Modern University

For Technology and Information

Electrical Engineering Department

Lectures Notes of
Digital
Communication
Systems
ELTE 312

Prepared By
Prof Dr: Elsayed Soleit
(2nd Edition 2021)

1
Digital Communication systems, 2nd Ed

Digital Communication Systems

By

Prof. Elsayed Soleit

Electrical Engineering Department

Faculty of Engineering
MTI University
Cairo 2021

1
Digital Communication systems, 2nd Ed

Vision
The vision of the Faculty of Engineering at MTI university is to be a
center of excellence in engineering education and scientific research
in national and global regions. The Faculty of Engineering aims to
prepare graduates meet the needs of society and contribute to
sustainable development.

Mission
The Faculty of Engineering MTI university aims to develop
distinguished graduates that can enhance in the scientific and
professional status, through the various programs which fulfill the
needs of local and regional markets. The Faculty of Engineering
hopes to provide the graduates a highly academic level to keep up the
global developments.

2
Digital Communication systems, 2nd Ed

Course contents

1 Introduction to digital communication 4

systems

2 System and signals 10

3 Sampling and Quantization 24

4 Pulse code Modulation 37

5 Differential Pulse Code Modulation, 44

DPCM

6 Baseband system transmission 59

7 Digital modulation 84

8 Spread Spectrum in Communication 95

Systems

9 References 108

3
Digital Communication systems, 2nd Ed

Chapter One
Introduction
1.1 Course Objective
The assigned course of digital communication systems aims to realize the
followings:
• Study and analysis of stationary and ergodic process
• Understanding the base band digital transmission
• Analyze the pulse code modulation
• Study of the FDMA, TDMA and CDMA.
• Give students an understanding of the benefits and challenges of digital
communication systems.
• Description of ASK, FSK, PSK, DPSK, QAM, QPSK modulation
schemes –
1.2 Fundamental of Communication System Components
The general block diagram of communication systems is shown in Fig.1

Fig.1 The general communication system model

1- Source: It is basic module that originates and prepare the source of message,
such as a human voice, a television images, a teletype message (used for
telegraph) or data signals.
Input transducer: It is the interface module between the users and the digital
transmitter modules. responsible to format the input signals and data to be
compatible with the digital transmission techniques. It converts the nonelectrical

4
Digital Communication systems, 2nd Ed

messages (e.g., acoustic voice, etc.) into electrical waveforms (signals) called
baseband or message signals.
2- Transmitter: The transmitter unit is to modify (adjust) the message
signals to make it possible (efficient) for digital transmission and processing.
The fundamental modules of the transmitter unit are the source encoder and the
digital modulation.
Source Encoder:
In digital communication we convert the signal from the source module into
digital signal. The point to use as few binary digits as possible to represent
the signal. In such a way this efficient representation of the source output
results in little or no redundancy. This sequence of binary digits is
called information sequence.
Digital Modulator:
The binary sequence (stream) which is coming from the transducer is passed
to digital modulator module. It in turns converts the binary information
sequence into discrete or continuous modulated carrier signals to be
transmitted through the communication channel. Hence, the digital modulator
techniques map the binary sequences into signal wave forms suitable to
transmit them through the communication channel.
For example, if we have binary data streams outputting from the source module
of “1”and “0”. These binary data are not suitable to be transmitted directly
through the communication channel. The well-known digital modulator
techniques are studied in chapter seven: They are ASK (Amplitude Shift
Keying), FSK (Frequency Shift Keying) and PSK (Phase Shift Keying).
3- Channel:
The communication channel is the physical medium that is used for
transmitting signals from transmitter to receiver. The communication channel
types may be wired or wireless channels.
• The wired channels: are suitable to transmit the analogue and digital
data messages such as telephone signals through the ground public
switched telephone networks (PSTN) Couper or Coaxial cables and
the wired computer networks. Also, the fibre optical communication
channel is using fibre optic transmission lines and the integrated
service digital network (ISDN) are example of data transmission
through the Wired channel.
• The Wireless channels: are suitable to transmit the analogue and the
digital signals. The different modulation techniques are used to
transmit them. The well-known analogue modulation techniques are
AM (amplitude modulation), FM (Frequency modulation) and PM
(Phase modulation). Moreover, the digital messages are transmitted
using the different digital modulation techniques as ASK, FSK and
PSK.

5
Digital Communication systems, 2nd Ed

• Distortion: is any deformation (alteration) of the signal due to

propagation in the channel.
Examples: attenuation or expansion in the signal width.
Noise: is undesirable (unwanted) signal.
Noise signal is random (probabilistic) and unpredictable
• Types of Noise
External noise:
Interference from other signals transmitted on nearby channels.
Automobile ignition radiation. Other sources power lines.
Internal noise:
Thermal motion of electrons in conductors. Random emission.
4- Receiver: re-processes the received signal by undoing the modification made
at the transmitter and channel.
Digital Demodulator:
The digital demodulator is to recover the digital transmitted signals by
inverse processing of the digital modulator. Hence; it converts these
waveforms into the sequence of numbers that represents estimates of the
transmitted data symbols
5- Source Decoder:
At the end, if an analogue signal is desired then the source decoder tries to
decode the information sequence. Then, it results an approximate replica of
the input at the transmitter end.

6- Output transducer: converts electrical signal into its original nonelectrical

baseband form (i.e. message).
6- Destination: is the unit to which the message is communicated (transmitted).

1.3 Advantages of the Digital Systems

1. Better encryption algorithms: Can not be done in analog
communication
2. More reliable data processing
3. Easily reproducible designs
4. Reduced cost
5. Easier data multiplexing
6. Facilitate data compression
1.4 Disadvantages:
1. Heavy signal processing
2. Synchronization is crucial
3. Larger transmission bandwidth
4. Non-graceful degradation

6
Digital Communication systems, 2nd Ed

Fig.3 Signal attenuation and Distortion

1.5 Goals in Communication System Design

1. To maximize transmission rate, R
2. To maximize system utilization, U
3. To minimize bit error rate, Pe
4. To minimize required systems bandwidth, W
5. To minimize system complexity, Cx
6. To minimize required power, Eb/No

7
Digital Communication systems, 2nd Ed

1.6 Organization of the book

This book includes eleven chapters. Chapter Two illustrates System and
signals. Chapter Three presents Sampling and Quantization. Chapter four
explains the pulse code modulation and Baseband system transmission is
demonstrated in Chapter Five. Chapter Six highlights Digital modulation and
Digital Demodulation is explained in Chapter Seven. Chapter Eight refers to
Delta Modulation. TDM and FDM Coding is focused in Chapter Nine.
Chapter Ten illustrates Analog & Digital Communication System behavior in
noise. Chapter Eleven explains the Digital Radio communication

8
Digital Communication systems, 2nd Ed

Exercise-1

1- Draw a simplified Scheme of the Digital Communication Systems

2- Explain the function of each block
3- Identify the function of the communication channel for voice and data
transmission.
4- Compare between the purposes of the modulation techniques to transmit the
analogue and digital messages.
5- Explain the advantage and the disadvantages and the disadvantages of the
digital communication system over the analogue one
6- Clarify the goals of applying the digital communication systems.

9
Digital Communication systems, 2nd Ed

Chapter Two

System and signals

2.1 base band signal

The base band signal is defined as the signal that is generated from the source
without any alteration of its nature. i.e its frequency spectrum does not change,
and it includes all the frequency components as the original one.
The change of the physical nature can be performed via the algorithms of the
signal processing as filtering, frequency conversions, time conversions,
modulation process, signal compression or expansions, etc.
Signal interference in both time and frequency domains may add extra
frequency components or undesired signals.

2.2 Classification of Signals

The base band signals may be classified due to its occurrence and generation
as deterministic or random signals.
1-Deterministic signals

A signal is deterministic means that there is no uncertainty with respect to its

value at any time of occurrence.
The deterministic signals may be periodical or non-periodical ones.
The periodical signals can repeat itself each time called periodical time. The
periodical Deterministic signal is modeled and defined by explicit
mathematical expression. Sinusoidal signals are examples of periodical
deterministic signals as:
x(t) = ACos(2𝜋𝑓𝑡 + 𝜑) (1)
The signal in eq. (1) has three deterministic parameters are well defined by
deterministic values from the beginning of generation and lasts without
change over a time period without intended change.
The signal parameters are:
- Amplitude, A
- Frequency, 𝑓 ,
- Periodical time, T=1/ 𝑓
- Phase, 𝜑 and the signal phase 𝜖{0,2𝜋}
- The angular frequency, 𝜔 = 2𝜋𝑓 (radian/sec) and the radian=
2𝜋 =(360o)

10
Digital Communication systems, 2nd Ed

𝜔
- frequency, 𝑓 =
2𝜋

Example 1:
A deterministic sinusoidal signal is expressed as an electrical signal:
𝑠(𝑡) = 5 sin(2𝜋1000𝑡 + 𝜋/3) (2)
Find out the amplitude, the frequency, the angular frequency, the periodical
time and periodical angles and the phase shift of the signal s(t).

Solution:
- The amplitude of the signal is 5V and
- the frequency, f=1000Hz and
-the angular frequency, 𝜔 = 2𝜋𝑓, = 2𝜋 ∗ 1000 = 2000𝜋
radian/sec
- The periodical time, T=1/𝑓=1/1000=1 msec.
𝑜
- The periodical angle=2𝜋1000𝑇 = 2𝜋1000/1000=2𝜋=360
𝜋
- The phase shift, 𝜑 = = 60𝑜
3

2- random signals
A signal is random means its occurrences possess degree of uncertainty
before the signal occurs.
Random waveforms/ Random processes when examined over a long period
may exhibit certain regularities that can be described in terms of
probabilities and statistical averages to( mean and autocorrelation and cross-
correlation functions).

3. Periodic and Non-periodic Signals

A signal x(t) is called periodic in time if there exists a constant time,
T0 > 0 such that:
𝑥(𝑡) = x(t + T0 ) for − ∞ < t < ∞ (2) (3)
t denotes time independent variable. T0 is the period of x(t).
Example 2:
Identify the periodicity of the following signals and find its periodical time if
it exists.
𝑠1(𝑡) = 5 sin(2𝜋𝑓𝑡 + 𝜋/3)
𝑠2(𝑡) = 5 cos(2𝜋𝑓𝑡 + 𝜋/3)
𝑠3(𝑡) = 5𝑎𝑡
𝑠4(𝑡) = log (𝑡 + 𝑎), t ≥0

11
Digital Communication systems, 2nd Ed

Solution:
The sinusoidal signals 𝑠1(𝑡), 𝑠2(𝑡) can be expressed respectively as:
𝑠1(𝑡) = 5sin (𝜃)
𝑠2(𝑡) = 5cos (𝜃)
Hence,
𝑠1(𝑡) = 5sin (𝜃 + 2𝜋)= 5sin (𝜃)
𝑠2(𝑡) = 5 cos(𝜃 + 2𝜋)=5cos (𝜃)
𝑎𝑛𝑑 𝑡ℎ𝑒 𝑝𝑒𝑟𝑖𝑜𝑑 = 2𝜋 in radian. 2𝜋𝑓𝑇 = 2𝜋
Then: 𝑇=1/𝑓 sec.
One concludes that any sinusoidal signal is periodical with a periodical time
is the inverse of its frequency. If the signal frequency is defined by 1000 Hz
as in the example 1, the periodical time, T=1/1000=1 msec.
• The exponential function s3 is not periodical because s3(t) ≠ s3(t+T)
• The logarithmic function s4 is not periodical because s4(t) ≠ s4(t+T)

4- Analog and Discrete Signals

An analog signal x(t) is a continuous function of time; that is, x(t) is
uniquely defined for all t from ±∞

A discrete signal x(kT) is one that exists only at discrete times; it is

characterized by a sequence of numbers defined for each time, kT, where
k is an integer
T is a fixed time interval and is called sampling time or period.

Example 3:
A sinusoidal signal can be expressed in time domain as :
𝑧(𝑡) = 5 sin(2000𝜋𝑡) is sampled at sampling frequency, Fs=4000 Hz.
Express it in a discrete time kTs, where k is integer number 0,1,2, …,N and
T is defined as the sampling time or sampling period.

