Scribd
Upload
71 views

Signals and Systems Using Matlab Chapter 4 - Frequency Analysis: The Fourier Series

This document discusses frequency analysis using the Fourier series. It begins by introducing eigenfunctions of linear time-invariant systems and how the frequency response is equal to the Fourier transform evaluated at a given frequency. Examples are then provided to illustrate these concepts, including a low-pass RC filter. The document goes on to explain complex exponential and trigonometric Fourier series representations of periodic signals. It derives Fourier coefficients from the Laplace transform and discusses properties of even and odd periodic signals. Additional examples are used to demonstrate concepts such as integration, time and frequency shifting, and signal modulation.

Uploaded by

Dilu
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
71 views

Signals and Systems Using Matlab Chapter 4 - Frequency Analysis: The Fourier Series

This document discusses frequency analysis using the Fourier series. It begins by introducing eigenfunctions of linear time-invariant systems and how the frequency response is equal to the Fourier transform evaluated at a given frequency. Examples are then provided to illustrate these concepts, including a low-pass RC filter. The document goes on to explain complex exponential and trigonometric Fourier series representations of periodic signals. It derives Fourier coefficients from the Laplace transform and discusses properties of even and odd periodic signals. Additional examples are used to demonstrate concepts such as integration, time and frequency shifting, and signal modulation.

Uploaded by

Dilu
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 22

SIGNALS AND SYSTEMS USING MATLAB

Chapter 4 Frequency Analysis: The Fourier Series


Luis F. Chaparro

Eigenfunctions
x(t) = e j0t , < t < , input to a causal and stable LTI system
j0 t
steady state
output
y
(t)
=
e
H(j0)
Z
H(j0) =
h( )e j0 d = H(s)|s=j0
0

frequency response at 0
x(t) = e j0t is eigenfunction of LTI system
Example: RC circuit, voltage source be vs (t) = 4 cos(t + /4), R = 1 ,
C = 1F
Vc (s)
1
transfer function H(s) =
=
Vs (s) s + 1

2
H(j1) =
/4 frequency response at0 = 1
2

steady-state output vc (t) = 4|H(j1)| cos(t + /4 + H(j1)) = 2 2 cos(t)

2 / 22

Example: Lowpass filter using RC circuit


Input vs (t) = 1 + cos(10, 000t) to series RC circuit (R = C = 1)
dvc (t)
dt
jt
if input vs (t) = e output vc (t) = e jt H(j), then in o.d.e.
vs (t) = vc (t) +

e jt = e jt H(j)(1 + j) H(j) =

1
1
=
1 + j
1 + 2

vs (t) = cos(0t) + cos(10, 000t)


1
cos(10, 000t /2) 1
vc (t) 1 +
10, 000
attenuates higher frequency component (i.e., low-pass filter)

3 / 22

Complex exponential Fourier series


Fourier Series of periodic signal x(t), of fundamental period T0, is infinite sum of
orthonormal complex exponentials of frequencies multiples of fundamental
frequency 0 = 2/T0 (rad/sec) of x(t):
x(t) =

Xk e jk0t

k=

1
FS coefficients Xk =
T0

t0 +T0

x(t)e jk0t dt

t0

{e jk0t } are ortho-normal Fourier basis


Z t0+T0
Z t0+T0
1
1
e jk0t [e j`0t ]dt =
e j(k`)0t dt
T0 t0
T
0 t0
0 k 6= ` orthogonal
=
1 k = ` normal

4 / 22

Line spectrum

Parsevals power relation


Px : power of periodic signal x(t) of fundamental periodT0
Z t0 +T0

X
1
2
|Xk |2 ,
for any t0
|x(t)| dt =
Px =
T0 t0
k=

Periodic x(t) is represented in frequency by

Magnitude line spectrum |Xk | vs k0


Phase line spectrum Xk vs k0
Power line spectrum |Xk |2 vs k0
Real-valued periodic signal x(t), of fundamental period T0 ,

