0% found this document useful (0 votes)
50 views71 pagesKeys
Digital modulation techniques enhance data capacity, security, and communication quality compared to analog methods. This document discusses key digital modulation techniques including Amplitude Shift Keying (ASK), Frequency Shift Keying (FSK), and Phase Shift Keying (PSK), detailing their operational principles and mathematical representations. Each technique is characterized by how it modulates the carrier signal based on the input digital data, with examples illustrating their applications and bandwidth requirements.
Uploaded by
tejasripendota1369Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
0% found this document useful (0 votes)
50 views71 pagesKeys
Digital modulation techniques enhance data capacity, security, and communication quality compared to analog methods. This document discusses key digital modulation techniques including Amplitude Shift Keying (ASK), Frequency Shift Keying (FSK), and Phase Shift Keying (PSK), detailing their operational principles and mathematical representations. Each technique is characterized by how it modulates the carrier signal based on the input digital data, with examples illustrating their applications and bandwidth requirements.
Uploaded by
tejasripendota1369Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
/ 71
UNIT2
DIGITAL MODULATION TECHNIQUES
Digital Modulation provides more information capacity, high data security, quicker system
availability with great quality communication. Hence, digital modulation techniques have a greater
demand, for their capacity to convey larger amounts of data than analog ones.
There are many types of digital modulation techniques and we can even use a combination of these
techniques as well. In this chapter, we will be discussing the most prominent digital modulation
techniques.
if the information signal is digital and the amplitude (IV of the carrier is varied proportional to
the information signal, a digitally modulated signal called amplitude shift keying (ASK) is
produced.
If the frequency (f) is varied proportional to the information signal, frequency shift keying (FSK) is
produced, and if the phase of the carrier (0) is varied proportional to the information signal,
phase shift keying (PSK) is produced. If both the amplitude and the phase are varied proportional to
the information signal, quadrature amplitude modulation (QAM) results. ASK, FSK, PSK, and
QAM are all forms of digital modulation:
M)= Vsin (Qe + fr 4 @)
{ tf
ASK ——-FSK PSK
eae
a simplified block diagram for a digital modulation system.
Amplitude Shift Keying
The amplitude of the resultant output depends upon the input data whether it should be a zero level
or a variation of positive and negative, depending upon the carrier frequency.
Amplitude Shift Keying (ASK) is a type of Amplitude Modulation which represents the binary
data in the form of variations in the amplitude of a signal.
Following is the diagram for ASK modulated waveform along with its input. oll Ld
ov
Input binary sequence
w
ov
-v
ASK Modulated output wave
Any modulated signal has a high frequency carrier. The binary signal when ASK is modulated,
gives a zero value for LOW input and gives the carrier output for HIGH input.
‘
Mathematically, amplitude-shift keying is
Vet) = [I + sso] Seo
where vask(t) = amplitude-shift keying wave
vm(t) = digital information (modulating) signal (volts)
A/2 = unmodulated carrier amplitude (volts)
coc= analog carrier radian frequency (radians per second, 2nfet)
In above Equation, the modulating signal [vm({)] is a normalized binary waveform, where + 1 V =
logic 1 and -1 V = logic 0. Therefore, for a logic 1 input, vm(t) = +1 V, Equation 2.12 reduces to
Yel) = [+ 1] Seosta) |
= A cos(w,t)
Mathematically, amplitude-shift keying is (2.12) where vask(t) = amplitude-shif keying wave
v(t) = digital information (modulating) signal (volts) A/2 = unmodulated carrier amplitude (volts)
2 c= analog carrier radian frequency (radians per second, 2nfot) In Equation 2.12, the modulating
signal [vm(t)] is a normalized binary waveform, where + 1 V = logic 1 and -1 V = logic 0.
Therefore, for a logic 1 input, vm(t) = + 1 V, Equation 2.12 reduces to and for a logic 0 input, vm(t)
=-1 V.Equation reduces to
A
Yast) = [1 i
ois
Thus, the modulated wave vask(t),is either A cos(act) or 0. Hence, the carrier is either "on “or
“off,” which is why amplitude-shift keying is sometimes referred to as on-off keying (OOK).
it can be seen that for every change in the input binary data stream, there is one change in the ASK.
waveform, and the time of one bit (tb) equals the time of one analog signaling element (t)
B=fb/l =f baud = fb/1 = fb
Example :
Determine the baud and minimum bandwidth necessary to pass a 10 kbps binary signal using
amplitude shift keying. 10Solution For ASK, N = 1, and the baud and minimum bandwidth are
determined from Equations 2.11 and 2.10, respectively:
B= 10,000 / 1 = 10,000
baud = 10, 000 /1 = 10,000
The use of amplitude-modulated analog carriers to transport digital information is a relatively low-
quality, low-cost type of digital modulation and, therefore, is seldom used except for very low=
speed telemetry circuits
ASK TRANSMITTER:
Mixer
Modulation signal ASK modulated wave
m(t) Sask(t)
Caner wave
ct) The input binary sequence is applied to the product modulator. The product modulator amplitude
modulates the sinusoidal carrier .it passes the carrier when input bit is ‘1’ .it blocks the carrier when
input bit is “0.”
Coherent ASK DETECTOR:
FREQUENCY SHIFT KEYING
The frequency of the output signal will be either high or low, depending upon the input data
applied.
Frequency Shift Keying (FSK) is the digital modulation technique in which the frequency of the
carrier signal varies according to the discrete digital changes, FSK is a scheme of frequency
modulation,
Following is the diagram for FSK modulated waveform along with its input.
w
’ LI
Input binary sequence time
w
—
at time
fi fa
FSK Modulated output wave
The output of a FSK modulated wave is high in frequency for a binary HIGH input and is low in
frequency for a binary LOW input. The binary 1s and 0s are called Mark and Space frequencies
‘SK is a form of constant-amplitude angle modulation similar to standard frequency modulation
(FM) except the modulating signal is a binary signal that varies between two discrete voltage levels
rather than a continuously changing analog waveform.Consequently, FSK is sometimes called
binary FSK (BFSK). The general expression for FSK is
4 vl) = Ve cos(2nf, + v9 AAT
where
vau(t) = binary FSK waveform
Ve = peak analog carrier amplitude (volts)
‘fe = analog carrier center frequency (hertz)
f peak change (shift)in the analog carrier frequency(hertz)
V(t) = binary input (modulating) signal (volts)
From Equation 2.13, it can be seen that the peak shift in the carrier frequency ( 1) is proportional to
the amplitude of the binary input signal (vat}), and the direction of the shift is determined by the
polarity.
The modulating signal is a normalized binary waveform where a logic | = + 1 V and a logic 0 = -1
V. Thus, for a logic | input, vat) = + 1, Equation 2.13 can be rewritten as,
For a logic o input, vat =-1, Bau fil Dearnke coslan(f. + Aft
ve) = V. cos[2n(f, Aft
With binary FSK, the carrier center frequency ({2) is shifted (deviated) up and down in the
frequency domain by the binary input signal as shown in Figure 2-3.
[wes —
FIGURE: ¥SK-inthe frequency domain
5 As the binary input signal changes from a logic 0 to a logic 1 and vice versa, the output frequency
shifts between two frequencies: a mark, or logic 1 frequency (fin), and a space, or logic 0 frequency
(4). The mark and space frequencies are separated from the carrier frequency by the peak frequency
deviation (f) and from each other by 2 f.
Frequency deviation is illustrated in Figure 2-3 and expressed mathematically as,
f= [fa £]/2 (2.14)
where f= frequency deviation (hertz)
[fan ~ fe] = absolute difference between the mark and space frequencies (hertz)
Figure 2-4a shows in the time domain the binary input to an FSK modulator and the corresponding
FSK output.
‘When the binary input (fi) changes from a logic | to a logic 0 and vice versa, the FSK output
frequency shifts from a mark ( fy) to a space (f,) frequency and vice versa.
In Figure 2-4a, the mark frequency is the higher frequency (f: + ) and the space frequency is the
lower frequency ({: =), although this relationship could be just the opposite.