Solution:
The signal z(t) can be written as :
𝑧(𝑡) = 5 sin(2𝜋𝑓𝑘𝑇𝑠)
𝑤ℎ𝑒𝑟𝑒 𝑡 = 𝑘𝑇𝑠=k/Fs
Substituting Fs=4000Hz, in Z(t) gives:
2000𝜋
𝑧(𝑘𝑇𝑠) = 5 sin( 𝑘)
𝜋
4000
𝑧(𝑘𝑇𝑠) = 5 sin( 𝑘),k=0,1,2,3,……………
2
𝜋 𝜋
Then: z(0)=0, z(1)=5 5 sin (2) = 5, z(2)= 5 sin(𝜋) = 0, z(3)=5 sin (3 2) = −5,
z(4)= 5 sin(2𝜋) = 0

12
Digital Communication systems, 2nd Ed

7- Energy and Power Signals

The performance of a communication system depends on the received signal

energy; higher energy signals are detected more reliably (with fewer errors)
than are lower energy signals
x(t) is classified as an energy signal if, and only if, it has nonzero but finite
energy (0 < Ex < ∞) for all time, where:
T/2 

lim  x 2 (t) dt  (t) dt

2
Ex = = x
T → −T / 2 − (4)
An energy signal has finite energy but zero average power.
Signals that are both deterministic and non-periodic are classified as energy
signals:

Example-4
For the given energy signal as:
𝑥 (𝑡) = 𝐴0sin (𝜔𝑡)
Calculate the signal energy for one period, T.
Solution:
Using the above equation, the energy signal can be computed as:
𝑇/2
2
𝐸𝑥 = 𝐴0 ∫ (sin (𝜔𝑡))2 𝑑𝑡
−𝑇/2

𝑇/2
𝑒 𝑗𝜔𝑡 − 𝑒 −𝑗𝜔𝑡 2
𝐸𝑥 = 𝐴20 ∫ ( ) 𝑑𝑡
−𝑇/2 2𝑗
𝑇/2
1
𝐸𝑥 = − 𝐴20 ∫ ( 𝑒 2𝑗𝜔𝑡 + 𝑒 −2𝑗𝜔𝑡 − 2)𝑑𝑡
4
−𝑇/2
1 1 𝑇/2
𝐸𝑥 = − 𝐴20 [ (𝑒 2𝑗𝜔𝑡 − 𝑒 −2𝑗𝜔𝑡 ) − 2𝑡]−𝑇/2
4 2𝑗𝜔

1 1 𝑇/2
𝐸𝑥 = − 𝐴20[ (sin (2𝜔𝑡) − 2𝑡)]−𝑇/2
4 𝜔
1 2 1 𝑇/2
𝐸𝑥 = − 𝐴0[ (sin (2𝜔𝑡) − 2𝑡)]−𝑇/2
4 𝜔
1 𝑇/2
𝐸𝑥 = 𝐴20[2𝑡)]−𝑇/2
4
1
𝐸𝑥 = 𝐴20 (2𝑇)
4

13
Digital Communication systems, 2nd Ed

1 2
𝐸𝑥 = 𝐴 𝑇
2 0
𝐼𝑡 𝑖𝑠 𝑐𝑙𝑒𝑎𝑟 𝑡ℎ𝑎𝑡 𝐸𝑥 𝑖𝑠 𝑓𝑖𝑛𝑖𝑡𝑒 𝑡ℎ𝑟𝑜𝑢𝑔ℎ 𝑡ℎ𝑒 𝑡𝑖𝑚𝑒 𝑝𝑒𝑟𝑖𝑜𝑑, 𝑇

8- The Unit Impulse Function

Dirac delta function δ(t) or impulse function is an abstraction an infinitely
large amplitude pulse, with zero pulse width, and unity weight (area under the
pulse), concentrated at the point where its argument is zero.
∞
∫−∞ 𝛿(𝑡) dt= 1 (5)
𝛿(𝑡) = 0 for t ≠ 0
𝛿(𝑡) is bounded at t = 0 (4)

9- Sifting or Sampling Property

∞
∫−∞ 𝑥(𝑡)𝛿(t-t 0 )dt= x(𝑡0 ) (6)

2.3 Auto correlation of an Energy Signal

Correlation is a matching process; autocorrelation refers to the matching of a

signal with a delayed version of itself.
Autocorrelation function of a real-valued energy signal x(t) is defined as:
∞
𝑅𝑥 (𝜏) = ∫−∞ 𝑥 (𝑡) x (t + 𝜏) dt for ∞ < 𝜏 < ∞ (12)
The autocorrelation function Rx(τ) provides a measure of how closely the
signal matches a copy of itself as the copy is shifted τ units in time.
Rx(τ) is not a function of time; it is only a function of the time difference τ
between the waveform and its shifted copy.
The autocorrelation function of a real-valued energy signal has the following
properties:
𝑅𝑥 (𝜏) =R 𝑥 (−𝜏) (13)
symmetrical in about zero
𝑅𝑥 (𝜏) ≤ 𝑅𝑥 (0) for all 𝜏 (14)
maximum value occurs at the origin.
𝑅𝑥 (𝜏) ↔ 𝜓𝑥 (𝑓)
Autocorrelation and ESD form a Fourier transform pair, as designated
by the double-headed arrows value at the origin is equal to
∞
𝑅𝑥 (0) = ∫−∞ 𝑥 2 (𝑡)dt (15)

14
Digital Communication systems, 2nd Ed

- Autocorrelation of a Power Signal

Autocorrelation function of a real-valued power signal x(t) is defined as:
T /2
1
R x ( ) = lim  x(t) x (t +  ) dt for - <  < 
T → T −T / 2
(16)
When the power signal x(t) is periodic with period T0, the
autocorrelation function can be expressed as:
1 𝑇 /2
𝑅𝑥 (𝜏) = ∫−𝑇0 /2 𝑥(𝑡)x (t + 𝜏) dt for -∞ < 𝜏 < ∞ (17)
𝑇0 0
Example-5

Find out the autocorrelation function and state its properties of the given
signal:𝑥(𝑡) = 𝑒 −𝑗𝜔𝑡 for the time, 0 ≤ 𝑡 ≤ 𝑇

Solution:
Using the autocorrelation equation in eq. (17)
1 𝑇 /2
𝑅𝑥 (𝜏) = ∫−𝑇0 /2 𝑒 −𝑗𝜔𝑡 𝑒 −𝑗𝜔(𝑡+𝜏) dt for -∞ < 𝜏 < ∞
𝑇0 0
Then:
1 𝑇0 /2 −2𝑗𝜔𝑡 −𝑗𝜔𝜏
𝑅𝑥 (𝜏) = ∫ 𝑒 𝑒 dt for -∞ < 𝜏 < ∞
𝑇0 −𝑇0/2
1 −𝑗𝜔𝜏 𝑇0 /2 −2𝑗𝜔𝑡
𝑅𝑥 (𝜏) = 𝑒 ∫ 𝑒 dt for -∞ < 𝜏 < ∞
𝑇0 −𝑇0 /2
𝑇0
1 −𝑗𝜔𝜏 −2𝑗𝜔𝑡 2
𝑅𝑥 (𝜏)= 𝑒 [𝑒 ] 𝑇0
−2𝑗𝑤𝑇0 −2
1
𝑅𝑥 (𝜏)= [sin (𝑤𝑇0)𝑒 −𝑗𝜔𝜏
𝑇0
1- Hence its maximum value is :
1
𝑅𝑥 (0) =
[sin (𝑤𝑇0 )
𝑇0
2- It is symmetric) since, | 𝑅𝑥 (𝜏)| = |𝑅𝑥 (−𝜏)|

Example-6

Given an energy signal x(t)=2t for the time, 0 ≤ 𝑡 ≤ 𝑇,

- Determine the autocorrelation function of the given signal x(t).
- Determine its maximum value at 𝜏 = 0 .
- Check its symmetry

15
Digital Communication systems, 2nd Ed

Solution:

The autocorrelation function of x(t) can be defined as:

𝑇
𝑅(𝜏) = ∫0 𝑥 (𝑡)𝑥(𝑡 + 𝜏)𝑑𝑡
Substituting by x(t) = 2t yields
𝑇
𝑅(𝜏) = ∫ 2(𝑡) ∗ 2(𝑡 + 𝜏)𝑑𝑡
0

𝑇
𝑅(𝜏) = 4 ∫ (𝑡 2 + 𝜏𝑡)𝑑𝑡
0
1 1
𝑅 (𝜏 ) = 4[𝑇0 3 𝑡 3 + 𝜏 2 𝑡 2]
4 3 2
𝑅(𝜏) = 6 (2𝑇 + 3𝜏𝑇 )
4
- The maximum value = 𝑅(0) = 3 𝑇 3
4
- 𝑅(−𝜏) = 6 (2𝑇 3 − 3𝜏𝑇 2
- R(𝜏) =≠ 𝑅(𝜏)
Hence, R(𝜏) is not symmetric

2.4 Ensemble Averages (mean)

The first moment of a probability distribution of a random variable X is
called mean value mX, or expected value of a random variable X
∞
𝑚𝑋 = 𝐸{𝑋} = ∫−∞ 𝑥𝑝𝑋 (𝑥) 𝑑𝑥 (18)

The second moment of a probability distribution is the mean-square value of x

as:
∞
𝐸{𝑋2 } = ∫−∞ 𝑥 2 𝑝𝑋 (𝑥)𝑑𝑥 (19)
Central moments are the moments of the difference between X and mX and the
second central moment is the variance of X
𝑣𝑎𝑟( 𝑋) = 𝐸{(𝑋 − 𝑚𝑋 )2}
∞
= ∫−∞(𝑥 − 𝑚𝑋 )2 𝑝𝑋 (𝑥)𝑑𝑥 (20)
Variance is equal to the difference between the mean-square value and the square
of the mean
𝑣𝑎𝑟( 𝑋) = 𝐸{𝑋2 } − 𝐸{𝑋}2 (21)

2.5 Statistical Averages of a Random Process

A random process whose distribution functions are continuous can be

described statistically with a probability density function (pdf).
A partial description consisting of the mean and autocorrelation function are
often adequate for the needs of communication systems.

16
Digital Communication systems, 2nd Ed

Mean of the random process X(t) :

∞
𝐸[𝑋(𝑡𝑘 )] = ∫−∞ 𝑥𝑝𝑋𝑘 (𝑥) 𝑑𝑥 = 𝑚𝑋 (𝑡𝑘 ) (22)
Autocorrelation function of the random process X(t)
𝑅𝑋 (𝑡1 , 𝑡2 ) = 𝐸[𝑋(𝑡1 )𝑋(𝑡2 )] (23)

2.6 Stationarity
A random process X(t) is said to be stationary in the strict sense if none of its
statistics are affected by a shift in the time origin.
A random process is said to be wide-sense stationary (WSS) if two of its
statistics, its mean and autocorrelation function, do not vary with a shift in
the time origin.
𝐸[𝑋(𝑡)] = 𝑚𝑋
𝑅𝑋 (𝑡1 , 𝑡2 ) = 𝑅𝑋 (𝑡1 − 𝑡2 ) (24)
Autocorrelation of a Wide-Sense Stationary Random Process
For a wide-sense stationary process, the autocorrelation function is only a
function of the time difference τ = t1 – t2; 𝑅𝑋 (𝜏) = 𝐸{𝑋(𝑡)𝑋(𝑡 +
𝜏)}𝑓𝑜𝑟 − ∞ < 𝜏 < ∞ (25)
Properties of the autocorrelation function of a real-valued wide-sense
stationary process are:
- Symmetrical in τ about zero
- Maximum value occurs at the origin
2.7 Time Averaging and Ergodicity

When a random process belongs to a special class, known as an ergodic

process, its time averages equal its ensemble averages.
The statistical properties of such processes can be determined by time
averaging over a single sample function of the process.
A random process is ergodic in the mean if
1 𝑇/2
𝑚𝑋 = 𝑙𝑖𝑚 ∫−𝑇/2 𝑋(𝑡)𝑑𝑡 (26)
𝑇𝑥→∞
It is ergodic in the autocorrelation function if
1 𝑇/2
𝑅𝑋 (𝜏) = 𝑙𝑖𝑚 ∫−𝑇/2 𝑋(𝑡)𝑋(𝑡 + 𝜏)𝑑𝑡 (27)
𝑇
𝑥→∞

Example-7

An ergodic digital system whose output sequences belong to the 4

levels line coding {-3, -1,1,3} and the output N=10, sequences are:
x(n)= {1, -1,1,3,-3,-1,-1,3,1,-1]
1- Determine its mean value (1st moment of the sequences, x(n)

17
Digital Communication systems, 2nd Ed

2- Find out the autocorrelation at 𝜏 = 0 ,𝑅0

3- Calculate the Variance

Solution:

The system is an ergodic process, then its ensemble averages =

statistical ones.
1- The mean value, E[x(n)] is defined by:
𝑁
1
𝐸 [𝑥 (𝑛)] = ∑ 𝑥(𝑛)
𝑁
𝑖=1
1
𝐸 [𝑥 (𝑛)] = (1 − 1 + 1 + 3 − 3 − 1 − 1 + 3 + 1 − 1)
10
2
𝐸 [𝑥 (𝑛)] = = 0.2
10
1
2- 𝑅 (0) = ∑𝑁 𝑖=1 𝑥(𝑛)
2
𝑁

=1/10(1+1+1+9+9+1+1+9+1+1]
35
= =3.5
10
3- The variance=E[𝑥(𝑛)2]-([E[x(n)])2
Substituting by the autocorrelation, R(0) and the mean value yields:
The variance=3.5-0.04=3.46

Example-8
Using the MATLAB program , Find out the autocorrelation
function of the signal given in Example-4 as:
𝑥 (𝑡) = 𝐴0sin (𝜔𝑡)
Calculate the signal energy for one period, T and the frequency
𝑓 = 2𝑘𝐻𝑧
T=1/f=1/2*103=0.5 msec.
Taking the step period 10-6 sec.
The number of samples=0.5*10-3/10-6=0.5*103=500 samples.

Program:2.1

% The discrete autocorrelation function, R

fileID = fopen('C:\Users\eseli\OneDrive\Desktop\output.txt', 'w');
for j=1:1000
y(j)=0.0;
end

18
Digital Communication systems, 2nd Ed

for j=1:80
R(j)=0.0;
end
pi=22/7;
for i=1:1000
% x=rand();
z=sin(2*pi*i/500);
y(i)= z;

%disp(z)
end
m=1;
N=400;
while (m<400)
sum=0.0;
for j=1:400
sum=sum+y(j)*y(j+m);
end
x=sum/N;
R(m)=x;
% disp(x)
m=m+5;
fprintf(fileID,'%12.1f \n',x);
%disp(sum);
end
fprintf(fileID,'%12.1f \n',m);
plot(R);
title('Figure o/p R vs m');
fclose(fileID);

19
Digital Communication systems, 2nd Ed

Fig.1 The autocorrelation function, R(m) versus time difference or

delay, m

Example-8

Given the exponential function , u(t)=EXP(-at), a=0.8. Calculate the

autocorrelation function 𝑅𝑋 (𝜏), using the Matlab program . The
sampling frequency is 2kHz. Hence, the input function can be
expressed in discrete time as:
𝑢(𝑛) = 𝐸𝑥𝑝(−0.8𝑛)
Program 2.2

% The discrete autocorrelation function, R

fileID=fopen('C:\Users\eseli\OneDrive\Desktop\output.txt', 'w');
for j=1:1000
y(j)=0.0;
end
for j=1:80
R(j)=0.0;
end
pi=22/7;
20
Digital Communication systems, 2nd Ed

for n=1:100
% x=rand();
z=10*exp(-0.2*n);
y(n)= z;
%disp(z)
end
m=1;
N=20;
while (m<40)
sum=0.0;
for j=1:20
sum=sum+y(j)*y(j+m);
end
x=sum/N;
R(m)=x;
% disp(x)
m=m+1;
fprintf(fileID,'%12.1f \n',x);
%disp(sum);
end
fprintf (fileID,'%12.1f \n',m);
plot(R);
title ('Figure o/p R vs m');
fclose (fileID);

21
Digital Communication systems, 2nd Ed

Fig.2 The autocorrelation function of the exponent anal signal versus, shift, 𝜏

It is obvious that the autocorrelation function is characterized by the following

characteristics.

