex Form of the Fourier Series
a0 1 T /2
c0
2 T T / 2
f (t )dt

1
cn ( an jbn )
2
1 T /2 T /2

T T / 2 T / 2
f (t ) cos n tdt j f (t ) sin n tdt
0 0
1 T /2 a0
f (t )(cos n0t j sin n0t )dt c0
T T / 2 2
1 T /2 1
f (t )e jn0t dt cn (an jbn )
T T / 2 2
1 1 1
T /2
c n (an jbn )
c n (an jbn )
2 T T / 2
f (t )e jn0t dt
2
 Complex Form of the Fourier Series
a0
Equation of exponential c0
x(t ) n
c e jn 0 t

n
Fourier series 2
1
cn (an jbn )
1 t 0 T where cn Fourier 2

jn 0t
cn x(t )e dt coefficient of exponential 1
T t0 Fourier series c n (an jbn )
2
cn | cn | e jn , c n cn* | cn | e jn
If f(t) is real,
1 2
| cn || c n | an bn2
c n cn* 2 n 1,2,3,
b 1
n tan 1 n c0 a0
an 2
Complex Frequency Spectra
cn | cn | e jn , c n cn* | cn | e jn b
n tan 1 n n 1,2,3,
1 2 an
| cn || c n | an bn2
2
|cn| amplitude
1 spectrum
c0 a0
2

phase
n spectrum

 Example
f(t)
A

t
T T d d T T

2 2 2 2

A d /2 A 1
cn e jn0t dt (2 j sin n0 d / 2)
T d / 2 T jn0
d /2 A 1
A 1 1 sin n0 d / 2
e jn0t T 2 n 0
T jn0 d / 2
nd
sin
Ad T
A 1 1 jn0 d / 2
e jn0 d / 2
e T nd
T jn0 jn0
T
Example
A/5

-120 -80 -40 0 40 80 120

-150 -100 -50 50 100 150

nd 1 1 d 1
sin d , T ,
Ad T 20 4 T 5
cn
T nd 2
0 8
T T
Example
A/10

-120 -80 -40 0 40 80 120

-300 -200 -100 100 200 300

nd 1 1 d 1
sin d , T ,
Ad T 20 2 T 5
cn
T nd 2
0 4
T T
Example
f(t)
A

t
T 0 d T

A d A 1
cn e jn0t dt (1 e jn0 d )
T 0 T jn0
d A 1 jn0 d / 2 jn0 d / 2 jn0 d / 2
A 1 e (e e )
e jn0t T jn0
T jn0 0
nd
sin
Ad T e jn0 d / 2
A 1 1
e jn0 d
T nd
T jn0 jn0
T
Fourier Series

Impulse Train
Dirac Delta Function
0 t0
(t )
t 0
and
(t )dt 1

Also called unit impulse function.

t
0
Property

(t )(t )dt (0) (t): Test Function

(t )(t )dt (t )(0)dt (0) (t )dt (0)

Impulse Train

3T 2T T 0 T 2T 3T t

T (t ) (t nT )
n
 Fourier Series of the Impulse Train

T (t ) (t nT )
n
2 T /2
a0 T (t )dt
T T / 2 T
2

2 T /2 2
an T (t ) cos( n0t )dt
T T / 2 T
2 T /2
bn T (t ) sin( n0t )dt 0
T T / 2
1 2
T (t ) cos n0t
T T n
 Complex Form
Fourier Series of the Impulse Train