Figure 2-4b shows the truth table for a binary FSK modulator. The truth table shows the input and
output possibilities for a given digital modulation scheme.
by
a
kK fete ie f toa bg binary frequency
weaiees Us = =
Ue tmite ilmite mite mite Saito
Jom mark Kequency; face frequency
@ ©
> s
FIGURE 2-4 FSK inthe time domain: (a) waveform: (b) truth
table FSK Bit Rate, Baud, and Bandwidth
In Figure 2-4a, it can be seen that the time of one bit (ts) is the same as the time the FSK output is a
mark of space frequency (t). Thus, the bit time equals the time of an FSK signaling element, and
the bit rate equals the baud,
The baud for binary FSK can also be determined by substituting N = | in Equation 2.11
baud = fp
fs
The minimum bandwidth for FSK is given as
B= |(f. — fy) — (fm — fi)
=|(f— fin)| + 2 fh,
and since |(f.~ f)| equals 2 f, the minimum bandwidth can be approximated as
B+ f+h) (2.15)
where
B= minimum Nyquist bandwidth (hertz)
f= frequency deviation |(far- {) (hertz)
fr = input bit rate (bps)
Example 2-2
Determine (a) the peak frequency deviation, (b) minimum bandwidth, and (¢) baud for a binary
FSK signal with a mark frequency of 49 kHz, a space frequency of 51 kHz, and an input bit rate of
2 kbps.
Solution
a, The peak frequency deviation is determined from Equation 2.14:
f |149kHz - 51 kHz|/2—1 kHz
b. The minimum bandwidth is determined from Equation 2.15:
B=2(100 2000)
=6 KHz c. For FSK, N = 1, and the baud is determined from Equation 2.11 as
baud = 2000 / 1 = 2000
FSK TRANSMITTER:
Figure 2-6 shows a simplified binary FSK modulator, which is very similar to a conventional FM
modulator and is very often a voltage-controlled oscillator (VCO).The center frequency (f:) is
chosen such that it falls halfway between the mark and space frequencies.
2 oe FSK output
ySJUUL FSK modulator
t > (vco) a
ky = Hav
“af taf
tm te
Logic 0
Logic 1
A logic I input shifts the VCO output to the mark frequency, and a logic 0 input shifts the VCO
output to the space frequency. Consequently, as the binary input signal changes back and forth
between logic 1 and logic 0 conditions, the VCO output shifts or deviates back and forth between
the mark and space frequencies,
az FSK ouput
inary SUL SK modulator
input en) >»
ieeten
say
te
Tiepeo
Loge t
FIGURE 26 FSK modulator
A VCO-FSK modulator can be operated in the sweep mode where the peak frequeney deviation is
simply the product of the binary input voltage and the deviation sensitivity of the VCO. ‘With the sweep mode of modulation, the frequency deviation is expressed mathematically as,
f= Vm(t)k (2-19)
V(t) = peak binary modulating-signal voltage (volts)
k, = deviation sensitivity (hertz per volt).
FSK Receiver
FSK demodulation is quite simple with a circuit such as the one shown in Figure 2-7.
‘Analog mark or
space frequency Rectified signal
t ~——
[apr] SY] Envelope | de
—— detector
FSK input .
SKinput | coltter| by 1, Batata
[pr] Y Envelope |_ de
LJ "| detector | Comparator
FIGURE 2-7 Noncoherent FSK demodulator
The FSK input signal is simultaneously applied to the inputs of both bandpass filters (BPFs)
through a power splitter. The respective filter passes only the mark or only the space frequency on to
its respective envelope detector.The envelope detectors, in turn, indicate the total power in each
passband, and the comparator responds to the largest of the two powers.This type of FSK detection.
is referred to as noncoherent detection.
Figure 2-8 shows the block diagram for a coherent FSK receiver. The incoming FSK signal is
multiplied by a recovered carrier signal that has the exact same frequency and phase as the
transmitter reference.
However, the two transmitted frequencies (the mark and space frequencies) are not generally
continuous; it is not practical to reproduce a local reference that is coherent with both of them,
Consequently, coherent FSK detection is seldom used.
utioter
~@ LE)
Power a
Fk input cater =
coer
Prmer| aguttiptior +4 Data out
“O ~ {cer
Calor
FIGURE 2-8 Coherent FSK demodulator
9 PHASE SHIFT KEYING:
The phase of the output signal gets shifted depending upon the input. These are mainly of two
types, namely BPSK and QPSK, according to the number of phase shifts. The other one is DPSK
which changes the phase according to the previous value.
Amo. ¥
Phase shilt keying (PSK)
Phase Shift Keying (PSK) is the digital modulation technique in which the phase of the carrier
signal is changed by varying the sine and cosine inputs at a particular time. PSK technique is widely
used for wireless LANs, bio-metric, contactless operations, along with RFID and Bluetooth
communications,
PSK is of two types, depending upon the phases the signal gets shifted, They are —
Binary Phase Shift Keying (BPSK)
This is also called as 2-phase PSK (or) Phase Reversal Keying. In this technique, the sine wave
carrier takes two phase reversals such as 0° and 180°,
BPSK is basically a DSB-SC (Double Sideband Suppressed Carrier) modulation scheme, for
message being the digital information.
Following is the image of BPSK Modulated output wave along with its input.
~ WWW We
10 Binary Phase-Shift Keying
The simplest form of PSK is binary phase-shift keying (BPSK), where N = 1 and M =
2.Therefore, with BPSK, two phases (2' = 2) are possible for the carrier.One phase represents a
logic 1, and the other phase represents a logic 0. As the input digital signal changes state (i.e., from
a1 toa 0 or from a 0 to a 1), the phase of the output carrier shifts between two angles that are
separated by 180°.
Hence, other names for BPSK are phase reversal keying (PRK) and biphase modulation. BPSK.
is a form of square-wave modulation of a continuous wave (CW) signal.
Binary
aan Ma [fesnearen Balanced ) | Bandpass| Modulated
n (pier) modulator fiter PSK output
% sin(o,t)
Butfer
7
Ay sinloxt
Reterence|
cartier
oscillator
FIGURE 2-12 BPSK transmitter
BPSK TRANSMITTER:
Figure 2-12 shows a simplified block diagram of a BPSK transmitter. The balanced modulator acts
as a phase reversing switch. Depending on the logic condition of the digital input, the carrier is
transferred to the output either in phase or 180° out of phase with the reference carrier oscillator.
Figure 2-13 shows the schematic diagram of a balanced ring modulator. The balanced modulator
has two inputs: a carrier that is in phase with the reference oscillator and the binary digital data, For
the balanced modulator to operate properly, the digital input voltage must be much greater than the
peak carrier voltage.
This ensures that the digital input controls the on/off state of diodes D1 to D4. If the binary input is
a logic I (positive voltage), diodes D 1 and D2 are forward biased and on, while diodes D3 and D4
a are reverse biased and off (Figure 2-13b). With the polarities shown, the carrier voltage is
developed across transformer T2 in phase with the carrier voltage across T
1. Consequently, the output signal is in phase with the reference oscillator.
If the binary input is a logic 0 (negative voltage), diodes DI and D2 are reverse biased and off,
while diodes D3 and D4 are forward biased and on (Figure 9-13c). As a result, the carrier voltage is
developed across transformer T2 180° out of phase with the carrier voltage across T 1
sinaxt Reference
cree Moduteted PK
Spur ‘ouput
sinangt ~" sinext
Caries
input output
+V (Binary 1)
wo
~ain ext
-V (Binaryo>
2 FIGURE 9-13 (a) Balanced ring modulator; (b) logic 1 input; (c) logic 0 input
a0
cant
nat sin wt
‘rar o
Binary input outa crase] (2. Pi
teoieo | 100
Logie 1 o
@
cont
on
wo
cont
© 0° Retoronce
Logica toge1
~con et
FIGURE 2-14 BPSK modulator: (a) truth table; (b) phasor diagram; (c) constellation
diagram
BANDWIDTH CONSIDERATIONS OF BPSK:
Ina BPSK modulator. the carrier input signal is multiplied by the binary data.
If + 1 V is assigned to a logic 1 and -1 V is assigned to a logic 0, the input carrier (sin wet) is
multiplied by either a+ or - I
The output signal is either + 1 sin tet or-1 sin wet the first represents a signal that is in phase with
the reference oscillator, the latter a signal that is 180° out of phase with the reference
oseillator.Each time the input logic condition changes, the output phase change
Mathematically, the output of a BPSK modulator is proportional to
BPSK output = [sin (21rf,t)] x [sin (211fet)] (2.20)
where
B f= maximum fundamental frequency of binary input (hertz)
f= reference carrier frequency (hertz)
Solving for the trig identity for the product of two sine functions,
0.5cos[2T1(fe — fa)t] - 0.Scos[2T1(fe + fa)t]
Thus, the minimum double-sided Nyquist bandwidth (B) is
fe + fa fe + fa
-f +f
-(fe + fa) or
and because f,- fiy/ 2, where fs = input bit rate,
where B is the minimum double-sided Nyquist bandwidth.
