Signals and Systems Using Matlab Chapter 7 - Fourier Analysis in Communications and Filtering

SIGNALS AND SYSTEMS USING MATLAB
Chapter 7 Fourier Analysis in Communications and Filtering

Luis F. Chaparro
Modulation systems
Given lowpass nature of most message signals, it is necessary to shift in frequency the
spectrum of the message to avoid using a very large antenna

This is attained by modulation: changing either the magnitude A(t) or the phase (t)
of a carrier
A(t) cos(2fc + (t)).
giving
Amplitude Modulation (AM): A(t) proportional to message, for constant phase
Frequency Modulation (FM), Phase modulation (PM): (t) changes with the
message
Communication system: cascade of transmitter, channel and receiver none LTI
2 / 21
AM modulation systems
AM Suppressed Carrier
message m(t), carrier cos(c t)

modulated signal s(t) = m(t) cos(c t)
c >> 2f0, f0 maximum frequency in m(t)
Fourier spectrum
1
S() = [M( c ) + M( + c )]
2
M() spectrum of m(t)
2 cos(c t)
cos(c t)
m(t)
Channel
Band-pass
filter
Transmitter
Receiver
AM-SC transmitter, channel and receiver
3 / 21
Low-pass
filter
m(t)
Demodulation: assuming output of bandpass filter is s(t)
R() = S( c ) + S( + c )
1
= M() + [M( 2c ) + M( + 2c )]
2
Output of LPF is M() or m(t)
Demodulation requires exact carrier frequency:
if demodulator uses c + , > 0:

r (t) = s(t) cos((c + )t)
R()
= S( c ) + S( + c + )
1
= [M( + ) + M( )]
2
1
+ [M( 2(c + /2)) + M( + 2(c + /2)] .
2
Output of LPF is distorted message
4 / 21
Commercial AM
s(t) = [K + m(t)] cos(c t) modulated signal

K chosen so that K + m(t) > 0
1
S() = K [( c ) + ( + c )] + [M( c ) + M( + c )] .
2
Example: Commercial AM modulation
1
0.8
0.6
0.4
y(t)
0.2
0
0.2
0.4
0.6
0.8
1
0
0.2
0.4
0.6
0.8
t (sec)
1.6
100
|Y()|
y(t)
1.4
150
0.5
0
0.5
1
0
1.2
0.005
0.01
0.015
0.02
0.025
0.03
0.035
50
0
2.5
0.04
1.5
0.5
1.5
0.5
t (sec)
0.5
1.5
2.5
4
x 10
0.5
1.5
2.5
4
x 10
f (Hz)
2
500
400
|Z()|
z(t)
1
0
1
2
0
300
200
100
0.005
0.01
0.015
0.02
0.025
0.03
0.035
0
2.5
0.04
f (Hz)
t (sec)
Message (top), part of original signal and corresponding AM modulated signal

(bottom left), spectrum of the original signal and of the modulated signal (top and
bottom right).
5 / 21
Single sideband modulation efficient use of spectrum by reducing bandwidth of
modulate signal
cos(c t)
m(t)
s(t)
BPF
H(j)
c + B
Upper side-band AM transmitter. c is the carrier frequency and B the bandwidth of

the message
Quadrature AM efficient use of spectrum by sending two messages on the same band
m1 (t)
m
1 (t)
cos(c t)
cos(c t)
+
s(t)
r(t)
shift
2
m2 (t)
LPF
2 shift
Transmitter
LPF
m
2 (t)
Receiver
QAM transmitter and receiver: s(t) is the transmitted signal and r (t) the received
signal
6 / 21
Frequency division multiplexing sharing the spectrum

cos(1 t)
m1 (t)
cos(1 t)
BPF
cos(2 t)
m2 (t)
LPF
m
1 (t)
cos(2 t)
BPF
Channel
cos(3 t)
m
2 (t)
LPF
cos(3 t)
m3 (t)
BPF
LPF
m
3 (t)
Frequency division multiplexing (FDM) system