1. Symmetry
𝑅𝑋 (𝜏) = 𝑅𝑋 (−𝜏)
2. Boundedness and maximum values

𝑅𝑋 (𝜏) ≤ 𝑅𝑋 (0)𝑓𝑜𝑟𝑎𝑙𝑙𝜏
3. Duality between time and frequency domains

𝑅𝑋 (𝜏) ↔ 𝐺𝑋 (𝑓)

4. The signal variance is equivalent to 𝑅𝑋 (0) at the origin

𝑅𝑋 (0) = 𝐸{𝑋2 (𝑡)}

22
Digital Communication systems, 2nd Ed

Exercise-2

1- Compare by drawing the wave forms between, the deterministic and

the stochastic (random) signals
2- Compare by drawing the wave forms between, the periodic and the
non-periodic signals
3- Define the deterministic signals
4- Given the following deterministic signals, 0 ≤ 𝑡 ≤ 𝑇
a- X(t)=cos(wt),
b- X(t)=𝑎−𝑡 sin(wt)
c- X(t)=log(t+1)
- Determine the men, the autocorrelation, and the variance of
the given signals

23
 Digital Communication systems, 2nd Ed

Chapter Three
Sampling and quantization
3.1 Introduction
The sampling of the analog time varying is an essential process to convert
it to digital formats to be processed by the digital signal processor or to
be stored in digital storage media. The sampling frequency should be
chosen according to Nyquist theorem to enable recovery of the original
signal after conversion. If the sampling frequency is lower than the
Nyquist frequency, the recovery of the original signal is accompanied by
the interference of the higher harmonics due to the aliasing effect.
3.2 Sampling processing
The continuous time signal can be converted into discrete time signals via a
sampling process. It is sampled at a regular time interval called sampling
time, 𝑻𝒔 or sampling intervals. as shown in Fig.3.1.

Quantized
Sampled signal
signal

Sampling Quantization Coding

Origin
al
signal 4. The as:
Analog Digital
and And Discrete
Discrete

Fig. 3.1 Sampling Processing

The sampling rate is defined as:

𝐹𝑠 =1/ 𝑇𝑠 (1)
Where Ts is the sampling time or period.

24
Digital Communication systems, 2nd Ed

3.3 Nyquist Theorem

Sampling theorem (Nyquist’s theorem) is used to determine minimum
sampling rate for any signal so that the signal will be correctly restored at the
receiver.

Nyquist’s theorem states that,

“The original information signal can be reconstructed at the receiver with
minimal distortion if the sampling rate in the pulse modulation system is equal
to or greater than twice the maximum information signal frequency”
sampling frequency. It states that the continuous signal, s(t) should be
sampled at a sampling rate ≥
twice the maximum bandwith of the original contiuous signal
Thus, if the maximum harmonic component of the frequency spectrum is f max
then, the minimum sampling rate is 2 fmax.
Moreover, the sampling frequency for a base band signal of maximum
bandwidth, Bw , is ≥ 2 (maxBw ). The sampling frequency is called Nyquist
rate if and only if it is equal to 2 Bmax .
Bmax means the maximum frequency components existing in the original
continuous signal.

Basic condition of sampling process

1) sampling at Fs =2fm(max)
When the modulating is sampled at a minimum sampling frequency, the
frequency spectrum is as shown in figure 3.2
In practice it is difficult to design a low pass filter, in order to restore the
original modulating signal
2) sampling at fs> 2fm(max)
This sampling rate creates a guard band between fm(max) and the lowest
frequency component (fs-fm(max)) of the sampling harmonic as shown in
figure 3.3
Therefore, a more practical LPF can be used to restore the modulating
signal.

3) Sampling at fs < 2fm(max)

When the sampling rate is less than the minimum value, distortion will
occurs. This distortion is called aliasing as shown in figure 3.4
Aliasing effect can be eliminated by using an anti-aliasing filter prior to
sampling and using a sampling rate slightly higher than Nyquist rate
(fs=2W).

25
Digital Communication systems, 2nd Ed

Fig. 3.2 Nyquist sampling spectrum, fs=2fm(max)

Fig. 3.3 Nyquist sampling spectrum, fs>2fm(max

Fig. 3.4 Nyquist sampling spectrum, fs<2fm(max

26
Digital Communication systems, 2nd Ed

3.4 Anti-aliasing filter

• Hence, the continuous time signal is applied to a limiting low pass to

reduce the higher frequency components before sampling. This filter is
known as anti-aliasing filter to reduce the required sampling frequency.
• This condition is necessary to enable recovering of the original signal
from the sampled one to avoid aliasing effect Band-limiting signals (by
filtering) before sampling

Fig.3.5 Anti-aliasing Filter

An anti-aliasing filter is placed before the A/D conversions to limit the
higher frequency components in the input signals before sampling. This filter
enables to use a low sampling frequency.

Anti-aliasing A/D
s(t) filter Conversion

T
Sampling
Fig.3.6 The antialiasing filter configuration

27
Digital Communication systems, 2nd Ed

3.5 Quantization Noise

Each sampled discrete signal, s(kT) can be converted into digital word of n
bits using a device known Analog to digital converter(ADC) as depicted in
fig.3.1. This process is called quantization process. The function of the ADC
is to convert the magnitude of the discrete signal into different levels. The
number of obtained levels, L depends on the precision of the ADC or the word
length of the output digital word signals.
The total number of levels, L can be evaluated as:
L=2𝑛 levels. (2)
n is the width or word length of the converted samples.
The step size, ∆ of the quantizer can be defined as:
2𝑉
∆= 𝑚𝑎𝑥 (3)
2𝑛
𝑉𝑚𝑎𝑥 is the maximum amplitude of the analog (continuous) time signal.
Let the sample value, q is equivalent to the digitized value of the continuous
time after conversion using the ADC.
This value is not exactly equivalent to the original sample or discrete value,
𝑓𝑠 (𝑘𝑇).
Let the quantization error , 𝑒𝑞 denoted by a uniform random variables and it is
∆
equivalent to ±
2
𝑒𝑞 =𝑉𝑚 -q (4)
Assume a uniform quantizer random error, with zero and variance, 𝜎𝑞 as:
E[𝜎𝑞 ]=0 (it is defined as the expectation of the quantized error noise.
The variance of the quantized error noise can be calculated in terms of the
step size ∆ as follows:
L=2𝑛 levels., n=𝑙𝑜𝑔2 L.
2𝑉𝑚𝑎𝑥
∆=
2𝑛
The quantization noise variance can be defined as:
𝜎𝑞2 =E[𝑒𝑞2 ]-(E[𝑒𝑞 ])2 (5)
The quantization noise can be considered zero mean and its variance in
eq( ) can be written as:
𝜎𝑞2 =E[𝑒𝑞2 ] (6)
2
Then, E[𝑒𝑞 ] can be expressed as:
∆/2 1 1 ∆/2
E[𝑒𝑞2 ]= ∫−∆/2 𝑒𝑞2 d𝑒𝑞 = ∫−∆/2 𝑒𝑞2 d𝑒𝑞
∆ ∆
1 ∆/2 1 ∆2
= [−∆/2𝑒𝑞3 /3]= (∆3/8+ ∆3/8)= (7)
∆ 3∆ 12

∆2
𝑇ℎ𝑢𝑠, 𝜎𝑞2 =
12

28
Digital Communication systems, 2nd Ed

2𝑉𝑚𝑎𝑥
Substituting by ∆= yields:
2𝑛
1
𝜎𝑄2 = 𝑉𝑚𝑎𝑥
2
2−2𝑛 (8)
3

Then the SNR of the quantization process can be written as:

𝑃
𝑆𝑁𝑅 = 2 (9)
𝜎𝑄
P is known as the max power of the sampling signal. Hence.
2
𝑉𝑚𝑎𝑥 22𝑛 3
SNR=3 2 = 22𝑛 (10)
2 𝑉𝑚𝑎𝑥 2
.
If n=n+1, the increase in SNR will be:
∆𝑆𝑁𝑅 = 4
∆𝑆𝑁𝑅(𝑑𝐵) =10 𝑙𝑜𝑔10 4=6dB,
Thus, the SNR increases 6db as the word length of the digitized word
increases by one bit.
If each a digitized sample is represented by B bits, then
Rs = Bit rate = B bits/sample x F samples/second

Exampl-1
- Calculate the minimum sampling rate if a signal, x(t) is
defined as:
X(t)=5 sin(10000𝜋𝑡).
- Deduce the minimum bit rate if each sample is quantized by
8 bits word
Solution:
- According to the Nyquist theorem, the minimum sampling
rate can be defined as;
Fs≥2fmax
2𝜋fmax = 10000𝜋 and fmax=5000Hz
Hence, minimum Fs=2x5000=10kHz
- The minimum bit rate =word length x Sampling frequency
- R=10000x8=80k bps (bit per second)
Example-2
A speech signal of maximum amplitude Vm =10 volt and it
is quantized by 12 bits word.
- Find out the step size in Volts
- Determine the variance of the quantization error
- Calculate the SNR in dB
-

29
Digital Communication systems, 2nd Ed

Solution
- The number of Levels L=212
Vm 10
- The step size, ∆= = Volt
212 212
∆2 1 100
- The variance,𝜎𝑞2 = =
12 12 224
3 2𝑛 3 24
- SNR= 2 = 2 =3𝑥223
2 2

Example-3

- A signal s(t) is defined by:

𝒔(𝒕) = 𝟓 𝒔𝒊𝒏 (𝟏𝟎𝟔 𝝅𝒕)

- Calculate the minimum sampling frequency, 𝑭𝒔 [5 Marks]
- Deduce the minimum bit rate if each sample is quantized by 8 bits word [5
Marks]
- Determine the improvement in SNR if the assigned word length is increased
by 2 bits. [3 Marks]
Ans:

- The minimum sampling frequency is equal to the equivalent Nyquist

frequency. Then, 𝑭𝒔 = 𝟐 𝒇𝒎𝒂𝒙
𝟏
𝑭𝒔 = 𝟐 ∗ 𝟏𝟎𝟔 ∗ 𝟐=𝟏𝟎𝟔 𝑯𝒛=1MHz
- The minimum bit rate =𝑭𝒔 ∗ 𝟖=8𝒙𝟏𝟎𝟔 bps=8 Mbps
- An increase in word length by 1 bit yields 6 dB improvement in SNR.
Hence ; 2 bits increase yields 12 dB improvement in SNR.

Example-4

The signal given in example-3 is sampled using the sampling frequency in the
following cases:

1- Under sampling as 𝐹𝑠 < 𝑁𝑒𝑞𝑢𝑖𝑠𝑡 𝑓𝑟𝑒𝑞𝑢𝑒𝑛𝑐𝑦 (2 𝑓𝑚𝑎𝑥 ),

𝐹𝑠 = 3𝑘𝐻𝑧

2- Neyquist sampling frequency, 𝐹𝑠 = (2 𝑓𝑚𝑎𝑥 )

𝐹𝑠 = 4𝑘𝐻𝑧
3- Oversampling as𝐹𝑠 ≫ 𝑁𝑒𝑞𝑢𝑖𝑠𝑡 𝑓𝑟𝑒𝑞𝑢𝑒𝑛𝑐𝑦 (2 𝑓𝑚𝑎𝑥 )

𝐹𝑠 = 8𝑘𝐻𝑧

30
Digital Communication systems, 2nd Ed

Using the Matlab program to analyse the above cases:

Under Sampling signal

100%
80%
60%
40%
20%
0%
1
5
9
13
17
21
25
29
33
37
41
45
49
53
57
61
65
69
73
77
81
85
89
93
97
-20%
-40%
-60%
-80%
-100%

Fs=3000Hz

Program 3.1

% The digital Sampling signal

fileID=fopen('C:\Users\eseli\OneDrive\Desktop\output.txt', 'w');
for j=1:1000
y(j)=0.0;
end
for j=1:80
R(j)=0.0;
end
pi=22/7;
fmax=2000;
Fsu=3000;
Fs=3000;
T=1/Fs
%x=10*sin(2*pi*f*t)
%x=10*sin(4000*pi*k*T)
for k=1:8000
% x=rand();
x=10*sin(4000*pi*k*T)
% if(k<=100)

31
Digital Communication systems, 2nd Ed

y(k)= x;
% end
disp(x)
fprintf(fileID,'%12.1f \n',x);
end
fprintf(fileID,'%12.1f \n',Fs);
plot(y);
title('Figure Undersampling signal');
fclose(fileID);

Neyquist Sampling
100%
80%
60%
40%
20%
0%
1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64 67
-20%
-40%
-60%
-80%
-100%

Fs=4000Hz

Oversampling
100%
80%
60%
40%
20%
0%
1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64 67
-20%
-40%
-60%
-80%
-100%

Fs=8000Hz

32
Digital Communication systems, 2nd Ed

Example-5

Multiple signals can be sampled using under , Nyquist and over sampling
frequency , Fs. In this example a Matlab program that executes the
sampling process of multiple different frequencies signals.