Figure 2-15 shows the output phase-versus-time relationship for a BPSK waveform. Logic | input
produces an analog output signal with a 0° phase angle, and a logic 0 input produces an analog
output signal with a 180° phase angle.
As the binary input shifts between a logic | and a logic 0 condition and vice versa, the phase of the
BPSK waveform shifts between 0° and 180°, respectively.
BPSK signaling element (t5) is equal to the time of one information bit (ts), which indicates that the
bit rate equals the baud.
Pow | w fo ow we | oe
FIGURE 2-15 Output phase-versus-time ret
tionship for a BPSK modulator
4 Example:
For a BPSK modulator with a carrier frequency of 70 MIz and an input bit rate of 10 Mbps,
determine the maximum and minimum upper and lower side frequencies, draw the output spectrum,
de-termine the minimum Nyquist bandwidth, and calculate the baud.
Solution
Substituting into Equation 2-20 yields
output = [sin (21fit)] x [sin (21rf.t)]; fa fi, / 2 = 5 MHz
=[sin 211(SMHz)t)] x [sin 271(70MH7)1)}
=0.5cos[211(70MHz — SMHz)t] — 0.5cos[271(70MHz + SMHz)t]
lower side frequency upper side frequency
Minimum lower side frequency (LSF):
LSF=70MHz - SMHz = 65MHz
Maximum upper side frequeney (USF):
USF = 70 MHz + 5 MHz = 75 MHz
Therefore, the output spectrum for the worst-case binary input conditions is as follows: The
minimum Nyquist bandwidth (B) is
0 MHz
65 MHz 70 MHz. 75 MHZ
(Suppressed)
B= 75 MHz - 65 MHz = 10 MHz
and the baud = fy or 10 megabaud,
BPSK receiver:
Figure 2-16 shows the block diagram of a BPSK receiver.
15 The input signal maybe+ sin wet or - sin Wet .The coherent carrier recovery circuit detects and
regenerates a carrier signal that is both frequency and phase coherent with the original transmit
carrier,
The balanced modulator is a product det
signal and the recovered carrier)
for; the output is the product d the two inputs (the BPSK
The low-pass filter (LPF) operates the recovered binary data from the complex demodulated signal,
zsin(o[ bay
= >| Balanced || | Level ee
ingat Ber modulator i convener cata | Le
output
T T
sin(og) i
{ Clock
recovery
Coherent
recovery
FIGURE 2-16 Block diagram of a BPSK receiver
Mathematically, the demodulation process is as follows.
For a BPSK input signal of + sin wet (logic 1), the output of the balanced modulator is
output = (sin Wet )(sin Wet) = sinwret
or sin?wct = 0.5(1 — cos 2un.t) = 0.5
filtered out
(2.21)
leaving
output = + 0.5 V = logic 1
It can be seen that the output of the balanced modulator contains a positive voltage (+{1/2]V) and a
cosine wave at twice the carrier frequency (2 Wt )
The LPF has a cutoff frequency much lower than 2 wet, and, thus, blocks the second harmonic of
the carrier and passes only the positive constant component, A positive voltage represents a
demodulated logic 1
For a BPSK input signal of -sin wt (logic 0), the output of the balanced modulator is
output = (-sin Wet )(sin wet) = sin’wet
or
sin’wat = -0.5(1 ~ cos 2wat) = 0.5€0,5e05 201.t >
16 Ne filtered out
leaving
output = 0.5 V = logic 0
The output of the balanced modulator contains a negative voltage (-[V/2]V) and a cosine wave at
twice the carrier frequeney (200).
Again, the LPF blocks the second harmonic of the carrier and passes only the negative constant
component. A negative voltage represents a demodulated logic 0.
QUADRATURE PHASE SHIFT KEYING (QPSK):
This is the phase shift keying technique, in which the sine wave carrier takes four phase reversals
such as 0°, 90°, 180°, and 270°,
If this kind of techniques are further extended, PSK can be done by eight or sixteen values also,
depending upon the requirement. The following figure represents the QPSK waveform for two bits
input, which shows the modulated result for different instances of binary inputs.
Carrier / Channel
Modulating value from two bits
° 2 1 3
(00) (10) (01) (14)
Modulated
Result
QPSK is a variation of BPSK, and it is also a DSB-SC (Double Sideband Suppressed Carrier)
modulation scheme, which sends two bits of digital information at a time, called as bigits.
Instead of the conversion of digital bits into a series of digital stream, it converts them into bit-pairs.
This decreases the data bit rate to half, which allows space for the other users.
2.5-2-1 QPSK transmitter.
v A block diagram of a QPSK modulator is shown in Figure 2-17Two bits (a dibit) are clocked into
the bit splitter. After both bits have been serially inputted, they are simultaneously parallel
outputted.
The I bit modulates a carrier that is in phase with the reference oscillator (hence the name "I" for "in
phase” channel), and theQ bit modulate, a carrier that is 90° out of phase.
Fora logic 1 = + 1 V anda logic 0= = 1 V, two phases are possible at the output of the I balanced
modulator (+sin wet and - sin wt), and two phases are possible at the output of the Q balanced
modulator (+cos wet), and (-cos wt).
When the linear summer combines the two quadrature (90° out of phase) signals, there are four
possible resultant phasors given by these expressions: + sin Wet + cos Wet, + sin Wet - cos Wet, -sin
cos Ue.
Wet + cos Wet, and -sin Wet
reams. el [soi
(aaa | ee
erly
Binary input eee
sin aa
a
ae
i -
=e Unear joutent:
summer [| 8PF
,
pm |
ee
way [
tress 7
ial
Q channel fy/2 modulator [cos at
Example:
For the QPSK modulator shown in Figure 2-17, construct the truthtable, phasor diagram, and
constellation diagram.
Solution
For a binary data input of Q = O and I= 0, the two inputs to the Ibalanced modulator are -1 and sin
wt, and the two inputs to the Q balanced modulator are -1 and cos wt.
Consequently, the outputs are
18 Tbalanced modulator =(-1)(sin wet) = -1 sin wet
Q balanced modulator ~(-1)(cos wet) =-1 cos wet and the output of the linear summer is
=] cos Wet - 1 sin Wet = 1.414 sim(wet - 135°)
For the remaining dibit codes (01, 10, and 11), the procedure is the same. The results are shown in
Figure 2-18a.
Scat cam, wll
Yo — ot
Base sats
inary ‘OPsK tents (0* reference)
= | =
a
ti] Sale” ate a!
fi) cats “iy
. ats Saath
‘
FIGURE 2-18 QPSK modulator: (a) truth table; (b) phasor diagram; (c) constellation
diagram
In Figures 2-18b and ¢, it can be scen that with QPSK cach of the four possible output phasors has
exactly the same amplitude, Therefore, the binary information must be encoded entirely in the
phase of the output signal
19 Figure 2-18b, it can be seen that the angular separation between any two adjacent phasors in QPSK
is 90°.Therefore, a QPSK signal can undergo almost a+45° or -45° shift in phase during
‘transmission and still retain the correct encoded information when demodulated at the receiver.
Figure 2-19 shows the output phase-versus-time relationship for a QPSK modulator.
owe =a an an a
input 1 ° 1 °
1
0
apsk
output
phase
FIGURE 2-19 Output phase-versus-time relationship for a PSK modulator
2-5-2-2 Bandwidth considerations of QPSK
With QPSK, because the input data are divided into two channels, the bit rate in either the I or the Q
channel is equal to one-half of the input data rate (fi/2) (one-half of fi/2 = fiv4).
QPSK RECEIVER:
The block diagram of a QPSK
iver is shown in Figure 2-21
The power splitter directs the input QPSK signal to the I and Q product detectors and the carrier
recovery circuit. The cartier recovery circuit reproduces the original transmit cartier oscillator
signal. The recovered carrier must be frequency and phase coherent with the transmit reference
cartier. The QPSK. signal is demodulated in the I and Q product detectors, which generate the
original I and Q data bits. The outputs of the product detectors are fed to the bit combining circuit,
where they are converted from parallel I and Q data channels to a single binary output data stream.
The incoming QPSK signal may be any one of the four possible output phases shown in Figure 2-
18, To illustrate the demodulation process, let the incoming QPSK signal be -sin wet + cos Wet.
Mathematically, the demodulation process is as follows.