Xk = Xk
or equivalently
(i) |Xk | = |Xk |, i.e., magnitude |Xk | is even function of k0 .
(ii) Xk = Xk , i.e., phase Xk is odd function of k0

5 / 22

Trigonometric Fourier series

Real-valued, periodic signal x(t), of fundamental period T0,


x(t) =

X0
|{z}

+2

dccomponent

X
k=1

|Xk | cos(k0 t + k )
|
{z
}
k th harmonic

[ck cos(k0 t) + dk sin(k0 t)]

= c0 + 2

0 =

k=1

2
T0

Fourier coefficients {ck , dk }


ck
dk

Z t0 +T0
1
=
x(t) cos(k0 t) dt
T0 t0
Z t0 +T0
1
=
x(t) sin(k0 t) dt
T0 t0

k = 0, 1,
k = 1, 2,

Sinusoidal basis functions { 2 cos(k0t), 2 sin(k0t)}, k = 0, 1, , are


orthonormal in [0, T0]

6 / 22

Example: x(t) = B + A cos(0t + ) periodic of fundamental period T0


trigonometric Fourier series: X0 = B; |X1 | = A/2, X1 =
exponential Fourier series:
i
A h j(0 t+)
j(0 t+)
x(t) = B +
e
+e
2
Ae j
Ae j

X0 = B, X1 =
, X1 = X1 =
2
2
|Yk |

|Xk |
B

A
2

A
2

A
2
k0

Xk

A
2
k0

0
Yk

k0
0

k0

Line spectrum of x(t) = B + A cos(0 t) and of y (t) = B + A sin(0 t) (right).

7 / 22

Fourier coefficients from Laplace

x(t), periodic of fundamental period T0


period: x1(t) = x(t)[u(t t0) u(t t0 T0)], any t0
2
1
(fundamental frequency), k = 0, 1,
Xk = L [x1(t)]s=jk0 0 =
T0
T0
Example: x(t) periodic, T0 = 2, x1(t) = u(t) u(t 1)
x(t) =

m=

x1(t 2m) =

k=
jk

1
1e
Xk = L [x1(t)]s=jk =
2
jk

8 / 22

Xk e jkt
= e jk/2

sin(k/2)
k/2

Reflection and even and odd periodic signals


x(t) periodic of fundamental period T0, Fourier coefficients {Xk }
Reflection: Fourier coefficients of x(t) are {Xk }
Even x(t): {Xk } are real
x(t) = X0 + 2

Xk cos(k0t)

k=1

Odd x(t): {Xk } are imaginary


x(t) = 2

jXk sin(k0t)

k=1

Any periodic signal x(t) then x(t) = xe (t) + xo (t), xe (t) and xo (t) even and odd

components

Xk = Xek + Xok
Xek = 0.5[Xk + Xk ]
Xok = 0.5[Xk Xk ]

9 / 22

Example: periodic pulse train x(t), of fundamental period T0 = 1


x(t)
2

1.25 0.75

0.25

0.25

0.75

1.25

T0 = 1

Integral formula:

Laplace transform:

Fourier series:

1
Xk =
T0

1
X0 =
T0

3T0 /4

x(t)e j0 kt dt =

T0 /4
3T0 /4

1/4

x(t)dt =
T0 /4

sin(k/2)
, k 6= 0
(k/2)

2dt = 1
1/4

x1 (t 0.25) = 2[u(t) u(t 0.5)], X1 (s) = 2(e 0.25s e 0.25s )


sin(k/2)
1
Xk = L [x1 (t)] |s=jk0 =
k 6= 0
T0
k/2

X
sin(k/2) jk2t
x(t) =
e
(k/2)
k=

10 / 22

period of x(t)

real X(k)
1

2
1.5

0.5
Xk

0.5
0
0

0.2

0.4

0.6

0.8

0.5

100

50

100

phase line spectrum

0.8

2
1

0.6

Xk

|Xk|

magnitude line spectrum

0
50
k0 (rad/sec)

0.4

0
1

0.2

3
100

50

0
50
k0 (rad/sec)