Frequency modulation (FM) nonlinear, timevarying modulation system
sFM (t) = cos(c t +
m( )d )
Narrowband FM: angle (t) small so that cos((t)) 1, sin((t)) (t)

d [c t + (t)]
d (t)
= m(t), IF (t) =
= c + m(t)
dt
dt
S() = F [cos(c t + (t))] = F [cos(c t) cos((t)) sin(c t) sin((t))]
1
[( c ) + ( + c )] + [( c ) ( + c )]
2j
() =
M()
j
7 / 21
Example: Narrowband FM
10
|M()|
m(t)
50
0
50
0
1
0.05
0.1
0.15
0
500
0.2
500
0
f
500
0.1
|X()|
x(t)
0.5
0
0.05
0.5
0
0.05
0.1
t (sec)
0.15
0
500
0.2
500
450
400
Frequency
350
300
250
200
150
100
50
0
0.6/7
0.7/7
0.8/7
0.9/7
1/7
1.1/7
1.2/7
1.3/7
0.2
Time
Top left message m(t) R= 80 sin(20t)u(t) and narrow-band FM signal

t
x(t) = cos(2fc t + 0.1 m( )d ); top-right magnitude spectra of m(t) and x(t).
Spectrogram of x(t) displaying evolution of its Fourier transform with respect to time.
8 / 21
Wideband FM
messages: m1(t) = 80 sin(20t)u(t), m2(t) = 2000tu(t)

instantaneous frequencies IFi (t) = 2fci + 50mi (t) i = 1, 2
fc1 = 2500, fc2 = 25Hz
200
150
m(t)
m(t)
50
0
100
50
50
0
0.02
0.04
0.06
0.08
0
0
0.1
0.04
0.02
0.04
0.06
0.08
0.1
0.06
0.08
0.1
0.5
x(t)
0.02
0.5
x(t)
0
0.5
0
0.5
1
0
0.02
0.04
0.06
0.08
0.1
t (sec)
t (sec)
10
|M()|
|M()|
10
5
0
200
0.015
150
100
50
50
100
150
0
1000
200
3
0.01
500
1000
2.5
|X()|
|X()|
500
3
x 10
0.005
2
1.5
0
5000
1
5000
5000
f
5000
5000
5000
4500
4500
4000
4000
3500
3500
3000
Frequency
Frequency
0
f
2500
2000
1500
3000
2500
2000
1500
1000
1000
500
500
0
0.02 0.04 0.06 0.08
0.10 0.12 0.14 0.16 0.18

Time
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
Time
Left sinusoidal message and right ramp message: messages, FM modulated signals,
spectra of messages, spectra of FM signals, and spectrograms of FM signals
9 / 21
Analog filtering
Use of eigenfunction property of LTI systems periodic and aperiodic signals have
Fourier representations consisting of sinusoids of different frequencies, the frequency

components of any signal can be modified by appropriately choosing the frequency
response of the LTI system or filter
LTI system
Aej(0 t+)
A|H(j0 )|ej(0 t++
! H(j0 )
H(s)
Eigenfunction property of continuous LTI systems

Appropriate filter for a certain application is specified using the spectral
characterization of the input and the desired spectral characteristics of the output
Classical approach in filter design is to consider lowpass prototypes, with normalized
frequency and magnitude responses, which may be transformed into other filters with
the desired frequency response
10 / 21
Filtering basics
B(s)
Filter transfer function H(s) =
(LTI system with specific frequency response)
A(s)
filter output Y () = X ()H(j)
Lowpass filter design
1
1 + f (2)
such that for low frequencies f (2) << 1 |H(j)|2 1,
for high frequencies f (2) >> 1 |H(j)|2 0
Choose magnitude squared function |H(j)|2 =
Issues to consider:
selection of the appropriate function f (.),
the factorization needed to get H(s) from the magnitude squared function
frequency transformation to convert LPF into other filters

Magnitude specifications
1 2 |H(j)| 1
0 |H(j)| 1
0 p (passband)
s (stopband)
11 / 21
()
|H(j)|
1
min
1 2
1
p
max
p s
Equivalent magnitude specifications for a lowpass filter

Loss specifications
Loss function () = 10 log10 |H(j)|2 = 20 log10 |H(j)| dBs

0 () max
0 p (passband)
() min
s (stopband)
max = 20 log10(1 2), min = 20 log10(1)
General case: (0) = 1 , 2 in the passband and 3 in the stopband
(0) = 1 dc loss
max = 2 1 maximum attenuation in passband
min = 3 1 minimum attenuation in stopband
12 / 21
Butterworth lowpass filter design
Magnitude response
N th order lowpass Butterworth filter

1
0 2
0
|HN (j )| =
=
1 + [0]2N
hp
hp half-power or 3dB frequency
Factorization
S = s/hp S/j = 0 = /hp

