D T

Spread-Spectrum

IN EN
IA
A , TM
Modulation

AT R
O E E
, K G D JE
LK PA
TY IN R
SI ER TTE
ER E A
by
IV IN H
N G C
U N A
R L E AV
PU A IT

Dr. Amitava Chatterjee
M
AV I A

Electrical Measurement and Instrumentation Laboratory,

Electrical Measurement and Instrumentation Laboratory,
Electrical Engineering Department,
C
D TR

Jadavpur University, Kolkata, India.
JA EC
EL
 Spread-Spectrum Modulation

D T
Whyy Spread-Spectrum
p p ??

IN EN
IA
A , TM
 In digital communication issues of major concern are

AT R
efficient usage of two primary communication resources i.e.

O E E
, K G D JE
LK PA
bandwidth and power.

TY IN R
SI ER TTE
 However,
However there may be situations,
situations where other design

ER E A
IV IN H
objectives will be more important than these two factors.

N G C
U N A
R L E AV
 For example, the system may be required to provide a
PU A IT

form of secure communication in a hostile environment

M

g
such that the transmitted signal is not easily
y detected or
AV I A

recognized by unwanted listeners.

C
D TR
JA EC

 Spread-spectrum modulation is such a technique which is

EL

used to satisfy this requirement.

 Spread-Spectrum Modulation

D T
Features

IN EN
IA
A , TM
 The primary advantage of a spread-spectrum

AT R
O E E
, K G D JE
LK PA
communication
i i system is
i its
i ability
bili to reject
j interference.
i f

TY IN R
SI ER TTE
ER E A
IV IN H
 There can be two types of interference. It may be
N G C
U N A
unintentional interference by another user simultaneously
R L E AV
attempting to transmit through the channel.
PU A ITM
AV I A
C

 Or it may be intentional interference by a hostile

D TR

transmitter attempting to jam the transmission.

transmission
JA EC
EL
 Spread-Spectrum Modulation

D T
Definition
f off Spread
Sp Spectrum
Sp …

IN EN
IA
A , TM
Spread spectrum is a means of transmission in which the data

AT R
O E E
sequence occupies a bandwidth in excess of the minimum

, K G D JE
LK PA
bandwidth required to send it.

TY IN R
SI ER TTE
The spectrum spreading is achieved before transmission through
the
h use off a code
d that
h isi iindependent
d d off the
h data
d sequence. The
Th

ER E A
IV IN H
same code is used in the receiver (operating in synchronism with

N G C
the transmitter) to despread the received signal so that the
U N A
R L E AV
original data sequence may be recovered.
recovered
PU A ITM

Two basic variants of spread-spectrum modulation are:

AV I A

direct-sequence technique and frequency-hopping technique.

C
D TR

Both techniques rely on the availability of a noiselike spreading code

JA EC

called a pseudo-random or pseudo-noise sequence.

EL
 Pseudo-Noise Sequences

D T
IN EN
 A pseudo-noise
d i (PN) sequence isi a periodic
i di bi
binary sequence with
ith a

IA
A , TM
noiselike waveform, usually generated by means of a feedback shift

AT R
O E E
register.

, K G D JE
LK PA
TY IN R
SI ER TTE
ER E A
IV IN H
N G C
U N A
R L E AV
PU A ITM
AV I A
C
D TR

Feedback shift register.

JA EC

The flip-flops are regulated by a single timing clock. At each clock

EL

pulse, the state of each flip-flop is shifted to the next one.

 Pseudo-Noise Sequences

D T
IN EN
IA
A , TM
AT R
O E E
, K G D JE
LK PA
TY IN R
SI ER TTE
ER E A
IV IN H
 Let sj(k) denote the state of the jth flip-flop after the kth clock pulse
N G C
U N A
(the state may be 0 or 1). The state of the shift register after the kth
R L E AV
clock pulse is defined by the set {s1(k), s2(k), …, sm(k)}, where k  0.
PU A ITM
AV I A
C
D TR
JA EC

s0(k): input applied to the first flip-flop after the kth clock pulse.
EL
 Pseudo-Noise Sequences

D T
Observations
v and conclusions …

IN EN
IA
A , TM
 With a total number of m flip-flops, the number of possible states of

AT R
O E E
g
the shift register is at most 2m. Then the PN sequence
q ggenerated byy a

, K G D JE
LK PA
feedback shift register must eventually become periodic with a

TY IN R
SI ER TTE
period of at most 2m.