20 =H {es} =
| T's"
[es Ss
FIGURE 2-21 QPSK receiver
The receive QPSK signal (-sin wet + cos wt) is one of the inputs to the T product detector. The
other input is the recovered carrier (sin wet). The output of the I product detector is,
I= (-sin wt + c08 wt) (sin wt)
‘QPSK input sina emer
(sin @,1)(sin @.£) + (cos o,1)(sin 01)
= —sin? wt + (cos w,£)(sin ,1)
= in(w, + w,)t + I)t
Airedoud equi
— tit fa t
I= —z + 700s 2s sin oa + zsind
Ly (logic 0)
= -4V (logic
ze (2.23)
Again, the receive QPSK signal (-sin Wet + cos wel) is one of the inputs to the Q product detector.
The other input is the recovered carrier shifted 90° in phase (cos et). The output of the Q product
detector is
21 Q = (-sin wt + C08 w,1) (cos w,1)
———__ __
‘QPSK input signal carrer
= cos? wt — w,t)(cos w,1)
4 + 608 20,2) — tsin(ex, + w,)1 ~ Asingw, ~ w,)¢
=
gg oat equals 0)
Ly Loops — bia 27!
i
1
2
in 203, —
V(logic 1) ea
The demodulated I and Q bits (0 and 1, respectively) correspond to the constellation diagram and
truth table for the QPSK modulator shown in Figure 2-18,
DIFFERENTIAL PHASE SHIFT KEYING (DPSK):
In DPSK (Differential Phase Shift Keying) the phase of the modulated signal is shifted relative to
the previous signal element, No reference signal is considered here. The signal phase follows the
high or low state of the previous element. This DPSK technique doesn’t need a reference oscillator.
The following figure represents the model waveform of DPSK.
om o a 1 Eo 1 Wel o Een: ie
It is seen from the above figure that, if the data bit is LOW ic., 0, then the phase of the signal is not,
reversed, but is continued as it was. If the data is HIGH i.e., 1, then the phase of the signal is
reversed, as with NRZI, invert on 1 (a form of differential encoding).
2 If we observe the above waveform, we can say that the HIGII state represents an M in the
modulating signal and the LOW state represents a W in the modulating signal.
The word binary represents two-bits. M simply represents a digit that corresponds to the number of
conditions, levels, or combinations possible for a given number of binary variables.
This is the type of digital modulation technique used for data transmission in which instead of one-
bit, two or more bits are transmitted at a time, As a single signal is used for multiple bit
transmission, the channel bandwidth is reduced.
DBPSK TRANSMITTER.
Figure 2-37a shows a simplified block diagram of a differential binary phase-shift keying
(DBPSK) transmitter. An incoming information bit is XNORed with the preceding bit prior to
entering the BPSK modulator (balanced modulator).
For the first data bit, there is no preceding bit with which to compare it, Therefore, an initial
reference bit is assumed. Figure 2-37b shows the relationship between the input data, the XNOR
output data, and the phase at the output of the balanced modulator. If the initial reference bit is
assumed a logic 1, the output from the XNOR circuit is simply the complement of that shown,
In Figure 2-37b, the first data bit is XNORed with the reference bit. If they are the same, the XNOR
output is a logic 1; if they are different, the XNOR output is a logic 0. The balanced modulator
operates the same as a conventional BPSK modulator; a logic I produces +sin wt at the output, and
A logic 0 produces -sin wet at the output.
2B a
‘Le
FIGURE 2-37 DBPSK modulator (a) block diagram (b) timing diagram
BPSK RECEIVER:
Figure 9-38 shows the block diagram and timing sequence for a DBPSK receiver. The received
signal is delayed by one bit time, then compared with the next signaling element in the balanced
modulator. If they are the same. J logic 1(+ voltage) is generated. If they are different, a logic 0 (-
voltage) is generated, [f the reference phase is incorrectly assumed, only the first demodulated bit is
in error. Differential encoding can be implemented with higher-than-binary digital modulation
schemes, although the differential algorithms are much more complicated than for DBPS K.
The primary advantage of DBPSK is the simplicity with which it can be implemented. With
DBPSK, no carrier recovery circuit is needed. A disadvantage of DBPSK is, that it requires
between 1 dB and 3 dB more signal-to-noise ratio to achieve the same bit error rate as that of
absolute PSK. inpxt Balanced covered
modulator
Tbe sina (sin od
bal (nag sm et
(snags (sin t
eo a a a
input Onase ateonce phase)
FIGURE 2-38 DBPSK demodulator: (a) block diagram; (b) timing sequence
COHERENT RECEPTION OF FSK:
The coherent demodulator for the coherent FSK signal falls in the general form of coherent
demodulators described in Appendix B. The demodulator can be implemented with two correlators
as shown in Figure 3.5, where the two reference signals are cos(27r ft) and cos(27r fit). They must
be synchronized with the received signal. The receiver is optimum in the sense that it minimizes the
error probability for equally likely binary signals. Even though the receiver is rigorously derived in
Appendix B, some heuristic explanation here may help understand its operation. When s 1 (0) is
transmitted, the upper correlator yields a signal 1 with a positive signal component and a noise
component. However, the lower correlator output 12, due to the signals’ orthogonality, has only a
noise component. Thus the output of the summer is most likely above zero, and the threshold
detector will most likely produce a 1. When s2(t) is transmitted, opposite things happen to the two
correlators and the threshold detector will most likely produce a 0. However, due to the noise nature
that its values range from -00 to m, occasionally the noise amplitude might overpower the signal
amplitude, and then detection errors will happen. An alternative to Figure 3.5 is to use just one
correlator with the reference signal cos (27r ft) - cos(2s £21) (Figure 3.6). The correlator in Figure
3.6 can be replaced by a matched filter that matches cos(27r fit) - cos(27r £2t) (Figure 3.7). All
25 implementations are equivalent in terms of error performance (see Appendix B). Assuming an
AWGN channel, the received signal is
r(t)=s,(t)+n(t), i= 1.2
where n(t) is the additive white Gaussian noise with zero mean and a two-sided power spectral
density A'/2, From (B.33) the bit error probability for any equally likely binary signals is
E, + E2 - 2p,
0 ( [Estee
where No/2 is the two-sided power spectral density of the additive white Gaussian noise. For
Sunde's FSK signals E1~ Ez ~ Eb, pl2 ~ 0 (orthogonal). thus the error probability is
where Eb ~ A2T/2 is the average bit energy of the FSK signal. The above Pb is plotted in Figure 3.8
where Pb of noncoherently demodulated FSK, whose expression will be given shortly, is also
plotted for comparison.
n
Threshold
cos(2nfit) Detector
ro) ri
eal lot} +
Jo
cos(2nfst) ™
(kIT ty
Cs
-
26 ter... Noncoherent BESK
nf
Coherent BFSK
Ey N,(dB)
Figure: Pb of coherently and non-coherently demodulated FSK signal.
|ON AND ERROR PERFORMANC
NONCOHERENT DEMODU!
Coherently FSK signals can be noncoherently demodulated to avoid the carrier recovery.
Noncoherently generated FSK can only be noncoherently demodulated. We refer to both cases as
noncoherent FSK. In both cases the demodulation problem becomes a problem of detecting signals
with unknown phases. In Appendix B we have shown that the optimum receiver is a quadrature
receiver. It can be implemented using correlators or equivalently, matched filters. Here we assume
that the binary noncoherent FSK signals are equally likely and with equal energies. Under these
assumptions, the demodulator using correlators is shown in Figure 3.9. Again, like in the coherent
case, the optimality of the receiver has been rigorously proved (Appendix B). However, we can
easily understand its operation by some heuristic argument as follows. The received signal
(ignoring noise for the moment) with an unknown phase can be written as
si(t.8) = Acos(2rfit+0), i=1,2
= Acos cos 2r fit ~ Asin@ sin 2x fit
7 The signal consists of an in phase component A cos 8 cos 27r f t and a quadrature component A sin
8 sin 2x ft sin 0. Thus the signal is partially correlated with cos 2s fit and partiah'y correlated with
sin 27r fit. Therefore we use two correlators to collect the signal energy in these two parts. The
outputs of the in phase and quadrature correlators will be cos 19 and sin 8, respectively. Depending.