1
HN (S)HN (S) =
1 + (S 2)N
D(S)D(S) = 1 + (S 2)N HN (S) = 1/D(S)
e j(2k1)
N 2N
j(2k1)
2N
Poles: (1) Sk = e
Sk = jN = e j(2k1+N)
e
Sk = e j(2k1+N)/(2N)
k = 1, , 2N
Poles in circle of radius 1

No poles on jaxis
Consecutive poles separated by /N radians
13 / 21
Filter design
() = 10 log10 |HN (/hp )|2 = 10 log10(1 + (/hp )2N )

0 () max
0 p
min () <
s
= p
p
hp
= s
s
hp
(p ) = 10 log10(1 + (p /hp )2N ) max so that

1/2N
0.1max
10
1
(s ) = 10 log10(1 + (s /hp )2N ) min so that
(100.1min 1)1/2N
half-power frequency:
s
p
hp
(100.1max 1)1/2N
(100.1min 1)1/2N
minimum order:
log10[(100.1min 1)/(100.1max 1)]
N
2 log10(s /p )
14 / 21
Chebyshev lowpass filter design
Normalized magnitude squared function
0
,
=
p
1 + 2CN2 (0)
N order of filter, ripple factor, CN (.) Chebyshev polynomials
|HN (j0)|2 =
Chebyshev polynomials
0
CN ( ) =
cos(N cos1(0))
|0| 1
cosh(N cosh1(0)) |0| > 1.
Three term difference equation:

CN+1(0) + CN1(0) = 20CN (0), N 0
initial conditions
C0(0) = cos(0) = 1
C1(0) = cos(cos1(0)) = 0
C0(0) = 1,
C1(0) = 0,
C2(0) = 1 + 202,
C3(0) = 30 + 403,
15 / 21
Filter design

0
2 2
( ) = 10 log10 1 + CN ( )
0
Ripple factor
p
= 100.1max 1,
=
p
0
RW = 1
1
1 + 2
Minimum order
i0.5
h 0.1
10 min 1
cosh1 10
0.1max 1
N
cosh1( ps )
Half-power frequency :
0
(hp ) = 10 log10(1 + 2CN2 (hp )) = 3 dB, then
2
1+
0
2
CN (hp )
= 100.3 2
1
0
= cosh(N cosh1(hp ))

1
1
= p cosh
cosh1
N
CN (hp ) =
hp
16 / 21
Factorization
0 = S/j, S = s/p
1
1
H(S)H(S) =
=
1 + 2CN2 (S/j) D(S)D(S)
Guillemins

1
1
a = sinh1
N
k = sinh(a) cos(k )
real part of pole
0k = cosh(a) sin(k )
imaginary part of pole
where 0 k < /2 are the angles corresponding to the Butterworth filters

(measured with respect to the negative real axis of the S plane)
17 / 21
Example: Lowpass filtering

x(t) = [2 cos(5t) + cos(10t) + 4 sin(20t)]u(t)
Design third-order lowpass Butterworth filter with a half-power frequency hp = 5
rad/sec, to attenuate frequency components of frequency 10 and 20
cos(10 t)2 cos(5 t)+4 sin(20 t)
1
0.7
0.6
x(t)
0.8
0
5
|H(j)|
10
t
15
20
10
t
15
20
0.5
0.4
0.3
0.2
y(t)
0.1
5
6
0.9
0
5
10
20
30
Signal x(t), top right figure; lowpass Butterworth filter with poles and magnitude
response shown on the left. Filtered signal, bottom right, is approximately the
lowfrequency component of x(t)
18 / 21
Example: Butterworth vs Chebyshev lowpass filters with hp = 5 rad/sec

Filtering x(t) = [2 cos(5t) + cos(10t) + 4 sin(20t)]u(t)
Specifications
(0) = 0 dB
max = 0.1 dB, p = 5 rad/sec
min = 15 dB, s = 10 rad/sec
cos(10 t)2 cos(5 t)+4 sin(20 t)
1
0.9
0.8
0.7
Chebyshev
Butterworth
Chebyshev
5
x(t)
0
5
0.6
|H(j)|
Butterworth
10
t
15
20
10
t
15
20
0.5
0.4
5
0.3
0.2
0.1
4
y(t)
0
5
10
20
30
19 / 21
Frequency transformations
s
Lowpass-Lowpass
S=
0
0
Lowpass-Highpass
S=
s
s 2 + 0
Lowpass-Bandpass
S=
s BW
s BW
Lowpass-Bandstop
S= 2
s + 0
S is the normalized and s the final variables
0 is a desired cut-off frequency and BW a desired bandwidth
0.8
0.8
|H1(j)|
|H(j)|
Example: Lowpass prototype filter (Butterworth)