T (t ) (t nT )
n
a0 1 T /2 1
c0
2 T T / 2 T (t )dt T
1 T /2 1

jn0t
cn T (t )e dt
T T / 2 T
1 jn0t
T (t ) e
T n
Fourier Series

Analysis of
Periodic Waveforms
Waveform Symmetry
Even Functions
x(t ) x(t )
Odd Functions

x(t ) x(t )
Decomposition
Anyfunction x(t) can be expressed as the
sum of an even function xe(t) and an odd f
unction xo(t).
x(t ) xe (t ) xo (t )
xe (t ) 12 [ x(t ) x(t )] Even Part

xo (t ) 12 [ x(t ) x(t )] Odd Part

 Example
e t t0
x(t )
0 t0

1 t Even Part
e t0
xe (t ) 2

1 t
2e t0

Odd Part
e t 1
t0
xo (t ) 1 t 2

2e t0
Half-Wave Symmetry
x(t ) x(t T ) and x(t ) x t T / 2

T/2 T/2 T
 Quarter-Wave Symmetry
Even Quarter-Wave Symmetry
x(t ) x(t ) and x(t ) x t T / 2

T/2 T/2 T

Odd Quarter-Wave Symmetry

x(t ) x(t ) and x(t ) x t T / 2
T/2 T/2
T
Hidden Symmetry
The following is a asymmetry periodic function:
A

T T
Adding a constant to get symmetry property.
A/2

T T
A/2
Fourier Coefficients of
Symmetrical Waveforms

The use of symmetry properties simplifies the c

alculation of Fourier coefficients.
Even Functions
Odd Functions
Half-Wave
Even Quarter-Wave
Odd Quarter-Wave
Hidden
Fourier Coefficients of Even Functions

x(t ) x(t )

x(t ) a 0 a n cos n 0 t
n 1

4 T /2
a n x(t ) cos(n 0 t )dt
T 0
Fourier Coefficients of Even Functions

x(t ) x(t )

x(t ) bn sin n 0 t
n 1

4 T /2
bn x(t ) sin( n 0 t )dt
T 0
Fourier Coefficients for Half-Wave Symmetry

x(t ) x(t T ) and x(t ) x t T / 2

T/2 T/2 T

The
The Fourier
Fourier series
series contains
contains only
only odd
odd harmonics.
harmonics.
Fourier Coefficients for Half-Wave Symmetry

x(t ) x(t T ) and x(t ) x t T / 2

x(t ) (a n cos n 0 t bn sin n 0 t )
n 1

0 for n even
an 4 T / 2

T 0 x (t ) cos( n 0 t ) dt for n odd

0 for n even
bn 4 T / 2

T 0 x (t ) sin( n 0 t ) dt for n odd
Fourier Coefficients for
Even Quarter-Wave Symmetry

T/2 T/2 T

x(t ) a 2 n 1 cos[(2n 1) 0 t ]
n 1

8 T /4
a 2 n 1
T
0
x(t ) cos[(2n 1) 0 t ]dt
Fourier Coefficients for
Odd Quarter-Wave Symmetry

T/2 T/2
T

x(t ) b2 n 1 sin[(2n 1) 0 t ]
n 1

8 T /4
b2 n 1
T 0
x(t ) sin[(2n 1) 0 t ]dt
Example
Even Quarter-Wave Symmetry
1
T/2 T/2
T T/4 T/4 T
1

8 T /4 8 T /4
a2 n 1
T
0
f (t ) cos[( 2n 1)0t ]dt
T
0
cos[( 2n 1)0t ]dt

T /4
8 n 1 4
sin[( 2n 1)0t ] (1)
(2n 1)0T 0
(2n 1)
 4 1 1
f (t ) cos 0t cos 30t cos 50t
3 5
Example
Even Quarter-Wave Symmetry
1
T/2 T/2
T T/4 T/4 T
1

8 T /4 8 T /4
a2 n 1
T
0
f (t ) cos[( 2n 1)0t ]dt
T
0
cos[( 2n 1)0t ]dt

T /4
8 n 1 4
sin[( 2n 1)0t ] (1)
(2n 1)0T 0
(2n 1)
Example
Odd Quarter-Wave Symmetry
1
T/2 T/2
T T/4 T/4 T
1

8 T /4 8 T /4
b2 n 1
T
0
f (t ) sin[( 2n 1)0t ]dt
T
0
sin[( 2n 1)0t ]dt

T /4
8 4
cos[( 2n 1)0t ]
( 2n 1)0T 0
(2n 1)
 4 1 1
f (t ) sin 0t sin 30t sin 50t
3 5
Example
Odd Quarter-Wave Symmetry
1
T/2 T/2
T T/4 T/4 T
1