Program-3.2

% The digital Sampling signal

fileID=fopen('C:\Users\eseli\OneDrive\Desktop\output.txt', 'w');
for j=1:1000
y(j)=0.0;
end
for j=1:80
R(j)=0.0;
end
pi=22/7;
fmax1=200;
fmax2=300;
fmax3=400;
Fsu=300;
Fs=800;
T=1/Fs;
r1=fmax1/Fs;
r2=fmax2/Fs;
r3=fmax3/Fs;
%x=10*sin(2*pi*f*t)
%x=10*sin(4000*pi*k*T)
for k=1:1000
% x=rand();
x=10*(cos(pi*r1*k)+cos(pi*r2*k)+cos(pi*r3*k));
%x=10*(cos(pi*r3*k));
y(k)= x;
%disp(x)
fprintf(fileID,'%12.1f \n',x);
end
%s=abs(fft(y));
%fprintf(fileID,'%12.1f \n',Fs);
plot(y);
title('Figure fft(fs=2fm)');
fclose(fileID);
The results are in the following cases:

33
Digital Communication systems, 2nd Ed

Fmax1=200Hz
Fmax2=300 Hz
Fmax3=400 Hz

Case:1
Fs=200Hz
Under sampling, Fs<2fmax
40

30

20

10

0
41

201

261

321

481

541
1
21

61
81
101
121
141
161
181

221
241

281
301

341
361
381
401
421
441
461

501
521

561
-10

-20

-30

Case:2
Fs=800Hz=2fmax (Nyquist frequency)
Sampling of 2multiple signals, Fs>2fmax
40

30

20

10

0
116
139
162
185
208
231
254
277
300
323
346
369
392
415
438
461
484
507
530
553
576
599
622
645
1
24
47
70
93

-10

-20

-30

34
Digital Communication systems, 2nd Ed

Case:3
Fs=1600Hz (Over sampling)

3 Multiple sampling signals, Fs>>2fmax

40

30

20

10

0
1

235

271

307
19
37
55
73
91
109
127
145
163
181
199
217

253

289

325
343
361
379
397
415
433
451
469
487
505
-10

-20

-30

Fig. The FFT of 3 Multiple signals with Fs=1600Hz =8fmax

35
Digital Communication systems, 2nd Ed

Exercise-3
1- Explain the relationship between the word length of the
quantized word and the quantization noise.
2- if a signal, x(t) is defined as:
X(t)=5 sin(106𝜋𝑡).
-Calculate the minimum sampling rate
-Deduce the minimum bit rate if each sample is
quantized by 8 bits word
3- A speech signal of maximum amplitude Vm =5 volt and it
is quantized by 16 bits word.
- Find out the step size in Volts
- Determine the variance of the quantization error
- Calculate the SNR in dB
- Determine the improvement in SNR if the word length is increased
by 2 bits.

36
Digital Communication systems, 2nd Ed

Chapter Four

Pulse Code Modulation

4.1 Introduction
Pulse modulation includes many different methods of converting
information into pulse form for transferring pulses from a source to a
destination. It is divided into two categories.
• Analog Pulse Modulation (APM)
• Digital Pulse Modulation (DPM
The analog information signal is sampled and Converting samples into
discrete pulses. Transport the pulses over physical transmission medium.
There are Four Methods of pulse modulation
1) Pulse amplitude modulation (PAM)
2) Pulse width modulation (PWM)
3) Pulse position modulation (PPM)
4) Pulse code modulation (PCM)
The types PAM, PWM and PPM are known as analogue pulse modulation
and the PCM represents the digital pulse modulation.
4.2 ANALOG PULSE MODULATION
• In Analog Pulse modulation (APM), the carrier signal is in the form of
pulse waveform, and the modulated signal (message or information) is
where one of the characteristics (either amplitude, width, or position) is
changed according to the modulating signal (Audio or music or data).
• The three common techniques of APM are: Pulse Amplitude Modulation
(PAM), Pulse Width Modulation (PWM) and Pulse Position
Modulation (PPM). Fig.4.1 Explains the wave forms of the PAM, PWM
and PPM.
4.2.1 Pulse Amplitude Modulation (PAM)
It is the simplest form of pulse modulation. The carrier signal (Amplitude,
Width and Position) is varied according to the amplitude of the modulating
signal.
Basically, the modulating signal is sampled by the digital train of pulses and the
process is based upon the sampling theorem. It is apparent from Fig.4.1 that the
amplitude of the PAM is varying according to the amplitude of the modulating
signal. The pulse train amplitudes within one period of the modulating signal
bears the information represented by the modulating signal.

37
Digital Communication systems, 2nd Ed

Fig. 4.1 The APM wave forms

4.2.2 Pulse Width Modulation (PWM)

Varying the width of the constant amplitude pulse proportional to the amplitude
of the modulation signal is as PWM. It is also known as Pulse Duration
Modulation (PDM).
Either the leading edge, trailing edge or both may be varied by the modulating
signal.
PWM gives better signal to noise performance than PAM.
PWM has an advantage, that it has a varying power content according the
variation of pulse duration time (pulse width). PWM still works if
synchronization between transmitter and receiver fails.

38
Digital Communication systems, 2nd Ed

Fig. 4.2 PWM pulse width modulation waveform

4.2.3 Pulse Position Modulation (PPM)

PPM is when the position of a constant-width and constant-amplitude pulse
within prescribed time slot (periodical time or period) is varied according to the
amplitude of the modulating signal.
PPM has the advantage of requiring constant transmitter power output, but the
disadvantage of depending on transmitter-receiver synchronization.
PPM has less noise due to the amplitude changes because the received pulses may
be clipped at the receiver, thus removing amplitudes changes caused by noise.
4.2.4 DIGITAL PULSE MODULATION (DPM)

Pulse Code Modulation (PCM)

PCM is a form of digital modulation where group of coded pulses are used to
represent the discrete analog signal after sampling. The analog signal is
sampled and converted to a fixed length (word length), serial binary number
for transmission. A block diagram of a PCM system is as shown in Fig 4.3.

39
Digital Communication systems, 2nd Ed

Fig. 4.3 The PCM generation

The continuous time message is applied t a Low pass filter to limit the
frequency contents and to remove the higher frequencies and thus, a
reasonable sampling frequency can be used. The output signal is sampled at a
sampling rate greater or equal to the Nyquist frequency. Each sample is
quantized by a fixed word length (8 bits or greater) of binary bits “0” and “1”
stream. The quantizer digital word is encoded and applied to the transmission
channel.
Due to the imperfection of the channel, the encoded pulses is distorted and
need regeneration and shaping via the repeater before the receiver.
The received signal is regenerated and decoded and applied to a reconstruction
filter to recover the original message.

4.3 Principles of PCM

Three main processes in PCM transmission are sampling, quantization, and
coding.
1. Sampling – is a process of taking samples of information signal at a rate of
Nyquist’s sampling frequency.
2. Quantization – is a process of assigning the analog signal samples to a
pre- determined discrete level. The number of quantization levels, L,
depends on the number of bits per sample, n, used to code the signal.

Where

40
Digital Communication systems, 2nd Ed

𝑳 = 𝟐𝒏 (1)
The magnitude of the minimum step size of the quantization levels is called
resolution,
It is equal in magnitude to the voltage of the least significant bit of the
magnitude step size of the digital to analog converter (DAC). The resolution
depends on the maximum voltage, Vmax, and the minimum voltage Vmin of
the information signal, where

Vmax − Vmin
V =
L −1 (2)
Quantization error or quantization noise is the distortion introduced during
the quantization process when the modulating signal is not an exact value of
the quantized level. It is the difference between original signal and the
quantized signal magnitude that is:
Quantization error, Qe = |x(t)| - |q(t)|
Where |x(t)| is the magnitude of original signal
Where |q(t)| is the magnitude of quantized signal
The maximum quantization error,
Quantization error can be reduced by increasing the number of quantization
level BUT this will increase the bandwidth required.
ENCODING
This is a process where each quantized sample is digitally encoded into n-
bits codeword, where
n = number of bits/samples
L = number of quantization levels
Transmission bit rate (R) is the rate of information transmission (bits/sec).
It depends on the sampling frequency and the number of bit per sample used
to encode the signal and is given by Transmission bit rate
Transmission Bandwidth.
𝒏 = 𝒍𝒐𝒈𝟐 𝑳
𝑹 = 𝒏 × 𝒇𝒔 𝒃𝒊𝒕𝒔/𝒔𝒆𝒄
𝑩 = 𝒏 × 𝒇𝒔 𝑯𝒛
Examples-1

A sinusoidal input wave of 3kHz is to be sampled at the lowest rate for

transmission as pulses. Calculate the minimum sampling frequency required,
so that all components of the wave can be reconstructed at the receiver.
Solution:

41
Digital Communication systems, 2nd Ed

The minimum sampling frequency=2fmax

Min fs= 2 x3000=6000Hz=6 kHz
Examples-2
The PCM sampled are encoded into 4-bits system. If the minimum sampling
rate used is 8kHz, calculate
the frequency of the information signal
the quantization levels.
the transmission rates
The transmission bandwidth
Solution:
The frequency of the information=4kHz
L=2n=24=16
Transmission rate=8000x4=32kHz
Transmission bandwidth,
𝑩 = 𝒏 × 𝒇𝒔 𝑯𝒛
=32kHz

Example-3
A speech signal of maximum frequency = 4200Hz and minimum frequency
component=200Hz. It is sampled and quantized by 8 bits word.
Find out the minimum sampling rate and the minimum bit rate to transmit
the digitized speech signal.
Solution:
The speech band width=fmax-fmin=4200-200=4000Hz
The min sampling frequency=2 B=2x4000=8000Hz
The minimum bit rate=fsx8=8x8000=64000bps=64kbps
Transmission bandwidth=64kHz

42
Digital Communication systems, 2nd Ed

Exercise-4
1- Explain and Compare between the PAM and PCM pulse modulation and
show the pulse wave forms
2- Mention the advantage and the disadvantage of the PWM and the PPM.
3- A speech signal of maximum frequency = 1000Hz and minimum
frequency component=500Hz. It is sampled and quantized by 12 bits
word.
-Find out the minimum sampling rate and the minimum bit rate to transmit
the digitized speech signal.
4- An image size of 200x100 pixels. Each pixel is quantized by 4 bits.
Calculate the minimum bit rate to transmit the image pixels serially.
5- The PCM sampled are encoded into 8-bits system. If the minimum
sampling rate used is 16kHz, calculate
• the frequency of the information signal
• the quantization levels.
• the transmission rates
• the transmission bandwidth

43
Digital Communication systems, 2nd Ed

Chapter Five

Differential Pulse Code Modulation (DPCM)

5.1 Introduction

 Based on the sampling theorem

 Each analog sample is assigned a binary code
Analog samples are referred to as pulse amplitude modulation (PAM)
samples
 The digital signal consists of block of n bits, where each n-bit number is
the amplitude of a PCM pulse

Pulse code modulation (PCM) is based on the sampling theorem, which states
that If a signal f(t) is sampled at regular intervals of time and at a rate higher than
twice the highest signal frequency, then the samples contain all the information
of the original signal The function f(t) may be reconstructed from these samples
by the use of a low-pass filter. If voice data are limited to frequencies below 4000
Hz, a conservative
procedure for intelligibility, 8000 samples per second would be sufficient to
characterize the voice signal completely. Note, however, that these are analog
samples, called pulse amplitude modulation (PAM) samples. To convert to
digital, each of these analog samples must be assigned a binary code.
Figure 6.15 shows an example in which the original signal is assumed to be band
limited with a bandwidth of B. PAM samples are taken at a rate of 2B, or once
every Ts = 1/2B seconds. Each PAM sample is approximated by being quantized
into one of 16 different levels. Each sample can then be represented by 4 bits. But
because the quantized values are only approximations, it is impossible to recover
the original signal exactly. By using an 8-bit sample, which allows 256 quantizing
levels, the quality of the recovered voice signal is comparable with that achieved

44
Digital Communication systems, 2nd Ed

via analog transmission. Note that this implies that a data rate of (8000 samples
per second) X (8 bits per sample) = 64 kbps is needed for a single
voice signal. Thus, PCM starts with a continuous-time, continuous-amplitude
(analog) signal, from which a digital signal is produced. The digital signal
consists of blocks of n bits, where each n-bit number is the amplitude of a PCM
pulse. On reception, the process is reversed to reproduce the analog signaL
Notice, however, that this process violates the terms of the sampling theorem. By
quantizing the PAM pulse, the original
signal is now only approximated and cannot be recovered exactly. This effect is
known as quantizing error or quantizing noise. The signal-to-noise ratio for
quantizing noise can be expressed as [GIBS93]
SNR dB = 20 log 2n + 1.76 dB = 6.02n + 1.76 dB
Thus, each additional bit used for quantizing increases SNR by about 6 dB, which
is a factor of 4.