0 = 40 and BW = 10
LP prototype
0.6
0.4
0.2
LPHP
0.6
0.4
0.2
10
20
0.8
0.8
0.6
0.4
0.2
0
0
60
80
30
40
LPBE
|H3(j)|
|H2(j)|
LPBP
40
0.6
0.4
0.2
0
10
20
30
40
10
20
Frequency transformations: (top-left) prototype lowpass filter, (top-right) lowpass to

highpass transformation, (bottom-left) lowpass to bandpass transformation,
(bottom-right) lowpass to band-eliminating transformation
20 / 21
Example: General filter design

Specifications
(0) = 0, max = 0.1, min = 60 dB
p = 10, s = 15 rad/sec
0.6
0.4
6
8
0.2
5
10
15
20
25
10
0
0.8
|H()|
<H()
|H()|
0.8
0
0
0
1
<H()
0.6
0.4
10
15
20
25
10
0
10
15
20
25
80
()
60
()
10
40
50
0
0
10
10
15
20
25
15
20
25
60
40
20
20
20
80
50
10
0
100
100
20
6
8
0.2
10
15
20
25
15
10
Elliptic (left) and Chebyshev2 (right) lowpass filter designs using analogfil function.
Clockwise, magnitude, phase, loss function and poles and zeros are shown for each design.
21 / 21

Signals and Systems Using Matlab Chapter 7 - Fourier Analysis in Communications and Filtering

Uploaded by

Copyright:

Available Formats

Signals and Systems Using Matlab Chapter 7 - Fourier Analysis in Communications and Filtering

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Signals and Systems Using Matlab Chapter 7 - Fourier Analysis in Communications and Filtering

Uploaded by

Copyright:

Available Formats

SIGNALS AND SYSTEMS USING MATLAB

Chapter 7 Fourier Analysis in Communications and Filtering

spectrum of the message to avoid using a very large antenna

Amplitude Modulation (AM): A(t) proportional to message, for constant phase

message m(t), carrier cos(c t)

AM-SC transmitter, channel and receiver

Demodulation: assuming output of bandpass filter is s(t)

if demodulator uses c + , > 0:

s(t) = [K + m(t)] cos(c t) modulated signal

Message (top), part of original signal and corresponding AM modulated signal

Single sideband modulation efficient use of spectrum by reducing bandwidth of

Upper side-band AM transmitter. c is the carrier frequency and B the bandwidth of

Frequency division multiplexing sharing the spectrum

Frequency division multiplexing (FDM) system

sFM (t) = cos(c t +

Narrowband FM: angle (t) small so that cos((t)) 1, sin((t)) (t)

Top left message m(t) R= 80 sin(20t)u(t) and narrow-band FM signal

messages: m1(t) = 80 sin(20t)u(t), m2(t) = 2000tu(t)

0.10 0.12 0.14 0.16 0.18

Fourier representations consisting of sinusoids of different frequencies, the frequency

A|H(j0 )|ej(0 t++

Eigenfunction property of continuous LTI systems

selection of the appropriate function f (.),

frequency transformation to convert LPF into other filters

Equivalent magnitude specifications for a lowpass filter

Loss function () = 10 log10 |H(j)|2 = 20 log10 |H(j)| dBs

Butterworth lowpass filter design

N th order lowpass Butterworth filter

S = s/hp S/j = 0 = /hp

Poles in circle of radius 1

() = 10 log10 |HN (/hp )|2 = 10 log10(1 + (/hp )2N )

(p ) = 10 log10(1 + (p /hp )2N ) max so that

Chebyshev lowpass filter design

Normalized magnitude squared function

Three term difference equation:

real part of pole

imaginary part of pole

where 0 k < /2 are the angles corresponding to the Butterworth filters

Example: Lowpass filtering

Example: Butterworth vs Chebyshev lowpass filters with hp = 5 rad/sec

Example: Lowpass prototype filter (Butterworth)

Frequency transformations: (top-left) prototype lowpass filter, (top-right) lowpass to

Example: General filter design

You might also like