ER E A
IV IN H
 A feedback shift register is said to be linear, when the feedback logic
N G C
U N A
consists entirely of modulo-2 adders. In such a case, the zero state
R L E AV
(the state for which all the flip-flops are in state 0) is not permitted.
PU A ITM
AV I A
C
D TR

 Consequently, the period of a PN sequence with m flip-flops cannot

JA EC

exceed 2m – 1. When the period is exactly 2m – 1, the PN sequence is

EL

called a maximal-length sequence or simply m-sequence.

 Pseudo-Noise Sequences
An Example …

D T
IN EN
IA
Let us consider a linear feedback shift register involving three flip-flops.

A , TM
AT R
O E E
, K G D JE
LK PA
TY IN R
SI ER TTE
ER E A
IV IN H
N G C
U N A
R L E AV
PU A ITM
AV I A
C

Maximal-length sequence generator for m = 3.

D TR
JA EC

Let the initial state of the shift register be 100 (reading the contents of the
EL

flip-flops from left to right). Then the successive states will be: 100, 110,
111, 011, 101, 010, 001, 100, …. So, the output sequence is: 00111010….
 Spread-Spectrum Modulation for
Baseband Transmission

D T
IN EN
IA
A , TM
Spread-spectrum modulation can provide protection against

AT R
O E E
, K G D JE
LK PA
externally generated interfering (jamming) signals with finite power
power.

TY IN R
SI ER TTE
Protection
P t ti against
i t jamming
j i waveforms
f i provided
is id d b
by purposely
l

ER E A
IV IN H
making the information bearing signal occupy a bandwidth far in
N G C
U N A
excess of the minimum bandwidth necessary to transmit it. This has
R L E AV
the effect of making the transmitted signal assume a noiselike
PU A IT

appearance so as to blend into the background.

M
AV I A
C
D TR

One method of widening the bandwidth of an information bearing

(data) sequence involves the use of modulation.
modulation
JA EC
EL
 Spread-Spectrum Modulation for
Baseband Transmission

D T
IN EN
IA
A , TM
AT R
O E E
, K G D JE
LK PA
TY IN R
SI ER TTE
ER E A
IV IN H
N G C
Transmitter for an idealized model of
U N A
baseband spread
spread-spectrum
spectrum system.
R L E AV
PU A IT

{bk}: a binary data sequence. {ck}: a PN sequence.

M
AV I A

b(t) and
d c(t):
(t) their
th i respective
ti polar
l NRZ waveform
f representations.
t ti
C
D TR

If the message signal b(t) is narrowband and the PN signal c(t) is wideband,
the product (modulated) signal,
signal m(t)
m(t)= b(t)c(t),
b(t)c(t) will have a spectrum that is
JA EC

nearly the same as the wideband PN signal. Hence, PN sequence performs

EL

the role of a spreading code.

Spread-Spectrum Modulation for
Baseband Transmission

D T
IN EN
IA
A , TM
AT R
O E E
, K G D JE
LK PA
TY IN R
SI ER TTE
ER E A
IV IN H
Illustration of the input and
N G C output waveforms in the
U N A
R L E AV
transmitter.
PU A ITM
AV I A
C
D TR
JA EC
EL
 Spread-Spectrum Modulation for
Baseband Transmission

D T
IN EN
IA
A , TM
AT R
m(t)

O E E
b(t)

, K G D JE
LK PA
x

TY IN R
SI ER TTE
ER E A
c(t)
()

IV IN H
mt   bt  ct 
N G C
U N A
R L E AV

M  f   B f * C  f 
PU A ITM
AV I A
C

B(f)
D TR

M(f)
JA EC
EL
 Spread-Spectrum Modulation for
Baseband Transmission

D T
IN EN
IA
A , TM
AT R
O E E
, K G D JE
LK PA
TY IN R
SI ER TTE
ER E A
IV IN H
N G C
Channel for an idealized model of
U N A
R L E AV
b
baseband
b d spread-spectrum
d t system.
t
PU A ITM

The received signal r(t) consists of the transmitted signal m(t) plus an
AV I A

additive
dditi iinterference
t f i( ) as shown
i(t), h iin th
the channel
h l model.
d l
C
D TR
JA EC
EL
 Spread-Spectrum Modulation for
Baseband Transmission

D T
IN EN
IA
A , TM
AT R
O E E
, K G D JE
LK PA
TY IN R
SI ER TTE
ER E A
IV IN H
N G C
Receiver for an idealized model of
U N A
R L E AV
b
baseband
b d spread-spectrum
d t system.
t
PU A ITM

To recover the original message signal b(t), the received signal r(t) is
AV I A

applied to a demodulator that consists of a multiplier, followed by an

C

integrator, and a decision device.