5.2 The linear prediction code

Fig.1 The Linear Prediction Code

Transversal Digital filter can perform the prediction process as shown in Fig.5.1
The output, z(nTs) of the Transversal Digital Filer can be expressed using the
linear digital convolution as:
z(nTs ) = ∑Ki=1 ai y(n − i)Ts (1)
For simplicity, we consider, Ts is normalized to 1
{ai } are the filter coefficients.
{ y(n − i)Ts } are known as the input observations.
Hence, the filter output, z(nTs) can be estimated from the delayed input
observations and knowing the optimal coefficients of the linear prediction code
scheme.
Equation (1) can be written in matrix notation as :

45
Digital Communication systems, 2nd Ed

z(nTs ) = 𝐴𝑇 𝐵 (2)
The coefficient vector, A is defined as:
𝐴𝑇 = [𝑎1 𝑎2 𝑎13 … … . 𝑎𝐾 ] (3)
And the observation vector is known as:

𝐵 𝑇 = [𝑦(𝑛 − 1)𝑦(𝑛 − 2) … … … … … 𝑦(𝑛 − 𝐾)] (4)

5.3 The DPCM principle

The error signal resulting from subtracting the output estimated signal, z(nTs ) of
the LPC filter from the original input signal y(nTs) as:
𝑒 (𝑛) = y(nTs)- z(nTs ) (5)

The error signal e(n) can be expressed using eq.(2) as:

𝑒(𝑛) = 𝐴𝑇 𝐵 - z(nTs ) (6)

The error signal is defined as DPCM which is quantized, encoded, and

transmitted through the communication channel to the receiver unit.
The coefficient vector, A is optimally designed such that the mean square error,
E [𝑒(𝑛)2 is minimized [5 ]. The optimal coefficient vector can be estimated as:

𝐴∗ = 𝑅 −1𝑃 (7)
The autocorrelation matrix, R is expressed as:

𝑦 2 (𝑛 − 1) … 𝑦(𝑛 − 1)𝑦(𝑛 − 𝐾)
𝑅 = 𝐸[ ⋮ … ⋮ ] (8)
2
𝑦 𝑛 − 1 𝑦(𝑛 − 𝐾)
( ) … 𝑦(𝑛 − 𝐾)

Where, P is defined as the cross-correlation vector and can be expressed as:

𝑧(𝑛)𝑦(𝑛 − 1)
𝑧(𝑛)𝑦(𝑛 − 2)
P=E[ ]
⋮
𝑧(𝑛)𝑦(𝑛 − 𝐾)

Furthermore, the optimal filter coefficients in eq. (7) can be obtained

iteratively using the adaptation algorithms [5,6]. The well-known Least
mean adaptation algorithm can be written as:
𝑎𝑖 (𝑛) = 𝑎𝑖 (𝑛 − 1) + 2𝜇𝑒 (𝑛)𝑦(𝑛 − 𝑖) (9)
𝑖 = 1,2, . . . . . . . . , 𝐾

46
Digital Communication systems, 2nd Ed

𝜇 is known as the step size of the adaptation algorithm. It controls the

adaptation speed and the resulting adaptation noise.

The transmitter and the receiver of the DPCM processor is depicted in

Figure.2. The received signals are decoded and applied to an Inverse LPC
predictor with same coefficients of that in the transmitter Unit.

The function of the Inverses LPC at the receiver site of the DPCM scheme can
be explained as follows:
The input signal to the Inverse LPC module is defined as:
S(n)=e(n)+x(n) (10)
Where e(n) is known as the transmitted sequences of the DPCM transmitter.
It is defined in eq. (5).
X(n) is the output of the Inverse LPC Filter which can be written as:
𝑥(𝑛) = ∑𝑁 𝑖=1 𝑎𝑖 𝑠(𝑛 − 𝑖) (11)
The output of the Inverse LPC predictor is a good estimate of the original
transmitted signals and it is applied to a smoothing, low pass filter, to cancel
the higher frequency components and to recover the original analog signals.

Fig.2 the transmitter and the receiver of the DPCM processor.

Instead of using one bit to indicate positive and negative differences, we can
use more bits to represent the quantization of the difference.

47
Digital Communication systems, 2nd Ed

Each bit code is used to represent the value of the difference. The more bits
the more levels yield the higher the accuracy.

The following MATLab program explains the software implementation of

the LPC module and the optimal coefficients are obtained after a learning
period. A 500-sample learning period is chosen.

Program-1

% Digital Realisation of the adaptive DPCM Transceiver

% Digital Realisation of adaptive LPC

N=5;
mu=0.05;
pi=22/7;
fileID=fopen('C:\Users\eseli\OneDrive\Desktop\error-output.txt', 'w');
fileID1=fopen('C:\Users\eseli\OneDrive\Desktop\output2.txt', 'w');
for j=1:30
yoe(j)=0.0;
xf(j)=0.0;
c(j)=0.0;
end
for i=1:3000
sumf=0.0;
sumb=0.0;
sume=0.0;
surmse=0.0;
x=rand(); % Data Source
if (x>=0.5) xf(1)=1;
else
xf(1)=-1;
end
% x=sin(pi*i/500);
% xf(1)=x;
for j=1:N %
sumf=sumf+c(j)*xf(j+1) ;
end
yo=sumf;
e=xf(1)-yo;
zo(i)=e;
e2=e*e;
%if (i>1000)
fprintf(fileID,'%12.1f \n',e);

48
Digital Communication systems, 2nd Ed

%end
z(i)=e2;
for j=1:N
c(j)=c(j)+2*mu*e*xf(j+1);
% fprintf(fileID,'%12.1f \n',c(j));
end
for j=1:20
xf(21-j+1)=xf(21-j);
end
end
% Inverse LPC
for j=1:3000
xo=zo(j);
sumbe=0.0;
for j=1:N %
sumbe=sumbe+c(j)*yoe(j+1) ;
end
yoe(1)=xo+sumbe;
%lpco(j)=10*yoe(1);
for j=1:20
yoe(21-j+1)=yoe(21-j);
end
% s=abs(fft(y));
fprintf(fileID1,'%12.1f \n',yoe(1));
end
for j=1:N
fprintf(fileID,'%12.1f \n',c(j));
end
% zz=fft(lpco);
plot(z);
title('Figure output e2 signal)');
fclose(fileID);
fclose(fileID1);

49
Digital Communication systems, 2nd Ed

LPC-Input
1.5

0.5

0
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53
-0.5

-1

-1.5

Fig.3 The input sequences {1,-1}

LPC output
3

0
217
229
241
253
265
277
289
301
313
325
1
13
25
37
49
61
73
85
97
109
121
133
145
157
169
181
193
205

-1

-2

-3

Fig.4 The adaptive LPC output for the input sequences {1,-1}

50
Digital Communication systems, 2nd Ed

Fig.5 The error square versus the iteration numbers, N=3000

Inverse LPC output

4

0
96

210

324

438

533
1
20
39
58
77

115
134
153
172
191

229
248
267
286
305

343
362
381
400
419

457
476
495
514

-1

-2

-3

-4

Fig.6 The output of the DPCM inverse LPC module

51
Digital Communication systems, 2nd Ed

The coefficients of the adaptive LPC module are:

Coefficients Value
A1 0.2
A2 -0.1
A3 -0.2
A4 0.1
A5 -0.3

5.4 Delta Modulation (DM)

In DM, analog input is approximated by staircase function moves up or

down by one quantization level () at each sampling interval.
The bit stream approximates derivative of analog signal (rather than
amplitude)
▪ 1 is generated if function goes up
▪ 0 otherwise
Two important parameters
▪ Size of step assigned to each binary digit ()
▪ Sampling rate
Accuracy is improved by increasing sampling rate, however, this increases
the data rate
Special type of DPCM with M = 2 (2 levels). It is inexpensive and simple to
implement. Fig.3 shows the Delta modulation scheme. The analog input signal
is applied to a LPF to limit the frequency band pass only the low frequency
bands. Then, it is sampled using a reasonable sampling rate. The comparator
device compares the sampled signals (PAM) with the output of the integrated
output of the accumulator unit, z(t). The Comparator output y(t)=1 if the PAM
is > z(t) and y(t)=0 if the PAM < z(t).

Fig.7 The Delta Modulation Scheme

52
Digital Communication systems, 2nd Ed

The delta modulator output DM is considered Pulse terrain with variable

duration as shown in Fig.4b. The DM signal is transmitted through the channel
to the receiver unit. The receiver integrates these signals to recover the analog
estimate of the original signal.

Fig.8 The input/output wave forms of the Delta modulation process

53
Digital Communication systems, 2nd Ed

DM implementation using MATLAB program

Program-2

% Digital Realisation of DM
N=4;
mu=0.05;
pi=22/7;
y=0.0;
fileID=fopen('C:\Users\eseli\OneDrive\Desktop\outputcomp.txt', 'w');
fileID1=fopen('C:\Users\eseli\OneDrive\Desktop\outputintegrator.txt', 'w');
for j=1:30
yc(j)=0.0;
xf(j)=0.0;
z(j)=0.0;
c(j)=0.0;
end
h(1)=1.0;
h(2)=0.5;
h(3)=0.1;
h(4)=0.0;
z(1)=1.0;
for i=1:3000
sumf=0.0;
sumb=0.0;
sume=0.0;
surmse=0.0;
y=0.0;
x=rand(); % Data Source
if (x>=0.5) xf(1)=1;
else
xf(1)=-1;
end
for j=1:N %
y=y+h(j)*xf(j) ;
end
if (y > z(1))yc(1)=1;
else yc(1)=0;
end
sum=0.0;
%_______________Integrator____________ %
% for j=1 :2
% sum=sum+yc(1)+yc(2);

54
Digital Communication systems, 2nd Ed

% end
z(1)=3*(yc(1)+yc(3))/2;
%z(1)=sum+z(2);
zc(i)=yc(1);
fprintf(fileID,'%12.1f \n',yc(1));
fprintf(fileID1,'%12.1f \n',z(1));
for j=1:20
xf(21-j+1)=xf(21-j);
yc(21-j+1)=yc(21-j);
z(21-j+1)=z(21-j);
end
end
% zz=fft(lpco);
plot(zc);
%title('Figure output-comparator signal)');
fclose(fileID);
fclose(fileID1);

Analog Input,w(t)
2

1.5

0.5

0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26
-0.5

-1

-1.5

-2

Fig.9a The Analog Input, w(t)

55
Digital Communication systems, 2nd Ed

1.2
Chart Title
1

0.8

0.6

0.4

0.2

0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26

Fig.10a The output of the Delta Modulation (DM), y(t) wave form

Analog I/P
2

1.5

0.5

0
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55
-0.5

-1

-1.5

-2

Fig.9b The Analog Input, w(t)

56
Digital Communication systems, 2nd Ed

Output DM
1.2

0.8

0.6

0.4

0.2

0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28

Fig.10b The output of the Delta Modulation (DM), y(t) wave form.

Integrator-O/P
4.5
4
3.5
3
2.5
2
1.5
1
0.5
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

Fig.11 The output of the Integrator or accumulator, z(t)

57
Digital Communication systems, 2nd Ed

Exercise-5
1- Explain the operation principle of the LPC digital filter
2- Explain the operation function of the DPCM scheme for both
transmission and receiver
3- Illustrate the function of the comparator device in the Delta modulation
scheme
4- Indicate the importance of the Delta and DPCM modulation
Examples of the last exams questions:

58
Digital Communication systems, 2nd Ed

Chapter Six

Base band system Transmission

6.1 Baseband binary Pulse amplitude modulation (PAM)

A baseband data transmission PAM system is shown in Fig, 6.1. A baseband

binary incoming binary sequence b0 consists of 1 and 0 symbols of duration, Tb.
is the input data source.

Figure 6.1 Baseband binary data transmission system.

The PAM changes this sequence into a new sequence of short pulses each
with amplitude ak, represented in polar form as:

{a k = 1.......if bk = 1}
{ak = −1.....if bk = 0} (1)

The bipolar modulated pulses {1, -1} are applied to the transmitting filter
whose impulse response is g(t). Then the output sequences, s(t) can be
obtained using the continuous and discrete convolution theorems as:

s (t ) =  a(t ) g (t −  )d
−
or
s (n) = kN=0 a (k ) g (n − k T b )
(2)

59
Digital Communication systems, 2nd Ed

s(t) is transmitted through the channel and it is modified by an impulse

response h(t) and then, the output signal is can be written in continuous time
as:
∞
𝑥0 (𝑡) = ∫−∞ 𝑠(𝑡)ℎ(𝑡 − 𝜏)𝑑𝜏 (3)
𝑜𝑟 𝑖𝑛 𝑑𝑖𝑠𝑐𝑟𝑒𝑡𝑒 𝑡𝑖𝑚𝑒 𝑎𝑠:
x0 (n) = k =0 s(k ) h(n − k T b )
M
(4)
A random white noise is added (AWGN channel model) to the useful signal
and the output signal of the transmission channel, x(t) is expressed as:
𝑥 (𝑡) = 𝑥0 (𝑡) + 𝑤(𝑡)
𝑜𝑟 𝑖𝑛 𝑑𝑖𝑠𝑐𝑟𝑒𝑡𝑒 𝑡𝑖𝑚𝑒 𝑎𝑠:
x(n) = kM=0 s (k ) h(n − k T b ) + w(n)
(5)
x(t) is the channel output of the noisy signal arriving at the receiver front
end.
A receive filter has an impulse response c(t) and the output signal, y(t) can
be expressed as:
y (n) = kM=0 x(k ) c(n − k T b )
(6)
y(t) is sampled synchronously with the transmitter clock signal that is
extracted from the receive filter output.
The reconstructed samples are compared to a threshold decision levels and
the output can be 1 or 0 according the signal level is >1 or <1.

The signal to noise ratio at the output of the receiving filter can be defined
as:
𝑬[𝒚(𝒕)𝟐 ]
𝑺𝑵𝑹 = (7)
𝑬[𝒘(𝒕)𝟐 ]

Example-1
Assume a causal system and the source PAM data, {1,-1} and the discrete
transmission channel is modeled by two coefficients, h(0)=0.5 and h(1)=1.
Assume that the channel is noise free.
• Determine the output signal of the channel, y(nT) using the discrete
linear convolution.
• Find out the signal output at n=T,2T as the transmitted data stream,
x(n) is {Start=1,1,-1,1,-1,-1,1,-1, End} and assume the starting time,
n=T, we consider T=1 for simplicity.