on the value of the unknown phase 8, these two outputs could be anything in (- 5, y). Fortunately
the squared sum of these two signals is not dependent on the unknown phase. That is
AT
F cose)? + (Ae
ere
sino)? = ¢
This quantity is actually the mean value of the statistics I? when signal si (t) is transmitted and noise
is taken into consideration, When si (t) is not transmitted the mean value of 1: is 0. The comparator
decides which signal is sent by checking these I?, The matched filter equivalence to Figure 3.9 is
shown in Figure 3.10 which has the same error performance. For implementation simplicity we can
replace the matched filters by bandpass filters centered at f and fi, respectively (Figure 3.1 1)
However, if the bandpass filters are not matched to the FSK signals, degradation to
Sample t
tar
Bandpass Envelope h
siteratfy [—?} Detector
Wa>b
O {choose 1
« Sample at | Comparatc Mb>t,
ager choose 0
Bancpass Envelope |__§. \\_F
finer tf Detetor bs
various extents will result, The bit error probability can be derived using the correlator demodulator
(Appendix B). Here we further assume that the FSK signals are orthogonal, then from Appendix B
the error probability is
n/2No
28 PART-2
DATATRANSMISSION
BASE BAND SIGNAL RECEIVER:
Consider that a binary encoded signal consists of a time sequence of voltage levels +V or -V.
if there is a guard interval between the bits, the signal forms a sequence of positive and
negative pulses. in either case there is no particular interest in preserving the waveform of the
signal after reception .we are interested only in knowing within each bit interval whether the
transmitted voltage was +V or —V. With noise present, the receives signal and noise together
will yield sample values generally different from +V. In this case, what deduction shall we
make from the sample value concerning the transmitted bit?
Suppose that the noise is gaussian and therefore the noise voltage has a
probability density which is entirely symmetrical with respect to zero volts. Then
the probability that the noise has increased the sample value is the same as the
probability that the noise has decreased the sample value. It then seems entirely
reasonable that we can do no better than to assume that if the sample value is
positive the transmitted level was +, and if the sample value is negative the
transmitted level was —V. It is, of course, possible that at the sampling time the
noise voltage may be of magnitude larger than V and of a polarity opposite to
the polarity assigned to the transmitted bit. In this case an error will be made as
indicated in Fig. 111-1. Here the transmitted bit is represented by the voltage
+ which is sustained over an interval T from t, to t;. Noise has been superim-
posed on the level + so that the voltage v represents the reccived signal and
noise. If now the sampling should happen to take place at a time f= 1, + Ar, an
error will have been made.
We can reduce the probability of error by processing the received signal plus
noise in such a manner that we are then able to find a sample time where the
sample voltage due to the signal is emphasized relative to the sample voltage due
to the noise. Such a processer (receiver) is shown in Fig. 11.1-2. The signal input
during a bit interval is indicated. As a matter of convenience we have set t= at
the beginning of the interval. The waveform of the signal s(t) before 1 = @ and
after t= T has not been indicated since, as will appear, the operation of the
receiver during each bit interval is independent of the waveform during past
and future bit intervals.
The signal s(t) with added white gaussian noise n(r) of power spectral density
n/2 is presented to an integrator. At time f= 0+ we require that capacitor C be
uncharged. Such a discharged condition may be ensured by a brief closing of
switch SM, at time r= 0 — , thus relieving © of any charge it may have acquired
during the previous interval. The sample is taken at the output of the integrator
by closing this sampling switch SW,. This sample is taken at the end of the bit
interval, at ¢— T. The signal processing indicated in Fig. 11.1-2 is described by
the phrase integrate and dump, the term dump referring to the abrupt discharge of
the capacitor after each sampling.
23 peta
ae 4 ‘
° nse Scag ia
od ive oe
to - aa
Peak Signal to RMS Noise Output Voltage Ratio
‘The integrator sields an output which is
IRC. Using ¢ = RC, we have
ran=2 [ros d=t Cara et (mod a
‘The sample voltage due to the signal is
tegral of its input multiplied by
wnat [Tea IE ann
The sample voltage due to the noise is
mcr= 2 fe a ous
‘This noise-sampling voltage 7,(7) is @ gaussian random variable in contrast with
a(t), which is 4 gaussian random process.
The variance of (7) was found in See. 7.9 [see Eq. (7.9-17)]) to be
res
nih =
any
and. as noted in Sec. 7.3, n,(T) has a gaussian probability density.
The output of the integrator, before the sampling switch, is rr) = s,(t)
+2n,(t). As shown in Fig, 11.1-3a, the signal output s,(f) is a ramp, in cach bit
interval, of duration T: At the end of the interval the ramp attains the voltage
s,{T) which is +UT/r or —WT/t, depending on whether the atorad, At
the end of each interval the switch SW, in Fig. 11.1-2 closes momentarily to dis-
charge the capacitor so that s,(f) drops to zero. The noise n(f). shown in
Fig. 11.1-3b, also starts cach interval with n,(0) = 0 and has the random value
n{T) at the end of each interval. The sampling switch SW closes briefly just
belore the closing of SH, and hence reads the voltage
ef TY = 57) nT) a
30 oa)
“ oe ee 5
Figure 11.1-3 (0) The signal awiput and {h) the noise output of the integrate of Fig. 11,1
We would naturally like the output signal voltage to be as large as possible
in comparison with the noise voltage, Hence a figure of merit of interest is the
signal-to-noise ratio
ATVI _ 2
tst7H eer (ttt -6)
This result is calculated from Eqs. (11.1-2) and (11.1-4). Note that the signal-to-
noise ratio increases with increasing bit duration 7 and that it depends on ¥"T
which is the normalized energy of the bit signal. Therefore. a bit represented by a
narrow, high amplitude signal and one by a wide, low amplitude signal are
equally effective, provided VT is kept constant,
Iis instructive to note that the integrator filters the signal and the noice such
that the signal voltage increases linearly with time, while the standard deviation
{ems value) of the noise increases more slowly, as \/7 Thus, the integrator
enhances the signal relative to the noise, and this enhancement increases with
time as shown in Eq. (11.1-6)
PROBABILITY OF ERROR
Since the function of a receiver of a data transmission is to ditinguish the bit 1 from the bit 0
in the presence of noise, a most important charcteristic is the probability that an error will be
made in such a determination.we now calculate this error probabilty Pc for the integrate and
dump receiver of fig 11.1-2
31 We have seen that the probability density of the noise sample »,(T) is gamse.
jan and hence appears as im Fig. 112-1. The density is therefore piven hy:
121
where a3. the variance.
WT) given by Faq. (111-8). Suppnse, then, that
during some bit interval the inputsignal voltage is held at, say, —V. Then, at the
sample time. the signal sample voltage is s{T) — —'T/z, while the noise sample
is nAT). If nT) is positive and larger in magnitude than VT /r, the total sample
voltage #,(T) — s9(T) + nf T) will be positive. Such a positive sample voltage will
Wn error, since ag noted earlier, we have instructed the receiver to inter
result in
Pret such a positive sample voltage to mean that the signal voltage was +1
during the bit interval. The probability of such a misinterpretation, that is, the
probability that nT) > T/z, is given by the area ef the shaded region in
Fig. 112-1. The probability of error is, using Eq. (11.2-1.
p= [S remcrn anny = [7 EE ery iad
Defining x = ng TH \/Zo,, and using Eq. (11.1-4), Bq. (1 1.2-2) may be rewritten as
12
2 Ve Jen
weve (vf) fete (22)
erte (=) P12
2 7
in which E, — V27 ie the signal oneray of a hit.
Tethe eignal voltage wore held loalend at 4+ during some
is clear from the symmetry of the situation that the probability of error woul
Sanin he given by Pin Eq. (11.2-3). Honec Bq. (112-3) gives. M_ quite generality
P, em dx
interval, then it
Haai7
Flgere 124 The gronsan probity enay of theo sample nf
32 The probability of error pe, as given in eq.(11.2-3),is plotted in fig.11.2-2.note that pe decreases
rapidly as Es/n increases. The maximum value of pe is ¥i.thus ,even if the signal is entirely lost in
the noise so that any determination of the receiver is a sheer guess, the receiver cannot bi wrong
more than half the time on the average.
THE OPTIMUM FILTER:
In the receiver system of Fig 11.1-2, the signal was passed through a filter(integrator),so that at the
sampling time the signal voltage might be emphasized in comparison with the noise voltage. We are
naturally led to risk whether the integrator is the optimum filter for the purpose of minimizing the
probability of error. We shall find that the received signal contemplated in system of fig 11.1-2 the
integrator is indeed the optimum filter. However, before returning specifically to the integrator
receiver.