D TR
JA EC

The multiplier is supplied with a locally generated PN sequence that is an

EL

exact replica of that used in the transmitter.

 EL
JA EC
D TR
AV I A
C
PU A ITM
R L E AV
U N A
N G C
IV IN H
ER E A
The multiplier output in the receiver:

SI ER TTE
TY IN R
, K G D JE
O E E
LK PA
AT R
Baseband Transmission

Also:

A , TM
IN EN
D T
Spread-Spectrum Modulation for

IA
 Spread-Spectrum Modulation for
Baseband Transmission

D T
IN EN
IA
A , TM
AT R
O E E
, K G D JE
LK PA
TY IN R
SI ER TTE
ER E A
Conclusion: The data signal b(t) is reproduced at the

IV IN H
N G C
multiplier output in the receiver, except for the effect of the
U N A
interference represented by the additive term c(t)i(t).
R L E AV
PU A ITM
AV I A

The
h ddata signal
i l b(t) is
i narrowband
b d and d the
h spurious
i
C

component c(t)i(t) is wideband. Hence the multiplier output

D TR

is applied to a low-pass filter with a bandwidth just large

JA EC

enough to accommodate the recovery of b(t) and thus the

EL

effect of interference is significantly reduced at the receiver

output.
 Spread-Spectrum Modulation for
Baseband Transmission

D T
IN EN
IA
A , TM
AT R
O E E
, K G D JE
LK PA
TY IN R
SI ER TTE
ER E A
In the receiver, the low-pass filtering is actually performed

IV IN H
N G C
by an integrator. The integration is carried out over the bit
U N A
interval 0  t  Tb.
R L E AV
PU A ITM
AV I A
C

The decision device makes a decision for the receiver, based

D TR

on the sign of v, the integrator output. If v < 0, the receiver

JA EC

i f
infers that
h symbolb l 0 was sent. Iff v > 0, the
h receiver
i iinfers
f
that symbol 1 was sent.
EL
 Spread-Spectrum Modulation for
Baseband Transmission

D T
IN EN
IA
Final Conclusion …

A , TM
AT R
O E E
, K G D JE
LK PA
The llonger the
Th h period
i d off the
h spreading
di code
d (with
( i h pseudo-
d

TY IN R
random properties), the closer will the transmitted signal be

SI ER TTE
to a truly random binary wave, and the harder it will be to

ER E A
detect.

IV IN H
N G C
U N A
R L E AV
A price
Any i paid
id ??
PU A ITM
AV I A

YES. The price paid for improved protection against

C

interference is increased transmission bandwidth, system

D TR

p y, and p
complexity, processing
g delay.
y
JA EC
EL
 Direct-Sequence Spread-Spectrum
with Coherent BPSK

D T
IN EN
IA
A , TM
AT R
O E E
, K G D JE
LK PA
TY IN R
SI ER TTE
ER E A
IV IN H
N G C
U N A
R L E AV
PU A IT

Transmitter for DS/BPSK scheme.

M
AV I A
C

The transmitted signal x(t) is a direct-sequence spread

D TR

binary phase
phase-shift-keyed
shift keyed (DS/BPSK) signal. The phase
JA EC

modulation (t) of x(t) has one of two values, 0 and ,

EL

depending on the polarities of b(t) and c(t).

 Direct-Sequence Spread-Spectrum
with Coherent BPSK

D T
IN EN
IA
A , TM
AT R
O E E
, K G D JE
LK PA
TY IN R
SI ER TTE
ER E A
IV IN H
N G C
U N A
R L E AV
Polarity of data sequence b(t)
Polarity of data sequence b(t) at 
at
PU A IT

time t
M

+ ‐
AV I A
C

Polarity of PN  + 0 
D TR

sequence c(t)

JA EC

at time t
i ‐ 0
EL

Truth Table for Phase Modulation (t) (in Radians)

Direct-Sequence Spread-Spectrum
with Coherent BPSK

D T
IN EN
IA
A , TM
AT R
O E E
, K G D JE
LK PA
TY IN R
SI ER TTE
Waveforms for the second
stage
t off modulation.
d l ti

ER E A
IV IN H
N G C
(a) Product signal
U N A m(t) = c(t)b(t)).
R L E AV

(b) Sinusoidal carrier.