60
Digital Communication systems, 2nd Ed

Solution:

• The output signal x(n) can be expressed as:

1

𝑦(𝑛) = ∑ ℎ (𝑖)𝑥(𝑛 − 𝑖)
0
𝑦(𝑛) = ℎ(0)𝑥 (𝑛) + ℎ (1)𝑥(𝑛 − 1)
Substituting by the channel impulse response coefficients gives:
𝑦(𝑛) = 0.5𝑥 (𝑛) + 𝑥(𝑛 − 1)
• Assuming a two-dimensional observation vector,
XT(n)=[x(n),x(n-1)], initially x(0)=x(-1)=0.
.. n=1, x(n)=1, x(n-1)=0 as:
𝑥(𝑛) 1
[ ]=[ ]
𝑥(𝑛 − 1) 0
y(1)=0.5,
𝑥(𝑛) 1
--n=2, X (2)= [ ]=[ ]
𝑥(𝑛 − 1) 1

y(2)=0.5x1+1x1=1.5. etc at n=3,4, …….

Example-2

If the impulse response of the equivalent channel, heq(n) can be expressed as

h(n)={1,-1.5,2}, n=0,1,2 and the source signal a(n)={start:1,1,-1,1,1,-1,1}.
Consider the system is causal.
Find the output signal at discrete time, n=1,2,3,4

Solution:

The channel output can be calculated using the linear convolution as:
2

𝑦(𝑛) = ∑ ℎ𝑖 𝑎(𝑛 − 𝑖)
𝑖=0
𝑦(𝑛) = ℎ0 𝑎(𝑛) + ℎ1 𝑎(𝑛 − 1) + ℎ2 𝑎(𝑛 − 2)
𝑦(𝑛) = 𝑎(𝑛) − 1.5𝑎(𝑛 − 1) + 2𝑎(𝑛 − 2)
𝑛=0
𝑦(0) = 𝑎(0) − 1.5𝑎 (−1) + 2𝑎(−2)
𝑎(−1) = 𝑎(−2) = 0
Then, y(0)=a(0)=1
𝑛 = 1, 𝑦(1) = 𝑎(1) − 1.5𝑎(0) + 2𝑎(−1)
𝑦(1) = 1 − 1.5 = −0.5
𝑦(2) = 𝑎(2) − 1.5𝑎 (1) + 2𝑎(0) = −1 − 1.5 + 2 = −0.5
𝑦(3) = 𝑎 (3) − 1.5𝑎(2) + 2𝑎 (1) = 1 + 1.5 + 2 = 4.5
61
Digital Communication systems, 2nd Ed

𝑦(4) = 𝑎 (4) − 1.5𝑎 (3) + 2𝑎(2) = 1 − 1.5 − 2 = −2.5

6.2 Matched Filter

The matched filter is an optimal linear filter for maximizing the signal-to-
noise ratio (SNR) in the presence of additive stochastic noise.
The role of the matched filter is to optimize the signal to noise ratio at the
receiver output.

Fig.6.2 The receiver modeling with a matching filter

The input signal to the matched filter g(t) is corrupted with a white additive
noise and the resulting signal x(t) represents the input signal to the Matched
filter with an impulse response, h(t).
The output signal, y(t) in discrete form of the matched filter in terms of its
impulse response can be expressed by:
∞
𝑦(𝑡) = ∫−∞ 𝑥 (𝑡)ℎ (𝑡 − 𝜏)𝑑𝜏 (8)
𝑜𝑟 𝑖𝑛 𝑑𝑖𝑠𝑐𝑟𝑒𝑡𝑒 𝑓𝑜𝑟𝑚 𝑎𝑠:
M
y n =  hl x(n − l )
l =0 (9)
Substituting eq (8 ) in eq(7) yields:

| − H ( f )G ( f ) exp( j 2fT )df
=
N0  2
2
 |
−
H ( f ) | df
(10)
Given G(f), the input signal in frequency domain and the spectrum intensity
of the additive white noise, 𝑁0 we have to find the transfer function of the
matched filter , H(f) that maximizes the SNR, to obtain the optimal
performance of the matching filter.

6.3 Inter-symbol Interference (ISI)

Since the communication channel is dispersive and bandlimited, some

frequencies of the received pulse are delayed which causes pulse distortion

62
Digital Communication systems, 2nd Ed

and interfered with each other (change in shape and delay). This phenomenon
is called inter-symbol interference, ISI. The existence of the ISI in digital
communication receiver makes the detection capability of the transmitted data
is not efficiently and adds extra errors. Hence, the removal of the ISI
occurrence is essential to minimize the probability of error at the receiver
decision circuit and to receive the transmitted data correctly.

6.4 Equalizer
When the channel is not ideal, or when signaling is not Nyquist, there is ISI
signals at the receiver side. In time domain, the Equalizer is efficient in
removing the ISR signals. It equalizes the imperfection of the channel
frequency response.
In frequency domain, Equalizer flats the overall frequency responses of the
channel and receiving filters. Its transfer function should be equivalent to the
inverse of the channel impulse response transfer functions.
In practice, we equalize the equivalent transmission channel and the receiver
filter responses using an Equalizer as depicted in Fig.5.3

Fig.6.3 The bas band digital transmission with Linear Equalizer

6.4.1 The operation principle of the Equalizer

The role of the linear digital Equalizer in Fig.5.3 is to reduce the ISI happened
due to the imperfection of the analog of communication channel.
Assuming that the modulation and the demodulation have been done
efficiently. The analog communication channel is band limited and it is not
flat for all frequency band. Moreover, the channel gain is fluctuating and time
varying.
The input signal to the receiver depicted in Fig.5.1 is represented by:
x(t)=x0(t)+w(t) (11)
Hence:
X(f)=B(f)G(f)H(f) (12)

63
Digital Communication systems, 2nd Ed

G(f) is the transfer function of the transmitter filter and H(f) is the transfer
function of the data communication channel. B(f) is original source data in
frequency domain.
𝑌(𝑓 ) = 𝐵(𝑓 )𝐻(𝑓 )𝐺(𝑓)C(f) (13)
C(f) is the transfer function of the Receiver.
Eq. (13) can be written as:
𝑌(𝑓 ) = 𝐵(𝑓 )𝐻𝑒𝑞(𝑓 ) (14)
Heq(f) is the equivalent transfer function of the cascaded connection of the
transfer functions of the transmitter, the communication channel and the
receiver.
The equalizer module is implemented at the output of the receiver unit before
the decision circuit.
Then, the equalizer output can be defined as:

𝐷(𝑓 ) = 𝐵 (𝑓 )𝐻𝑒𝑞(𝑓 )𝐸(𝑓) (15)

If the equalizer operates properly, its output will be good estimate and
equivalent to the transmitted source data, B(f). Hence.

𝐵 (𝑓) = 𝐵 (𝑓)𝐻𝑒𝑞 (𝑓)𝐸(𝑓) and Then

𝐸 (𝑓) = 𝐻𝑒𝑞(𝑓)−1 (16)

It is apparent from eq. (15) that the transfer function of the optimally designed
equalizer is equivalent to the inverse of the feedback communication channel
from the source until the receiver and the output data is a good estimate
(approximate ) of the transmitted data. Thus, the decision errors are minimized
and the probability of error at the decision circuit is minimized.

Example-3

Assume that the digital communication channel is modeled by:

H(z)=𝒉𝟎 + 𝒛−𝟏 𝒉𝟏
- Find out the equivalent transfer function of the optimal digital equalizer
- Determine the difference equation that describes the optimal digital equalizer
in discrete time domain.

Solution:

Using eq.(15), the equivalent transfer function of the Equalizer is defined by:
E(z)=H(z)−1
64
Digital Communication systems, 2nd Ed

Then;
𝐳(𝐳) 𝟏
E(z)= =
𝐱(𝐳) 𝐡𝟎 +𝐳 −𝟏 𝐡𝟏
And the Direct form difference equation can be written as:
ℎ0 𝑧(𝑛) = 𝑥(𝑛) + ℎ1 𝑧(𝑛 − 1)
If h0 = 1, the equalizer output is defined by:
𝑧(𝑛) = 𝑥 (𝑛) + ℎ1 𝑧(𝑛 − 1)

Example-4

If the equivalent channel is given as:

1
𝐻 (𝑧 ) =
1 + 0.5𝑧 −1
Determine the equalizer transfer function and its implementation in
discrete domain.

Solution:

The equalizer transfer function is deduced as:

E(z)=H(z)−1 and E(z) = 1 + 0.5z −1
The equalizer Direct form implementation can be obtained by :

𝑧(𝑛) = 𝑥(𝑛) + 0.5𝑥(𝑛 − 1)

6.5 The adaptive equalizer

The equivalent transfer function is unknown and then, the design of the optimal
equalizer is impossible and impractical. Hence, the design of the adaptive
equalizer is introduced. Fig.5.4 explain a Simplified schem of the adaptive
equalaizer.

Fig. 6.4 The adaptive Equalizer

65
Digital Communication systems, 2nd Ed

The adaptive equalizer comprises of the following modules:

1- The adaptive filter module
2- The decision device
3- The adaptation algorithm
4- The training sequence generator in the learning mode.
The filtering section is also known as the convolution module. It is
considered a time varying filter as depicted in Fig.5.5.

Fig.6.5 The adaptive transversal filter

The adaptive transversal filter depicted in Fig. 5.5 consists of a tapped delay
line and its outputs are multiplied by the corresponding weight coefficients.
The Z-1 is considered unit delay which delay one period, T. The filter
comprises of N coefficients and the output , y(n) of the filter can be written
as:
N
y (n) =  wi x(n − i )
i =0 (17)
The filter output signal is obtained by convolving the input sequence with
the filter coefficients.
The error signal , e(n) is calculated as:
e(n)=d(n)-y(n) (18)
d(n) is known as the desired response.
The filter coefficients are updated such that the mean square of the error
signal is minimized. The equalizer has two modes of operation: the learning
and the permanent modes.
The transmitted sequences, a(n) is known to the receiver and it is considered
the desired response of the adaptive filter. The error signal is defined as:
e(n)=a(n)-y(n) (19)
The filter coefficients are updated using an adaptation algorithm such that
the mean square error is minimized using an adaptation algorithm.

66
Digital Communication systems, 2nd Ed

The filter coefficients converge to the optimal values after a certain

convergence time by the end of the learning period (learning mode).
At the end of the learning period, the output of the decision device can be
considered a good estimate of the transmitted sequence (desired response).
The error signal becomes:

e( n ) = a ( n ) − y ( n ) (20)
6.5.1 The equalizer modes of operation:
1- The learning (training) mode
2- Permanent mode
In the learning mode, the desired response, d(n) is equivalent to the same
transmitted sequences in the period of the learning or training mode.
The error signal in eq (18) can be written in terms of the filter coefficients
as:
N
e(n) = d (n) −  wix(n − i )
i =0
=
e( n ) = d ( n ) − W T X ( n ) (21)
Where, W and X(n) are defined as the coefficient and observation vectors
respectively:
The filter coefficients and observation signals vectors are defined
T
W w w = [ 0 , 1,......., N ]
w
T
respectively as
X (n) = [ x(n), x(n − 1),........, x(n − N )] (22)
The mean square of the error signal e(n) can be expressed by the ensemble
average of the squared error signal, e(n) as:
 n = E[e(n)
2
]
(23)
 n = E[(d (n) − W T X (n) ) ]
2

 n = E[(d (n) ) − 2d (n) X (n)

2 T T
W + W T X (n) X (n) W ]

 n = E[(d (n) )] − 2 PT W + W T RW
2
(24)
P and R refer the cross-correlation vector and the autocorrelation matrix of
the input observation signals.
Differentiating the MSE error with respect to the filter coefficients and
equating the gradient vector to zero yields:
67
Digital Communication systems, 2nd Ed

𝝏𝒆(𝒏 )𝟐 ]
𝜵= = −𝟐𝑷 + 𝟐𝑹𝑾 (25)
𝝏𝑾
Equating the gradient to zero results:
𝝏𝒆(𝒏 )𝟐 ]
• 𝜵= = −𝟐𝑷 + 𝟐𝑹𝑾Equating the gradient to zero results:
𝝏𝑾

 −1
W =R P (26)
Eq (26) is known Winer-Hoff orthogonal equation.
Example-5
Refer to the scheme denotes to an adaptive transversal equalizer in
Fig. 5.5 with two coefficients (impulse response) 𝑎𝑠:

WT= [𝑤0 , 𝑤1 ]
• Determine the output signal y(n) using the discrete convolution of the
input signal x(n) and the equalizer coefficients.
• Calculate the error signal e(n) and show how the output mean square
error E[(e(n)2] is minimized.