We assume that the received signal is a binary waveform. One binary digit is represented by
a signal waveforms; (1) which persists for time T, while the4 other bit is represented by the
waveform S2(t) which also lasts for an interval T. For example, in the transmission at baseband, as
shown in fig I1.1-2 Si(t)=+V; for other modulation systems, different waveforms are transmitted.
for example for PSK signaling , $:(t)=Acoswot and S2(t)=-Acoswot;while for FSK,
Si()=Acos(worant
As shown in Fig. 11.3- the input, which is s,(¢) or s,(f), is corrupted by the
addition of noise n(1). The noise is gaussian and has a spectral density G{f). [In
most cases of interest the ne white, so that G(s) = 7/2. However, we shall
assume the more general poss since it introduces no complication to do
so] The signal and noise are filtered and then sampled at the end of each bit
interval. The output sample is either (7) = s..(T)+ mT) or rT) = s.3(T)
+n{7T). We assume that immediately after each sample, every energy-storing
element in the filter has been discharged.
We have already considered in Sec. 2.22, the matter of signal determination
im the presence of noise. Thus, we note that in the absence of noise the ouiput
sample would be nF) — #4(T) or s,:(T). When noise is present we have shown
that to minimize the probability of error one should assume that s,(¢) has been
transmitted if u,(7) is closer to s,,(T) than to s,.(7). Similarly, we assume s,(1)
has been transmitted if 0,(7) is claser to s,,(T). The decision boundary is there-
fore midway between s,,(T) and s,,(T). For example, in the baseband system of
Fig. 111-2, where s,.(T} = VT/e and s,.(T) = —V Tix, the decision boundary is
v fT) = 0. in general, we shall take the decision boundary to be
in)
ay
The probability of error for this general case may be deduced as an extension
of the considerations used in the baseband case. Suppose that 8,,(T) > 3,37) and
that 5,(1) was transmitted. If, at the sampling time, the noise n,(7) is positive and
larger in magnitude than the voltage difference H[s,4(T) + aya(TH — sot T) an
error will have been made. That is, an error [we decide that s\(1) is transmitted
rather than s,(7)] will result if
sal) = 5.07)
mene (32,
33 Hence probability of error is
e (3)
= Jaingnan
te LBL gins firey
z te
Far 134 Arar iro nating
If we make the substitution x = n,(T)//20,, Eq. (113-3) becomes
2 J sp
~12 ot ae (113-46)
2 Se Ni -nsararten
A(T) = 557
AT) — sof) (134
a 2
Note that for the case s,,(7) = VI/1 and s,.(7) = —WVT/t, and, using Ea. (
4), Eq. (113-46) reduces to Eq, (11.2-3) as expected.
The complementary error function is a monotonically decreasing function of
its argument. (Sec Fig. 11.2-2.) Hence, as is to be anticipated, P, decreases as the
difference 5,3(7) — 547) becomes larger and as the rms noise voltage a, becomes
smailer, The optimum filter, then, is the filter which maximizes the ratio
y — Sel) = Sel)
We now calculate the transfer function MS) of this optimum filter, As a matter of
mathematical convenience we chall actually maximize 7? rather than y.
Calculation of the Optimum-Filter Transfer Function H(/)
The fundamental requirement we make of a binary encoded data receiver is that
it distinguishes the voltages 5,(¢) + mr) and s,(r)-+ a(t), We have seen that the
ability of the receiver to do so depends on how large a particular receiver can
make y. It is important to note that 7 is proportional not to s,(1) nor to s(t). but
Father to the difference between them. For example, in the baseband system we
represented the signals by voltage levels + and —V. But clearly, if our only
imterest was in distinguishing levels, we would do just ac well to use +2 volts and
0 volt, or +8 volts and +6 volts, etc. (The +¥ and —V levels, however. have
the advantage of requiring the least average power to be transmitted.) Hence,
while s;(1} or s(t) is the received signal, the signal which is to be compared with
the noise, ne. the signal which ts relevant in all our error-probability calculations,
is the difference signal
PA) = 5,60) — 5360) 113-6)
‘Thus, for the purpose of calculating the minimum error probability, we shall
assume that the input signal to the optimum filter is p{e). The corresponding,
output signal of the Gilter is then
Pat) = 84) — Seal) (113-7)
34 We shall let P(f) and P,(f) be the Fourier transforms, respectively, of ptr) and
re)
If HCP) is the transfer function of the filter,
PLS) = HSPs) i
and p(T) = C PLS eT f= e HOPPE df (1.39)
3-8)
The input noise to the optimum filter is mfr). The output noise is ,{¢ which
has a power spectral density G,(f) and is rclated to the power spectral density of
the input noise GLA) by
GA) = | HOO PG.AD, (113-10)
5), we find that the normalized output noise
Using Parseval's theorem (Eq. |
power, ie., the noise variance a2, is
a= [ GN) are f° LAN P GN dF (31D
From Eqs. (11.3-9) and (11.3-11) we now find that
G 2eTF ap]?
Lt MUP TF af (113-12)
FPO GAD) af
Equation (11,3-12) is unaltered by the inclusien or deletion of the absolute value
sign in the numerator since the quantity within the magnitude sign p{T) is a
positive real number. The sign has been included, however, in order to allow
further development of the equation through the use of the Schwarz inequality.
The Schwarz inequality states that given arbitrary complex functions X()
and Y(f) of a common variable f, then
£ xcrvirrar| =f" cron ar {7 IYonP ar ans
The equal sign applies when
XU = REVS) a
where K is an arbitrary constant and ¥*%(f) is the complex conjugate of ¥{/’).
‘We now apply the Schwarz inequality to Eq. (11-3-12) by making the ident
sation
X(N = MGA HUY (11.3-15)
1
and YU) = == Pifiel™s 11.3-16)
VGA §
‘Using Eqs, (11.3-15) and (11.3-16) and using the Schwarz inequality, Eq. (113-13),
we may rewrite Eq. (11.3-12) as
PHT) _ Uf22 XW AAP C Lyon? af t
0 |XLOP af
n
35 or. using Eq. (11.3-16),
AD < [" ronP ara £
Lipo)?
Ma a3
Guy canna)
The ratio pi(T)/o3 will attain its maximum value when the equal sign in
Eq. (11.3-18) may be employed as is the case when X() = KY¥*(1. We then find
from Eqs. (11.3-15) and (113-16) that the optimum filter which yields such a
maximum ratio p(T a2 has a transfer function
- PP) . ert
AUP) K GAP) 11,3-19),
‘Correspondingly, the maximum ratio is, from Eq. (11.3-18),
- PUP
~ on & ae
n succeeding sections we shall have occasion te apply Eqs. (11.13-19) and
(11.13-20) to a number of cases of interest.
1L4 WHITE NOISE:
HE MATCHED FILTER
An optimum filter which yields a maximum ratio pi(T)e2 is called a marched
filter when the input noise is white. In this case Gf) = 1/2, and Eq. (113-19)
hecomes
ny = «ED ones Ray
‘The impulsive response of this filter, ie, the response of the filter to a unit
strength impulse applied at ¢ — 0, is
MO = FTE = FE ("pete erreten ag (114-24)
= 2” prinetenen ay (11.4-26)
A physically realizable filter will have an impulse response which is real,
complex. Therefore f(r) = h*(9. Replacing the right-hand member of Ea. (
by its complex conjugate, an operation which leaves the equation unaltered, we
have
2K f°
mi 7 f PU fle =O af (114-34)
2K
kr (
7 wt ) (114-30)
Finally, since p() = 5,0) ~ (0 Lee Eq, (113-6), we have
m2 for 9s) (144
36 The significance of these results for the matched filter may be more readily
appreciated by applying them to a specific example. Consider then,
Fig. 114-1
that 5)(¢) is @ triangular waveform of duration T, while s2(t) as
-1h, is of identical form except of reversed polarity. Then pli) is
as shown in Fig, 11.4-1c and pl—#) appears in Fig. 11.4-1d. The waveform /{—1)
is the waveform p{¢) rotated around the axis — 0. Finally, the waveform p(T — #)
called for as the impulse response of the filter in Eq. (11.4-3b) is this rotated
waveform p—1) translated in the positive ¢ direction by amount T. This last
translation ensures that h(/) = for 1 -< Qas is required for a causal filter.
In general, the impulsive response of the matched filter consists of p(t) rotated about t=0 and
then delayed long enough(ice, a time T) to make the filter realizable. We may note in passing, that
any additional delay that a filter might introduce would in no way interfere with the performance of
the filter ,for both signal and noise would be delayed by the same amount, and at the sampling time
(which would need similarity to be delayed)the ratio of signal to noise would remain unaltered,
@
Ce) Flgure 41 The sana (2 8) 200 ad
0) — 50 Ud) ose aout The
(0 (9) The waveform nd) ranalated
the righ by amcunt 7.