100

100

50

0
50
k0 (rad/sec)

100

Top: period of x(t) and real Xk vs k0 ; bottom magnitude and phase line spectra

11 / 22

Example: Non-symmetric periodic signals


x(t)

1
2

y(t)

z(t) = x(t + 0.5), even, period: z1 (t) = 2[u(t + 0.5) u(t 0.5)]
2
Z1 (s) = [e 0.5s e 0.5s ]
s
1 2 jk/2
sin(0.5k)
Zk =
[e
e jk/2 ] =
realvalued
2 jk
0.5k
i
X
Xh
jk0 (t0.5)
jk/2
x(t) = z(t 0.5) =
Zk e
=
Zk e
e jkt
{z }
k
k |
Xk

Xk complex since x(t) neither even nor odd

12 / 22

Even and odd components of the period of y (t), 1 t 1


y1e (t)

y1o (t)

2
1

y1e (t) = [u(t + 1) u(t 1)] + [r (t + 1) 2r (t) + r (t 1)]


|
{z
} |
{z
}
rectangular pulse
triangular pulse
y1o (t) = [r (t + 1) r (t 1) 2u(t 1)] [u(t + 1) u(t 1)]
|
{z
} |
{z
}
triangular pulse
rectangular pulse
1
1 (1)k
Yek =
Y1e (s) |s=jk0 =
k 6= 0, Ye0 = 1.5
T0
(k)2
1
(1)k
Yok =
Y1o (s) |s=jk0 = j
k 6= 0, Yo0 = 0
T
k
0

Ye0 + Yo0 = 1.5 + 0 = 1.5
k=0
Yk =
Yek + Yok = (1 (1)k )/(k)2 + j(1)k /(k) k 6= 0

13 / 22

Example: Integration

Z
y (t) =

x(t)dt

x(t)

0.5

y(t) =

x(t)dt

Integral does not exist if the dc is not zero


Period

Period
0.5

0.5

0.4

0.3

y(t)

x(t)

0.5

0.1

1
0

0.5

1.5

0.5

1
t

1.5

0.4

0.3

0.2
0.1
0
0

0
1

20

40

60

2
0

<Yk

0.3

|Yk|

0.4

<Xk

|Xk|

0.2

0.2
0.1

20

40

60

0
0

14 / 22

2
1

20

40

60

0
0

20

40

60

Convergence of Fourier series


For Fourier series of x(t) to converge, it should:
be absolutely integrable,
have a finite number of maxima, minima and discontinuities.
FS equals x(t) at every continuity point and 0.5[x(t + 0+) + x(t + 0)] at every
discontinuity point
Example: Approximate train of pulses with x2(t) = + cos(0t) by
Minimize

Z
1
E2 =
|x(t) x2 (t)|2 dt, w.r.t. ,
T0 T0
Z
dE2
1
=
2[x(t) ]dt = 0
d
T0 T0
Z
dE2
1
=
2[x(t) cos(0 t) cos2 (0 t)]dt = 0
d
T
Z 0 T0
1
=
x(t)dt,
T0 T0
Z
2
=
x(t) cos(0 t)dt
T0 T0

15 / 22

Time and frequency shifting

Periodic signal x(t)


Time-shifting: x(t0 ) remains periodic of the same fundamental period
x(t) {Xk } x(t t0) Xk e jk0t0 = |Xk |e j(Xk k0t0)
only change in phase
Frequency-shifting:
x(t)e j1 t is periodic of fundamental period T0 if 1 = M0 , for an integer M 1,
for 1 = M0 , M 1, the Fourier coefficients Xk are shifted to frequencies
k0 + 1 = (k + M)0
the modulated signal is real-valued by multiplying x(t) by cos(1 t).

16 / 22

Example: Modulating cos(20t) with


a periodic train of square pulses

x1(t) = 0.5[1 + sign(sin(t))] =
with a sinusoid

1 sin(t) 0
0 sin(t) < 0

x2(t) = sin(t).