Solution:
𝑒(𝑛) = 𝑑(𝑛) − 𝑦(𝑛)
1

𝑦(𝑛) = ∑ 𝑤𝑖 𝑥(𝑛 − 𝑖) = 𝑤0 𝑥 (𝑛) + 𝑤1𝑥(𝑛 − 1)

0
𝜉 =𝐸[𝑒(𝑛 )2 ] = 𝐸[(𝑑(𝑛) − 𝑦(𝑛) )2 ]
Substituting by y(n) yields:
2
𝜉 = 𝐸 [𝑑(𝑛) ] − 2𝐸[𝑑(𝑛)(𝑤0𝑥 (𝑛) + 𝑤1𝑥(𝑛 − 1))]+E[(𝑤0𝑥 (𝑛) + 𝑤1𝑥(𝑛 −
1))2
=
𝐸 [𝑑(𝑛)2 ] − 2𝑤0 𝐸 [𝑑 (𝑛)𝑥 (𝑛)] − 2𝑤1 𝐸 [𝑑 (𝑛)𝑥 (𝑛 −
1)]+𝑤02 𝐸[𝑥(𝑛)2 ]+ 𝑤12 𝐸[𝑥(𝑛 − 1)2 +2𝑤0 𝑤1 𝐸[𝑥(𝑛)𝑥(𝑛 − 1)]
Differentiating the above function with respect to filter coefficients,
𝑤0 𝑎𝑛𝑑 𝑤1 results:
𝜕𝜉
= −2𝐸[𝑑(𝑛)𝑥 (𝑛)]+2𝑤0 𝐸[𝑥 (𝑛)2 ] + 2𝑤1 𝐸[𝑥 (𝑛)𝑥(𝑛 − 1)]
𝜕𝑤 0
And
𝜕𝜉
= −2𝐸[𝑑(𝑛)𝑥 (𝑛 − 1)]+2𝑤1 𝐸[𝑥(𝑛−1)2 ] + 2𝑤0 𝐸[𝑥(𝑛)𝑥(𝑛 − 1)]
𝜕 𝑤1
Where E[x(n)x(n-1)] is assumed equal to zero because x(n) and x(n-1)
are assumed statistically independent and are considered orthogonal.
Equating the derivatives to zero, yields:
−𝐸[𝑑(𝑛)𝑥 (𝑛)]+𝑤0 𝐸[𝑥 (𝑛)2 ] + 𝑤1 𝐸 [𝑥 (𝑛)𝑥(𝑛 − 1)] = 0

68
Digital Communication systems, 2nd Ed

−𝐸[𝑑(𝑛)𝑥 (𝑛 − 1)]+𝑤1 𝐸[𝑥(𝑛 − 1)2 ] + 𝑤0 𝐸[𝑥 (𝑛)𝑥(𝑛 − 1)] = 0

The above equation can be written in matrix notation as:
𝐸[𝑥 (𝑛)2 ] 𝐸 [𝑥 (𝑛)𝑥(𝑛 − 1)] 𝑤0
[ ] [𝑤 ]=
𝐸 [𝑥 (𝑛)𝑥(𝑛 − 1)] 𝐸[𝑥 (𝑛 − 1)2 ] 1
𝐸[𝑑(𝑛)𝑥(𝑛)]
⌈ ⌉.
𝐸[𝑑(𝑛)𝑥(𝑛 − 1)]
−1
𝑤0 𝐸[𝑥 (𝑛)2 ] 𝐸 [𝑥 (𝑛)𝑥 (𝑛 − 1)] 𝐸 [𝑑 (𝑛)𝑥 (𝑛)]
[𝑤 ] [ ] ⌈ ⌉
1 𝐸 [𝑥 (𝑛)𝑥 (𝑛 − 1)] 𝐸[𝑥 (𝑛 − 1)2 ] 𝐸 [𝑑 (𝑛)𝑥 (𝑛 − 1)]
Or
𝑤0 𝑟11 𝑟12 −1 𝑝1
[𝑤 ] = [𝑟 𝑟22 ] ⌈𝑝2 ⌉
1 21
Which is equivalent to Wiener-Hoff equation (26) as:

 −1
W =R P
It is obvious that in case of considering the transmitted stream of data
sequences is statistically independent and uncorrelated and zero mean. The
autocorrelation matrix becomes diagonal as:
𝐸[𝑥(𝑛)2 ] 0
[ ]
0 𝐸[𝑥(𝑛 − 1)2 ]
Then, the optimal weight coefficients are given as:
𝐸[𝑑 (𝑛)𝑥(𝑛)] −1
𝑤0 = = 𝑟11 𝑝1
𝐸[𝑥(𝑛)2 ]
𝐸[𝑑(𝑛)𝑥(𝑛−1)] −1
𝑤1 = 2 =𝑟22 𝑝2
𝐸[𝑥(𝑛−1) ]
The optimum filter coefficient vector can be obtained by multiplying the
inverse of the autocorrelation matrix by the cross-correlation vector of
desired signal, d(n) and the observation vector X(n).
The optimum filter coefficients are designed off-line as in eq (26), the filter
output signal y(t) becomes a good estimate of the transmitted sequence ,
d(n). Hence, the output of the decision device can be used instead of the
actual transmitted sequence, d(n)
In the permanent mode, d(n) is replaced by the output signal of the
decision device which is considered good estimate of the transmitted
sequences.

69
Digital Communication systems, 2nd Ed

6.5.2 The adaptation algorithm

The solution of the Winer-Hoff equation needs heavy processing due to off-
line block implementation of the inverse of the autocorrelation matrix; R and
computing the cross-correlation vector; P.
Generally, the impulse response of the communication channel is time
varying and hence, the equalizer coefficients should adapt and track the
channel variation.
Hence ; the block implementation of the fixed coefficient equalizer is not
efficient in this case and the adaptive solution of the winer-Hopf is
introduced.
6.5.3 The gradient adaptation algorithm
The filter coefficients are varying iteratively such that the mean square error
is minimized using the gradient techniques.
Then, the filter coefficients are updated as:
W n = W n−1 −  n−1 (27)
𝜇called the step size of the adaptation algorithm and controls the adaptation
speed and the stability of the adaptation algorithm.
It also, controls the adaptation noise and the misalignment of the filter
coefficients. Differentiating the mean square error in eq (21) with respect to
the filter coefficients and substituting in eq (22 )one obtains:
The gradient adaptation algorithm is defined as:
2
E[e(n) ]
W n = W n−1 −  W n −1
 e(n)
W n  W n−1 − 2e(n) W n −1 (28)
6.5.4 The Least mean square (LMS) adaptation algorithm
At the beginning of the adaptation process, we consider the ensemble of the
instantaneous square error is equivalent to the instantaneous error square
according to Widrow postulate. Hence, the filter coefficients are updated
according to LMS adaptation algorithm as:
W n = W n−1 + 2e(n) X (n − 1) (29)
wi,n = wi,n−1 + 2e(n) x(n − i)
i = 0,1,2,........, N (30)
6.6 Decision Feedback Equalizer (DFE)
The transversal equalizer has the draw back that it requires a large number of
the coefficients to model communication channel. The frequency response of

70
Digital Communication systems, 2nd Ed

the Transversal equalizer is poor which increases the probability of error.

Hence, the decision feedback equalizer is introduced to provide a channel
modeling with few numbers of coefficients and better frequency response. The
resulting detection precision is high, and the error rate is minimized.
The decision feedback equalizer Consists of feedforward, feedback, and
decision sections (nonlinear) as shown in Fig. (5.6). The output of the decision
device can be considered a good estimate of the transmitted sequences a(n). It
is fed back input to the backward section.

Fig.6.6 The Decision Feedback Equalizer

DFE outperforms the linear equalizer when the channel has severe amplitude
distortion or shape out off. The equalizer output is the sum of the outputs of
the forward and backward sections as:
And the filter coefficients are updated using the LMS adaptation algorithm
as:
wi,n = wi,n−1 + 2e(n) x(n − i)
i = 0,1,2,........, N

b j ,n = b j ,n−1 + 2e(n) d (n − j )
j = 1,2,........, M (31)
The output signal y(n) is applied to a decision device (circuit) to convert this
signal to the nearest estimate of the original transmitted data at the
transmitter, 𝑑 ^(n).
Then, 𝑑 ^(n)=±1 𝑖𝑓 𝑡ℎ𝑒 𝑜𝑟𝑖𝑔𝑖𝑛𝑎𝑙 𝑡𝑟𝑎𝑛𝑠𝑚𝑖𝑡𝑡𝑒𝑑 𝑑𝑎𝑡𝑎 𝑎(𝑛)𝑖𝑠 𝑃𝐴𝑀 = ±𝟏.
Assuming that the transmitted sequences are statistically independent and
zero mean. The autocorrelation matrix becomes diagonal which results the
optimal filter coefficients to be written as:

Example-6

Explain the operation mode of the adaptive equalizer

71
Digital Communication systems, 2nd Ed

Solution:
1- The learning (training) mode
In the learning period , the transmitted data is known a priory to the
receiver and the error signal is defined as in eq(18) and the filter
coefficients are updated using the LMS adaptation algorithm as in eq(29)
2- Permanent mode
In the permanent period , the transmitted data is not known a priory to the
receiver and the output of the decision circuit is considered a good
estimate of the transmitted sequences. The error signal is defined as in eq
(19) and the filter coefficients are updated using the LMS adaptation
algorithm as in eq(30).

Example-7
Assume an adaptive decision feed back equalizer with one coefficient in
forward section and one coefficient in the backward section with the
following specifications.
• The equalizer output y(n) in response to the input received signal,
x(n)= {Start:0.5,1.2, -1.2,-0.5,1.5].
• The desired response signal a(n) in the learning mode is known to the
equalizer and is defined as the frame data sequence as
a(n)={Start:1,1,-1,1,-1,-1,1, 1} .
• The LMS adaptation algorithm is used to update the equalizer
coefficients and 𝜇 = 0.01. y(-1)=y(0)=0 and the filter coefficients are
initially zero.
It is required to calculate the output signal of the equalized and its
corresponding output of the decision circuit for three iterations

Solutions:

The equalizer output y(n) can be expressed as:

𝑦(𝑛) = 𝑤0 (𝑛)𝑥(𝑛) + 𝑏1 (𝑛)𝑦(𝑛 − 1)
𝑒(𝑛) = 𝑎(𝑛) − 𝑦(𝑛)
𝑛=1
𝑦(1) = 𝑤0 (1)𝑥 (1) + 𝑏1 𝑦(0) = 0
𝑒(1) = 𝑎(1) − 𝑦(1)
=1−0=1

72
Digital Communication systems, 2nd Ed

The LMS adaptation algorithm:

𝑤0 (𝑛 + 1) = 𝑤0(𝑛) + 2𝜇𝑒(𝑛)𝑥(𝑛)
𝑏1 (𝑛 + 1) = 𝑏1 (𝑛) + 2𝜇𝑒(𝑛)𝑦(𝑛 − 1)
𝑤0(2) = 𝑤0 (1) + 2𝜇𝑒(1)𝑥(1)
𝑤0 (2) =0+2x0,01x1x0.5=0.01
𝑏1 (2) = 𝑏1(1) + 2𝜇𝑒(1)𝑦(0)
=0+0=0
…n=2,
𝑦(2) = 𝑤0 (2)𝑥 (2) + 𝑏1 (2)𝑦(1)=0.01x1.2+0=0.012
…e(2)=a(2)-y(2)=1-0.012=0.988
𝑤0(3) = 𝑤0 (2) + 2𝜇𝑒(2)𝑥(2)
𝑤0 (3)=0.01+0.02x0.988x1.2=0.023712
𝑏1 (3) = 𝑏1 (2) + 2𝜇𝑒(2)𝑦(1) = 0
…n=3
𝑦(3) = 𝑤0 (3)𝑥(3) + 𝑏1(3)𝑦(2)
= 0.023712x-1.2=-0.0284544
e(3) =a(3)-y(3)
=-1+0.0284544=-0.9715456

73
Digital Communication systems, 2nd Ed

Program-1

The Implementation of the adaptive Transversal Equalizer using the Matlab

program.

% Digital Realisation of adaptive Transversal Equalizer

c(1)=1.0;
c(2)=0.5;
c(3)=-0.2;
%b(1)=0.5;
%b(2)=-0.6
N=3;
M=2;
Ne=3;
mu=0.02;
fileID=fopen('C:\Users\eseli\OneDrive\Desktop\output.txt', 'w');
for j=1:10
y(j)=0.0;
xe(j)=0.0;
xf(j)=0.0;
ce(j)=0.0;
D(j)=0.0;
end
for i=1:1000
sumf=0.0;
sumb=0.0;
sume=0.0;
surmse=0.0;
x=rand(); % Data Source
if (x>=0.5) xf(1)=1;
else
xf(1)=-1;
end
for j=1:N %
sumf=sumf+c(j)*xf(j) ;
end
%for j=1:M %
% sumb=sumb+b(j)*y(j+1) ;
%end
yopc = sumf;
%+sumb;% output of communication channel
% Equalizer Module

74
Digital Communication systems, 2nd Ed

xe(1)=yopc;
%disp(yopc);
for j=1:Ne %
sume=sume+ce(j)*xe(j) ;
end
ye=sume;
if (ye > 0.0) D = 1;
else
D = -1;
end
%disp(ys)
% fprintf(fileID,'%12.1f \n',ye);
if (i <500) e=xf(1)-ye;
else
e=D-ye;
end
e2(i)=e*e;
if(i>800)
surmse=surmse+e2;
end
% fprintf(fileID,'%12.1f \n',e2(i));
% display(e);
%y1(i)=y(2);
% y2(i)=y(3);
% y3(i)=y(4);
for j=1:Ne
ce(j)=ce(j)+2*mu*e*xe(j);
end
%disp(ce(1));
%disp(ce(2));
%disp(ce(3));
for j=1:3
xf(4-j+1)=xf(4-j);
xe(4-j+1)=xe(4-j);
end
for j=1:3
y(4-j+1)=y(4-j);
end
end
%end
% s=abs(fft(y));
%for j=1:N
% fprintf(fileID,'%12.1f \n',c(j));

75
Digital Communication systems, 2nd Ed

% end
%for j=1:M
% fprintf(fileID,'%12.1f \n',b(j));
%end
RMSE=surmse/200;
for j=1:Ne
fprintf(fileID,'%12.1f \n',RMSE);
end
plot(e2);
title('Figure MSE vs N)');
%plot(ce(1));
%title('Figure ce(1) vs N)');
fclose(fileID);