11S PROBABILITY OF ERROR OF THE MATCHED FILTER
The probability of error which results when employing a matched filter, may be
found by evaluating the maximum signal-to-noise ratio (p'(TVa21,,
. given by
Fg. (113-20), With GCP) = n/2. Eq. (11.3-20) becomes
ape
=| inane a (say
lan Jaw
37 From parseval's theorem we have
{ irnear= [" Pad
| rode (115.2)
In the last integral in Eq, (11,5-2), the limits take account of the fact that p(t) per-
sists for only a time T. With pl) = sx — sale), and using Eq. (tt-s-2) we may
write Ea. (115-1) as
[2] 2! bs.) syoP at 115-30)
2 (T. 7 fr
aL, oes 0) a ~2 | s(ts,(f) dt (115-36)
2
aia + Ba 26) (sx
Here F,, and F,; are the energies, respectively, in s;(e) and s,(c), while E,,2 is the
energy due to the correlation between #,() and »,(4).
Suppose that we have selected «,(r), and let s(t) have an energy E,,. Then it
can be shown that if s(t) is to have the same energy, the optimum choice of s(t)
= ~3() assy
The choice is optimum in that it yields a maximum output signal p3(T) for a
given signal energy. Letting s(t) = —s,(t), we find
and Eq, (11.5-3e) becomes
(115-5)
Rewriting Eq (11.3-48) using p(T) = 5,,(T) ~sT), we have
P. = Gove [20] = Fone [HO
2roJ 2
" (115-6)
we find that the minimum errr probability
‘Combining Eq. (11.5-6) with (11.5
(POmin COFFesponding to a maximum value of p3(T)/e?
(r= Sof [ 2807] 3” ii
o«(2) use
We note that Eq. (11.5-8) establishes more generally the idea that the error
probability depends only on the signal energy and not on the signal waveshape.
Previously we had established this point only for signals which had constant
voltage levels.
38 We note also that fq. (115-8) BIYeS (Ime OF the case of the matched titer
and when 5)(f) — —s2(t). In See. 11.2 we Considered the case when s(t) — +
and s.(t)}—= —V and the filter employed was an integrator. There we found
LEg. (11.2-3)) thar the result for P, was identical with (PJmin given in Eq. (115-8).
This agreement Ieads us to suspect that for an input signal Where s,(c) = + V and
s,() — — V, the integrator is the matched filter. Such is indeed the case, For when
we have
ostsT (115-92)
OstsT (1s.9m)
the impulse response of the matched filter is, from Eq, (11.448),
w= tro s{T—0) (11.5410)
The quantity (7 — 2) —s(T —1 is a pulse of amplitude 2V extending from
£= Oto r= Tand may be rewritten, with u(r) the unit step.
wa
arian — 4t—T)) ans
‘The constant factor of proportionality 4KV/y in the expression for h(2) (that is,
the gain of the filter) has no effect on the probability of error since the gain affects
signal and noise alike. We may therefore sclect the cocMficient K in Eq. (11.5-11)
so that 4K Vim ‘Then the inverse transform of h(t), that is, the transfer fune—
tion of the filter, becomes, with s the Laplace transform variabl
HG)
as
12)
The first term in Eq. (1.5.12) represents an integration beginning at ¢
while the second term represents an integration with reversed polarity begin
att =. The overall response of the matched filter is an integration from +
to f= T and a zero response thereafter. In a physical system, as already
described. we achieve the effect of a zero response after ©= T by sampling at
¢ = T. so that so far as the determination of one bit is concerned we ignore the
response after r= T.
COHERENT RECEPTION :CORRELATION:
We discuss now an alternative type of receiving system which, as we shall see, is
entical in performance with the matched filter receiver. Again, as shown in
Fig. 116-1, the input is a binary data waveform #,(¢) of az{¢) corrupted by nuise
n(). The bit length is T, The received signal plus noise fe) is multiplied by a
locally generated waveform s,(1) — s(t). The output of the multiplier is passed
through an integrator whose output is sampled at £— T. As before, immediaiely
after each sampling, at the beginning of each new bit interval, all energy-storing
lements in the integrator are discharged. This type of receiver is called a correla-
"or, since we are correlating the received signal and noise with the wavelorm 50)
‘The output signal and noise of the correlator shown in Fig. 11.6-1 are
sdos(t) ~ 340] de
(11.6-1)
39 wine! [etsy san
(11.6-2)
Where si(t) is either si(t) or s2(1),and wthere mis the constant of the integrator(i.e.,the integrator
output is I/x times the integral of its input).we now compare these outputs with the matched filter
outputs.
tnt | “
ving A frm HE
coer
Fig:11.6-1Coherent system of signal reception
IF h(t) is the impulsive response of the matched filter ,then the output of the matched filter vo(t) can
be found using the convolution integral. we have
tuo = [°scane ayaa [vane ~ ayaa
is (11.6-3)
The limits on the integral have been charged to 0 and T since we are interested in the filter response
to a bit which extends only over that interval. Using Eq.(11.4-4) which gives h(t) for the matched
filter, we have
Wt) = = Es(T =) —sT - 01
Mt) =P ET = 9 SAT — (116-4)
so that Ae = [s(T 04) —s(T 14 aD (11.65)
sub 11.6-5 in 11.6-3
= 2 nAedT — 14a) — se — 1-4 ayaa (11.66)
Since vf) =
) + nf) and 0,0) = sit) + m0, setting t= T yields
2K
it) [sats sa a
te (11.6-7)
where (2s eq to (2) oF (2) Similarly we find that
nit) = [ene = sa) da
" (11.6-8)
Thus so(T) and no(T), as calculated from egs.(11.6-1) and (11.6-2) for the correlation receiver, and
as calculated from eqs.(11.6-7) and (1.68) for the matched filter receiver, are identical hence the
performances of the two systems are identical. The matched filter and the correlator are not simply
40 two distinct, independent techniques which happens to yield the same result. In fact they are two
techniques of synthesizing the optimum filter h(t)
a Sty iter Probability of ASK
Amplitude Shift Keying (ASK), some number of carrier cycles are transmitted to
send '1' and no signal is transmitted for binary ‘0’. Thus,
Binary ‘I’ => x, (f) = 2; cos(2nfot) and
Binary ‘0 => x(t) = 0 (ie. no signal) =» (13.1)
Here F, is the normalized power of the signal in 12 load. ie. power P, =
Hence A =,2P,. Therefore in above equation for x; (!) amplitude ‘A’ is replaced iy
VR.
We know that the ~— of error of the optimum filter is given as,
Fan) w= (5132)
226
7 xor
“Tart
‘The above equations can be applied to matched filter when we consider white
Gaussian noise. The power spectral density of white Gaussian noise is given as,
sun = No
Putting this value of S,, (f) in above equations we get,
[ouD=x0 7° a fxg KOR ay
~ -
- (5.13.3)
ss ST 2
wp {xn af
Parseval’s power theorem states that,
fxoey = frou
Hence equation 5.13.3 becomes,
: 2 =
[aOaeO? - 2 Fama
Ll Janae
42 We know that x(f) is present from 0 to T. Hence limits in above equation can be
changed as follows
- 52 Tr
[2 (T)— 02 ) = et 2d - 6.13.4)
e Jomx Noy
We know that x() =x; (22 (0 For ASK x2) is. zero, hence x(é)= 2x; (1). Hence
above equation becomes,
r
= Pf doa
9%
Putting equation of x; () from equation 5.13.1 in above equation we get,
Pt 2
= wo fb? cos 2nfot) |” at
BPE cog?
. Ne cos? (2Rfy f) dt
We know that cos? @=2*°°59 Here applying this formula to cos? (2rfy!) we get,
2 Jmx No a
[sO=s20f 2 Af Lroostalt
Lact sntagal
o 0
J
7 [isin fot)”
Uo *|axfo
sin dn fT
‘a | an fo
We know that T is the bit period and in this one bit period, the carrier has integer
number of cycles. Thus the product fy T is an integer. This is illustrated in Fig. 5.13.1
43 Fig. 5.13.1 In one bit period T, the
carrier completes its two cycles, The
carrier has frequency fe. From figure we can
write,
1d
ll | iat
ty 2
one $ ie T=?
we
pated A Tod Gemgenect uted
As shown in above figure, the carrier completes two cycles in one bit duration.