0.5

0.5
x2(t)

x1(t)

0.5

0.5

2
t

2
t

0.2
|X2k|

|X1k|

0.2
0.1
0
0

50

100

150

0.1
0
0

50

100

150

Modulated square-wave x1 (t) cos(20t) (left) and modulated cosine x2 (t) cos(20t)

17 / 22

Response of LTI systems to periodic signals

Periodic input x(t) of causal and stable LTI system, with impulse response h(t), by
eigenfunction property of LTI systems

Fourier series x(t) = X0 + 2

X
k=1

yss (t) = X0|H(j0)| + 2

X
k=1

|Xk | cos(k0t + Xk )

0 =

2
T0

|Xk ||H(jk0)| cos(k0t + Xk + H(jk0))

where H(jk0) = |H(jk0)|e jH(jk0)H(s)|s=jk0


frequency response of the system at k0

18 / 22

Example: Lowpass filtering using RC circuit with


1
1 + s/100
sin(k/2) j2kt
e
k/2

transfer function H(s) =


input x(t) =

|Ha(j)|

k=,6=0

y(t)
x(t)

0.5

0.4

0.5

0.3
0.2

20

40

60

80

100

0.1
x(t), y(t)

0
0

0
0.1

<H(j )

0.2
0.3

0.5
0.4
0.5

1
0

20

40

60
(rad/sec)

80

100

0.2

0.4

0.6

0.8

1
t (sec)

1.2

1.4

1.6

1.8

Left: magnitude and phase response of the low-pass RC filter at harmonic frequencies. Right: response
due to the train of pulses x(t). Actual signal values are given by the dashed line, and the filtered signal
is indicated by the continuous line

19 / 22

Derivatives and integrals of Periodic Signals

Derivative:

Derivative dx(t)/dt of periodic signal x(t) is periodic of the same


fundamental period. If {Xk } are the coefficients of the Fourier series of x(t), the
Fourier coefficients of dx(t)/dt are
jk0Xk ,

0 fundamental frequency ofx(t)

Integral: Zero-mean, periodic signal y (t) with Fourier coefficients {Yk },

integral z(t) =

y ( )d

Yk
k integer 6= 0
jk0
X
2
1
0 = .
Z0 =
Ym
jm0
T0

Zk =

m6=0

20 / 22

Example: Derivative
period of x(t) : x1 (t) = 2r (t) 4r (t 0.5) + 2r (t 1), 0 t 1, T0 = 1
dx(t)
Gk
g (t) =
k 6= 0
Xk =
dt
jk0
period of g (t) : g1 (t) = dx1 (t)/dt = 2u(t) 4u(t 0.5) + 2u(t 1)
(1)(k+1) (cos(k) 1)
Gk
=
Xk =
jk0
2k 2
X0 = 0.5 from plot of x(t)

k 6= 0

Integral
Z

x(t) =

g ( )d, (G0 = 0)

(1)(k+1) (cos(k) 1)
Gk
=
Xk =
j0 k
2k 2

X
X
Gm
X0 =
= 0.5
j2m
m=,m6=0

k 6= 0
(1)m+1

m=,m6=0

21 / 22

sin(m/2)
(m/2)

2

Basic Properties of Fourier Series


Signals and constants
Linearity
Parsevals power relation
Differentiation

x(t), y (t) periodic


with period T0 , ,
x(t) + y (t)
Z
1
Px =
|x(t)|2 dt
T0 T0
dx(t)
Zdt
t

Integration
Time shifting
Frequency shifting
Symmetry

Multiplication

Xk , Yk
Xk + Yk
X
Px =
|Xk |2
k

jk0 Xk

x(t 0 )dt 0 only if X0 = 0

x(t )

e jM0 t x(t)
x(t) real

z(t) = x(t)y (t)

Xk
k 6= 0
jk0
e j0 Xk
XkM
|Xk | = |Xk | even
function of k
Xk = Xk odd
function of k
X
Zk =
Xm Ykm
m

22 / 22

You might also like