PU A ITM

(c) DS/BPSK signal.

AV I A
C
D TR
JA EC
EL
 Direct-Sequence Spread-Spectrum
with Coherent BPSK

D T
IN EN
IA
A , TM
AT R
O E E
, K G D JE
LK PA
TY IN R
SI ER TTE
ER E A
IV IN H
N G C
U N A
R L E AV

Receiver for DS/BPSK scheme.

PU A ITM
AV I A
C

Here also, demodulation is performed in two stages:

D TR

Stage 1: reverses the phase shift keying applied to the

JA EC

transmitted signal.
EL

Stage 2: performs spectrum despreading.

 Direct-Sequence Spread-Spectrum
with Coherent BPSK

D T
IN EN
Model
d l for
f Analysis
l …

IA
A , TM
AT R
O E E
, K G D JE
LK PA
TY IN R
SI ER TTE
ER E A
IV IN H
N G C
U N A
R L E AV
PU A IT

Model of DS/BPSK system.

M
AV I A

For model analysis

analysis, it is more convenient to interchange
C

the order of two stages of modulation. Similarly the two

D TR

stages of demodulation are also interchanged.

JA EC

For the interchange operation to be feasable, the incoming

EL

data sequence and the PN sequence must be synchronized.

 Direct-Sequence Spread-Spectrum
with Coherent BPSK

D T
IN EN
Model
d l for
f Analysis
l …

IA
A , TM
AT R
O E E
, K G D JE
LK PA
TY IN R
SI ER TTE
ER E A
IV IN H
N G C
U N A
R L E AV
PU A ITM

The channel output:

AV I A
C
D TR
JA EC

s(t):
( ) binary
bi PSK signal.
i l c(t):
( ) PN signal.
i l j(t):
j( ) interfering
i f i signal.
i l
EL
 Direct-Sequence Spread-Spectrum
with Coherent BPSK

D T
IN EN
Model
d l for
f Analysis
l …

IA
A , TM
AT R
O E E
, K G D JE
LK PA
TY IN R
SI ER TTE
ER E A
IV IN H
N G C
U N A
R L E AV
The coherent
PU A IT

detector input:
M
AV I A
C
D TR

The coherent detector input u(t) consists of a binary PSK signal s(t)
embedded in additive code-modulated
code modulated interference,
interference c(t)j(t).
c(t)j(t) The modulated
JA EC

nature of the latter component forces the interference signal to spread its
EL

spectrum such that the information bits at the receiver output can be
detected more reliably.
 Direct-Sequence Spread-Spectrum
with Coherent BPSK

D T
IN EN
Constraints off Direct-Sequence Technique
T h …

IA
A , TM
The use of a PN sequence to modulate a phase-shift-keyed signal

AT R
O E E
, K G D JE
LK PA
achieves instantaneous spreading of the transmission bandwidth.
bandwidth

TY IN R
SI ER TTE
The ability of such a system to combat the effects of jammers is
determined by the processing gain of the system, which is a

ER E A
function of the PN sequence period.

IV IN H
N G C
The p
processinggg U N A
gain can be made larger
g by y employing
p y g a PN
R L E AV
sequence with narrow chip duration. However, the capabilities of
PU A IT

physical devices used to generate the PN spread-spectrum signals

M

impose a practical limit on the attainable processing gain.

AV I A

Implication …
C
D TR

The processing gain may turn out not large enough to overcome
JA EC

the effects of some jammers of concern. An alternative solution is

EL

to use frequency-hop spread-spectrum technique.

Frequency-Hop Spread-Spectrum Technique

D T
IN EN
Here the
H th data-modulated
d t d l t d carrier
i hops
h randomly
d l from
f one

IA
A , TM
frequency to another. In effect, the spectrum is spread
sequentially rather than instantaneously. Here the term

AT R
O E E
, K G D JE
LK PA
sequentially refers to the pseudo-random-ordered
pseudo random ordered sequence of
frequency hops.