Hence
foT = 2
‘Therefore, in general if carrier completes ‘k’ number of cycles, then,
foT =k (Here k is an integer)
‘Therefore the sine term in equation 5.13.5 becomes, sin4nk and is integer.
For all integer values of k, sin 4nk «0, Hence equation 5.13.5 becomes,
[xoM)-z0(]? _ 2RT
[acaem) a ws (5.13.6)
[rnM-x0D] _
| - 6.137)
Putting this value in equation 5.132 we get error probability of ASK using
matched filter detection as,
-~ (5.13.8)
This is the expression for error probability of ASK using matched filter detection.
44 Error Probability of Binary FSK
The observation vector x has two elements x, and x, that are defined by, respectively,
Ty
m= [ x(t), (t) dt (6.92)
Te
= [, x(t)halt) di (6.93)
where x(r) is the received signal, the form of which depends on which symbol was trans
mitted, Given that symbol 1 was transmitted, x(2) equals 5,(#) + 10(t), where w(t) is he
sample function of a white Gaussian noise process of zero mean and power spectral densi?
No/2. If, on the other hand, symbol 0 was transmitted, x(¢] equals s2(t) + w(t). .
Now, applying the decision rule of Equation (5.59), we find that the observaner
space is partitioned into two decision regions, labeled Z, and Z, in Figure 6.25. 7
decision boundary, separating region Z, from region Z; is the perpendicular bisector
the line joining the two message points. The receiver decides in favor of symbol 1 if the
received signal point represented by the observation vector x falls inside region Z,, This
occurs when x, > x2. If, on the other hand, we have x < x3, the received signal point
falls inside region Z, and the receiver decides in favor of symbol 0. On the decision
boundary, we have x1 = x2, in which case the receiver makes a random guess in favor of
symbol 1 or 0.
Define a new Gaussian random variable Y whose sample value y is equal to the
difference between x, and x35 that is,
yeu (6.94)
The mean value of the random variable ¥ depends on which binary symbol was trans-
mitted, Given that symbol 1 was transmitted, the Gaussian random variables X; and X,,
whose sample values are denoted by x; and x, have mean values equal to VE,, and zero,
respectively. Correspondingly, the conditional mean of the random variable Y, given that
symbol 1 was transmitted, is
E[Y|1] = B[X,|1] — E[X,|1]
6.95)
=+VE, (3s)
On the other hand, given that symbol 0 was transmitted, the random variables X, and Xz,
have mean values equal to zero and VE,, respectively. Correspondingly, the conditional
mean of the random variable Y, given that symbol 0 was transmitted, is
45 ELY|0] = E[X,|0] — E[X2|0]
-VE,
The variance of the random variable Y is independent of which binary symbol was trans-
mitted, Since the random variables X, and X; are statistically independent, each with a
variance equal to No/2, it follows that
var[¥] = var[Xi] + varlX]
= No
Suppose we know that symbol 0 was transmitted. The conditional probability density
function of the random variable Y is then given by
1 (y + VE
0)= = exp| — 279
Fely|0) TaN = 2No
Since the condition x, > x3, or equivalently, y > 0, corresponds to the receiver making a
decision in favor of symbol 1, we deduce that the conditional probability of error, given
that symbol 0 was transmitted, is
(6.96)
(6.97)
(6.98)
Pio = P(y > O|symbol 0 was sent)
=[/ solo % ; (6.99)
1 c _(y + VE, d
Vian, Jo P| 2N |
+ VE,
= (6.100)
0
‘Then, changing the variable of integration from y to z, we may rewrite Equation (6 9)
as follows:
tf
Dio = +! : exp(—z*) dz
— (6.101)
= (ax)
Similarly, we may show the Pox, the conditional probability of error given that symbol 1
was transmitted, has the same value as in Equation (6.101). Accordingly, averaging p,,
and poy we find that the average probability of bit error or, equivalently, the bit error rte
for coherent binary FSK is (assuming equiprobable symbols)
46 x
Bes ete fe) (6.10)
Comparing Equations (6.20) and (6.102), we sec that, in a coherent binary F5K
system, we have to double the bit energy-to-noise density ratio, Ep/No, to maintain the
same bit error rate as in a coherent binary PSK system. This result is in perfect accord with
the signal-space diagrams of Figures 6.3 and 6.25, where we see that in a binary PSK
system the Euclidean distance between the two message points is equal to 2VE,, where
in a binary FSK system the corresponding distance is \/2E,. For a prescribed F,, the
minimum distance d,,;, in binary PSK is therefore V2 times that in binary FSK. Recall
from Chapter 5 that the probability of error decreases exponentially a8 din hence the
difference between the formulas of Equations (6.20) and (6.102).
Error Probability of QPSK
In a coherent QPSK system, the received signal x(z) is defined by
Q<:> 1, we may ignore the quadratic term on the right-hang
side of Equation (6.33), so we approximate the formula for the average probability of
symbol error for coherent QPSK as
[E-
Pie ere Jo (6.34)
The formula of Equation (6.34) may also be derived in another insightful way, using
the signal-space diagram of Figure 6.6. Since the four message points of this diagram are
circularly symmetric with respect to the origin, we may apply Equation (5.92), reproduce
here in the form
1g dix 4
P, == > erfel = for alli (6.35
Py a |
al
Consider, for example, message point m, (corsesponding to dibit 10) chosen as the trans.
mitted message point. The message points », and m, (corresponding to dibits 00 and 11}
are the closest to m,. From Figure 6.6 we readily find that »m, is equidistant from mand
‘mg in a Euclidean sense, as shown by
dy = dis
VIE
49 Assuming that E/Np is large enough to ignore the contribution of the most distant message
point m: (corresponding to dibit 01) relative to am, we find that the use of Equation (6.35)
yields an approximate expression for P, that is the same as Equation (6.34). Note that ia
mistaking either nr, of mts for m1, single bit error is made; on the other hand, in mistaking
my for m,, two bit errors are made. For a high enough E/No, the likelihood of both bits
of a symbol being in error is much less than a single bit, which is a further justification
for ignoring ms in calculating P, when m, is sent,
In a QPSK system, we note that since there are two bits per symbol, thie transmuted
signal energy per symbol is twice the signal energy per bit, as shown by
E=2E, (6.36)
Thus expressing the average probability of symbol error in terms of the ratio B,/No, we
may write
[Ew
Pi = — 6.37)
orfe (J na (
With Gray encoding used for the incoming symbols, we find from Equations (631!
and (6.36) that the bit error rate of QPSK is exactly
1 [Ey
ber =4 erte{ [Ex (638)
2 VNo
‘We may therefore state that a coherent QPSK system achieves the same average probably
of bit error as a coherent binary PSK system for the same bit rate and the same Fs/No
but uses only half the channel bandwidth. Stated in a different way, for the same ExINo
and therefore the same average probability of bit error, a coherent QPSK system transmtt
information at twice the bit rate of a coherent binary PSK system for the same chant
bandwidth, For a prescribed performance, QPSK uses channel bandwidth better than bi-
nary PSK, which explains the preferred use of QPSK over binary PSK in practice.
ERROR PROBABILITY OF BINARY PSK:
To realize a rule for making a decision in favor of symbol 1 or symbol 0,.e partition the signal
space into two regions:
> The set of points closest to message point 1 at +VEy.
® The set of points closest to message point 2 at —VEy
50 This is accomplished by constructing the midpoint of the line joining these two message
points, and then marking off the appropriate decision regions. In Figure 6.3 these decision
regions are marked Z, and Z,, according to the message point around which they are
constructed.
The decision rule is now simply to decide that signal s;(t) (ie., binary symbol 1) was
transmitted if the received signal point falls in region Z,, and decide that signal s,() (ie.
binary symbol 0) was transmitted if the received signal point falls in region Z,. Two kinds
of erroncous decisions may, however, be'made. Signal s,(2) is transmitted, but the noise is
such that the received signal point falls inside region Z, and so the receiver decides in favor
of signal s,(r). Alternatively, signal s,(t) is transmitted, but the noise is such that the re-
ceived signal point falls inside region Z, and so the receiver decides in favor of signal ss).
To calculate the probability of making an error of the first kind, we note from Figure
6.3 that the decision region associated with symbol 1 or signal s(t) is described by
Zu0 0.Equalitycanonlybeachievedif -PU(u)log2 PU(w)=0
Proof. Since 0 < Py(u) <1, we have
=0 if Py(u) =1
>0 fO0