8 T /4 8 T /4
b2 n 1
T
0
f (t ) sin[( 2n 1)0t ]dt
T
0
sin[( 2n 1)0t ]dt

T /4
8 4
cos[( 2n 1)0t ]
( 2n 1)0T 0
(2n 1)
Fourier Series

Half-Range
Expansions
Non-Periodic Function Representation

A non-periodic function f(t) defined over (0, ) c

an be expanded into a Fourier series which is d
efined only in the interval (0, ).
Without Considering Symmetry

A non-periodic function f(t) defined over (0, ) c

an be expanded into a Fourier series which is d
efined only in the interval (0, ).
Expansion Into Even Symmetry

T=2

A non-periodic function f(t) defined over (0, ) c

an be expanded into a Fourier series which is d
efined only in the interval (0, ).
Expansion Into Odd Symmetry

T=2

A non-periodic function f(t) defined over (0, ) c

an be expanded into a Fourier series which is d
efined only in the interval (0, ).
Expansion Into Half-Wave Symmetry

T=2

A non-periodic function f(t) defined over (0, ) c

an be expanded into a Fourier series which is d
efined only in the interval (0, ).
Expansion Into
Even Quarter-Wave Symmetry

T/2=2
T=4

A non-periodic function f(t) defined over (0, ) c

an be expanded into a Fourier series which is d
efined only in the interval (0, ).
Expansion Into
Odd Quarter-Wave Symmetry

T/2=2 T=4

A non-periodic function f(t) defined over (0, ) c

an be expanded into a Fourier series which is d
efined only in the interval (0, ).
Fourier Series

Least Mean-Square
Error Approximation
Approximation a function
a0 k
Use S k (t ) an cos n0t bn sin n0t
2 n 1
to represent f(t) on interval T/2 < t < T/2.

Define k (t ) f (t ) S k (t )
1 T /2 Mean-Square
Ek [ k (t )] dt
2

T T / 2 Error
Approximation a function
Show that using Sk(t) to represent f(t) has
least mean-square property.
1 T /2
Ek [ k (t )]2 dt
T T / 2
2
1 a0k

T / 2 f (t ) 2 an cos n0t bn sin n0t dt
T /2

T n 1

Proven by setting Ek/ai = 0 and Ek/bi = 0.

 Approximation a function
1 T /2
Ek
T T / 2
[ k (t )]2 dt
2
1 a0 k


T /2

T T / 2 f (t )
2 n 1
a n cos n 0 t bn sin n t
0

dt

Ek a0 1 T /2 Ek 2 T /2
a0

2 T
T / 2
f (t )dt 0
an
an
T T / 2
f (t ) cos n0tdt 0

Ek 2 T /2
bn f (t ) sin n0tdt 0
bn T T / 2
Mean-Square Error
1 T /2
Ek
T T / 2
[ k (t )]2 dt
2
1 a0 k


T /2

T T / 2 f (t )
2 n 1
a n cos n 0 t bn sin n t
0

dt

k2
1 a 1
T / 2 [ f (t )] dt 4 2
T /2
Ek 2 0
( an bn )
2 2

T n 1
Mean-Square Error
1 T /2
Ek
T T / 2
[ k (t )]2 dt
2
1 a0 k


T /2

T T / 2 f (t )
2 n 1
a n cos n 0 t bn sin n t
0

dt

k 2
1 a 1
T / 2 [ f (t )] dt 4 2
T /2
2
(an bn )
0 2 2

T n 1
Mean-Square Error
1 T /2
Ek
T T / 2
[ k (t )]2 dt
2
1 a0 k


T /2

T T / 2 f (t )
2 n 1
a n cos n 0 t bn sin n t
0

dt

2
1 a 1
T / 2 [ f (t )] dt 4 2
T /2
2
(an bn )
0 2 2

T n 1