TY IN R
SI ER TTE
A common modulation format for FH systems is that of MM-ary
ary

ER E A
frequency-shift keying (MFSK). The combination of these two

IV IN H
N G C
techniques is called FH/MFSK.
U N A
R L E AV
F
Frequency-hop
h T
Techniques
h i
PU A ITM
AV I A
C

Slow-frequency hopping Fast-frequency hopping

D TR

(symbol rate Rs of the (hop rate Rh is an integer

JA EC

MFSK signal is an integer multiple of symbol rate Rs

EL

multiple of hop rate Rh) of the MFSK signal)

 EL
JA EC
D TR
AV I A
C
PU A ITM
R L E AV
U N A
N G C
IV IN H
ER E A
SI ER TTE
TY IN R
, K G D JE
O E E
LK PA
AT R
A , TM
Slow-Frequency Hopping

IN EN
D T
Transmitter for frequency-hop spread M-ary FSK.

IA
 EL
JA EC
D TR
AV I A
C
PU A ITM
R L E AV
U N A
N G C
IV IN H
ER E A
SI ER TTE
TY IN R
, K G D JE
O E E
LK PA
AT R
A , TM
Slow-Frequency Hopping

IN EN
Receiver for frequency-hop spread M-ary FSK.

D T
IA
 Slow-Frequency Hopping
An Example …

D T
IN EN
(a) Frequency variation for one

IA
A , TM
complete period of the PN
sequence.

AT R
O E E
, K G D JE
LK PA
(b) Variation of dehopped

TY IN R
frequency with time.

SI ER TTE
ER E A
Number of bits per MFSK

IV IN H
symbol = 2  M = 4

N G C
U N A Rs = Rb/2
R L E AV
Rc = max(Rh, Rs) = Rs
PU A IT

Length of PN segment
M
AV I A

per hop (k) = 3

Total number of
C

frequency hops = 2k = 8
D TR
JA EC

Parameters of the
EL

FH/MFSK signal.
 Fast-Frequency Hopping
Features …

D T
IN EN
IA
A , TM
In a fast FH/MFSK system, there are multiple hops per M-ary

AT R
O E E
symbol. Hence, in a fast FH/MFSK system, each hop is a

, K G D JE
LK PA
chip. Usually fast frequency hopping is used to defeat a

TY IN R
smart jammer’s tactic.

SI ER TTE
H to
How t recover data
d t att the
th Receiver
R i ??

ER E A
IV IN H
N G C
The data recovery can be performed by noncoherent
U N A
R L E AV
detection at the receiver. The detection procedure can be
PU A IT

implemented in two ways:

M
AV I A

Procedure
d 1: For each
h FH/MFSK
/ symbol,
b l separate decisions
d i i are
C

made on the K frequency-hop chips received and a majority vote

D TR

based rule is used to estimate the dehopped MFSK symbol.

JA EC

Procedure 2: For each FH/MFSK symbol, likelihood functions are

EL

computed as functions of the total signal received over K chips

and the largest one is selected. A receiver so designed is optimum.
 Fast-Frequency Hopping
An Example …

D T
IN EN
(a) Variation of the

IA
A , TM
transmitter frequency with
time.

AT R
O E E
, K G D JE
LK PA
(b) Variation of dehopped

TY IN R
frequency with time.

SI ER TTE
Number of bits p
per MFSK

ER E A
symbol = 2  M = 4

IV IN H
N G C
Rs = Rb/2
U N A
R L E AV
Rc = max(Rh, Rs) = Rh
PU A IT

Length of PN segment
M

per hop (k) = 3

AV I A

Total number of
C

frequency hops = 2k = 8
D TR

Parameters of the
JA EC

FH/MFSK signal.
EL
 References

D T
IN EN
IA
A , TM
 Simon Haykin, Communication Systems. 4th Edition, Wiley India

AT R
O E E
, K G D JE
LK PA
Edition 2008.
Edition, 2008

TY IN R
SI ER TTE
ER E A
IV IN H
 Bernard Sklar, Digital Communication: Fundamentals and

N G C
Applications. 2nd Edition, Pearson Education, 2007.
U N A
R L E AV
PU A ITM
AV I A
C
D TR
JA EC
EL
 EL
JA EC
D TR
AV I A
C
PU A ITM
R L E AV
U N A
N G C
IV IN H
ER E A
SI ER TTE
TY IN R
, K G D JE
O E E
LK PA
AT R
A , TM
IN EN
D T
IA