The Results:

1- Forward coefficients after 1000 iterations

Ce1=0.9 , Ce2=-0.4 , Ce3= 0.3

Fig.1 The learning curve of the transversal equalizer(MSE versus iterations)

76
Digital Communication systems, 2nd Ed

Fig.2 The input data {1,-1} to the communication channel

The output data at the permanent mode (after the training mode)

Decision Circuit Data O/P

1.5

0.5

0
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39
-0.5

-1

-1.5

Fig.3 The binary output data {1,-1)

77
Digital Communication systems, 2nd Ed

Input Data
1.5

0.5

0
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39
-0.5

-1

-1.5

Fig.4 The binary input data {1,-1)

Program-2
The Implementation of the adaptive DFBE Equalizer using the Matlab
program

% Digital Realisation of adaptive DFBE equalizer

c(1)=1.0;
c(2)=0.5;
c(3)=-0.2;
%b(1)=0.5;
%b(2)=-0.6
N=3;
M=2;
Ne=3;
Me=4;
mu=0.01;
Le=500;
fileID=fopen('C:\Users\eseli\OneDrive\Desktop\output.txt', 'w');
for j=1:10
y(j)=0.0;
yd(j)=0.0;
zd(j)=0.0;
xe(j)=0.0;
xf(j)=0.0;
ce(j)=0.0;
be(j)=0.0;
D(j)=0.0;

78
Digital Communication systems, 2nd Ed

end
for i=1:1000
%*********************************
%***** Ttransmitter Module *******
%*********************************
sumf=0.0;
sumb=0.0;
sume=0.0;
surmse=0.0;
sumd=0.0;
x=rand(); % Data Source
if (x>=0.5) xf(1)=1;
else
xf(1)=-1;
end
z(i)=xf(1);% input data
%*********************************
%** Communication Channel Module**
%*********************************

for j=1:N %
sumf=sumf+c(j)*xf(j) ;
end
%for j=1:M %
% sumb=sumb+b(j)*y(j+1) ;
%end
yopc = sumf;
% output of communication channel
%*********************************
% Receiever and Equalizer Module**
%*********************************
xe(1)=yopc;
%disp(yopc);
for j=1:Ne %
sume=sume+ce(j)*xe(j) ;
end
if (i > Le)
for j=1:Me
sumd=sumd+be(j)*yd(j+1) ;
end
for j=1:Me
sumd=sumd+be(j)*xf(j+1) ;
end

79
Digital Communication systems, 2nd Ed

end
ye=sume+sumd;
yd(1)=ye;
%*********************************
%**Decision Circuit***************
%*********************************
if (ye > 0.0) D = 1;
else
D = -1;
end
if(i>500)
yd(1)=D;
end
if (i <500) e=xf(1)-ye;
else
e=D-ye;
end
e2(i)=e*e;
if(i>800)
surmse=surmse+e2;
end
% fprintf(fileID,'%12.1f \n',xf(1));
% fprintf(fileID,'%12.1f \n',e2(i));
% display(e);
%y1(i)=y(2);
% y2(i)=y(3);
% y3(i)=y(4);
%*********************************
%** Adaptation Algorithm **
%*********************************
for j=1:Ne
ce(j)=ce(j)+2*mu*e*xe(j);
end
if (i > Le)
for j=1:Me
be(j)=be(j)+2*mu*e*yd(j+1);
end
for j=1:Me
be(j)=be(j)+2*mu*e*zd(j+1);
end
end
%disp(ce(1));
%disp(ce(2));

80
Digital Communication systems, 2nd Ed

%disp(ce(3));
for j=1:3
xf(4-j+1)=xf(4-j);
xe(4-j+1)=xe(4-j);
end
for j=1:4
yd(5-j+1)=yd(5-j);
zd(5-j+1)=zd(5-j);
end
end
%*********************************
%** Performance Measure Module *
%*********************************
% s=abs(fft(y));
for j=1:Ne
fprintf(fileID,'%12.1f \n',ce(j));
end
for j=1:Me
fprintf(fileID,'%12.1f \n',be(j));
end
RMSE=surmse/200;
%for j=1:Ne
%fprintf(fileID,'%12.1f \n',RMSE);
%end
plot(e2);
title('Figure MSE vs iteration');
%plot(ce(1));
%title('Figure ce(1) vs N)');
fclose(fileID);

81
Digital Communication systems, 2nd Ed

Fig.5 Learning Curve of DFBE, step=0.1

Forward and Backward Coefficients

Forward C1=1.0
Forward C2= -0.4
Forward C3= 0.2
Backward B1= -0.1
Backward B2= 0.1
Backward B3= -0.1

82
Digital Communication systems, 2nd Ed

Exercise-6

Consider the scheme of the digital transmission system shown in Fig.5.2.

answer the followings
1- Assuming the analogue channel is modelled by a IIR, infinite impulse
response filter given by:
𝟏
H(z)=
𝟏−𝟎.𝟓𝒛−𝟏 +𝒛−𝟐
a- Determine the transfer function of the linear digital equalizer
b- Find the equalizer output response in discrete domain
c- State the type of the implemented equalizer
d- Calculate the output signal y(n) at n=1,2,3 if the input signal stream,
x(n)={Start:1,3,-1,-3,1,-1}
2- Assume that an adaptive transversal equalizer with two coefficients is
used to equalize the communication channel mentioned in Problem-1.
Choose, the step size, 𝜇 = 0.02 and the transmitted data sequences are
{Start:1,1,-1,1,-1,1,1}.
1- Find the equalizer output response
2- Determine the mean square error criterion in terms of the filter
coefficients
3- Find out the optimal filter coefficients
4- Using the LMS adaptation algorithm to update the filter coefficients.
(assuming the coefficients are initially zero)
5- Determine the decision device outputs
6- Compare between the Transversal and Decision feedback equalizers from
the following points:
a- The direct form implementation in time domain
b- The input/output responses
c- The operation principle in learning and permanent modes
d- The LMS adaptation algorithms used to update the filter coefficients

83
Digital Communication systems, 2nd Ed

Chapter Seven

Digital Modulation
7.1 Introduction

The digital modulation techniques are illustrated in this chapter.

The well-known types are:
 Digital-to-analog
Digital data and digital signals must be converted to analog signals for
digital transmission
1- The amplitude shift keying (ASK)
2- The frequency shift keying (FSK)
3- The phase shift keying (PSK)
4- The differential phase shift keying (DPSK)

• Analog-to-analog
Baseband signals must be modulated onto a higher-frequency carrier for
transmission. The basic techniques are amplitude modulation (AM), the
frequency modulation (FM) and the phase modulation (PM) as studied in
Analog Communication Systems.
• Analog data to digital signal
- Pulse code modulation (PCM)
- Delta modulation (DM)
These types of modulation are processed on the data source before applying
to the transmitter modulator and propagated through the communication
channel. In the Receiver unit, inverse processing is done to demodulate the
modulated signals and recover the data source. The measure that determines
how successful a receiver will be in interpreting an incoming signal?
84
Digital Communication systems, 2nd Ed

- Signal-to-noise ratio
- Data rate
- Bandwidth
• An increase in data rate increases bit error rate
• An increase in SNR decreases bit error rate
• An increase in bandwidth allows an increase in data rate
7.2 The ASK modulation
The amplitude shift keying modulation is changing of the amplitude of a carrier
signal according to the amplitude of a baseband data signal {0,1} as depicted in
Fig. 6.1. If the baseband (modulating) data incoming from the data source is “1”
bit, the ASK modulator passes the carrier signal during the duration of “1” bit.
On the other hand, if the incoming bit “0”, the output of the modulator is zero
signal. It means that the ASK modulator interrupts the amplitude of the carrier
signal according to the applied baseband data signal.

Fig.7.1 The ASK modulation wave forms

It is apparent that the output signals of the ASK modulator is interrupted carrier
signals as:
𝑥 (𝑡) = 𝐴0𝑠𝑖𝑛 (2𝜋𝑓𝑡 ) 𝑑𝑎𝑡𝑎 “1” (1)
The disadvantage of this method is that the missing time of signal is corrupted
with additive noise and interference signals and the receiver may consider them
as useful signal at the decision device.
 Used to transmit digital data over optical fiber
 Susceptible to sudden gain changes
 Inefficient modulation technique
7.3 The FSK modulation
This modulation technique considers all states of the source data signals “1” or
“0”. It assigns two frequencies, 𝑓1 𝑎𝑛𝑑 𝑓2 . If the incoming data signal is “1” , 𝑓1
is considered, but If the incoming data signal is “0” , 𝑓2 is considered as shown
in Fig.6.2.

85
Digital Communication systems, 2nd Ed

Fig. 6.2 The FSK modulation wave forms

This process of a Binary FSK transmitter is explained in Fig.6.3

Fig.6.3 A Binary FSK modulator

If a binary data “1” is applied, it will be multiplied by the sinusoidal signal,
𝑐𝑜𝑠(2𝜋𝑓1 𝑡) and passes to the FSK output, S1(t) = 𝑐𝑜𝑠(2𝜋𝑓1𝑡). The inverter in the
lower branch converts “1” to “0” and the second frequency, 𝑐𝑜𝑠(2𝜋𝑓2 𝑡) does
not pass. Hence, S1(t) = 𝑐𝑜𝑠(2𝜋𝑓1 𝑡) if the input data is “1’’ and S1(t)
= 𝑐𝑜𝑠(2𝜋𝑓2 𝑡) if the input data is “0”.
 Less susceptible to error than ASK
 Used for high-frequency (3 to 30 MHz) radio transmission

7.3.1 Multiple Frequencies (MFSK)

 More than two frequencies are used in FSK
 More bandwidth efficient
 Used for frequency hopping in spread spectrum
si (t ) = A cos 2f i t 1  i  M (2)
f i = fc + (2i − 1 − M ) fd
M = number of different signal elements = 2 L
L = number of bits per signal element
Example7-1
With fc=250 kHz, fd=25 kHz and M=8 (L=3 bits),

86
Digital Communication systems, 2nd Ed

f i = fc + (2i − 1 − M ) fd
we have the following frequency assignments for each of the 8 possible 3-bits
data combinations:
f1= 75 kHz 000 f2=125 kHz 001 f3=175 kHz 010
f4=225 kHz 011 f5=275 kHz 100 f6=325kHz 101
f7=375 kHz 110 f8=425 kHz 111

7.4 The PSK modulation

The PSK modulation senses the transition from “0’’ to “1” or from “1” to “0” in
the baseband binary input data, x(t). The phase of the carrier signal is changed
by 𝜋 if a transition is detected and no change if there is no transition. The PSK
modulation is shown in Fig. 6.4.

Fig. 7.4 The PSK modulation Wave form

7.5 Differential Phase Shift Keying, DPSK
DPSK principle
• to send symbol 0, we advance the phase of the current signal waveform
by 180 degrees,
• to send symbol 1, we leave the phase of the current signal waveform
unchanged.
Generation of DPSK:
The differential encoding process at the transmitter input starts with an
arbitrary first bit, serving as reference.
DPSK is a method of BPSK, where there is no reference phase signal. Here,
the signal which is transmitted is used as a reference signal. The DPSK
modulator diagram is shown below. This modulation encodes two separate
signals namely the carrier signal as well as the modulating signal. The phase
shift of each signal is 180°.

87
Digital Communication systems, 2nd Ed

Fig.7.6 DPSK Modulation

In the above figure, the serial input data can be applied to the XNOR gate &
the o/p of the logic gate is fed back again to the input via 1-bit delay. Both the
carrier signal as well as XNOR gate output is applied to the balanced
modulator so that the modulated signal of DPSK can be generated.
DPSK Demodulation
In this demodulator, both the previous bit and the reversed bit are compared
with each other. The DPSK demodulator block diagram is shown below. From
the above block diagram, it is clear that the DPSK signal is applied to a
balanced modulator using a 1-bit delay input.

Fig.7.7 DPSK Demodulation

That signal is ready to release in the direction of lower frequencies using a low
pass filter. After that, it is transmitted toward a shaper circuit for improving
the unique binary data like the output. Here shaper circuit is a Schmitt trigger
or comparator circuit.
DPSK Advantages and Disadvantages
The advantages of DPSK include the following.
• This modulation doesn’t need the carrier signals at the end of the
receiver circuit. Therefore, compound circuits are not required.
• The BW of DPSK requirement is low evaluated to BPSK modulation.
• Non-consistent receivers are simple and inexpensive to construct,
therefore extensively used in wireless communication.
The disadvantages of DPSK include the following.
• The bit error rate or chance of error is high in DPSK contrast to BPSK.
• The interference of noise in DPSK is more.
• This modulation employs two consecutive bits intended for its response.
Thus, error in primary bit makes error within a subsequent bit as well as
consecutively error spreads.
88
Digital Communication systems, 2nd Ed

The applications of DPSK include the following.

• The applications of differential phase-shift keying mainly include wireless
communications like RFID, WLANs, and Bluetooth. The famous
application among them is Bluetooth wherever alternatives of DPSK has
been used like 8-DPSK, and π/4 – DQPSK modulation
• Thus, DPSK is a general type of phase modulation and it is used to transmit
data through the carrier wave by changing its phase. This type of PSK
removes the need for a consistent reference signal at the end of the receiver
by adding two fundamental operations at the end of the transmitter. Here
is a question for you, what is the difference between DPSK and BPSK?
7.6 Quadrature Phase Shift Keying, QPSK
Quadrature Phase Shift Keying (QPSK) can be interpreted as two
independent BPSK systems (one on the I-channel and one on Q-channel),
and thus the same performance but twice the bandwidth (spectrum)
efficiency. Fig.6.8 illustrates the QPSK Constellation Diagram.

Fig.7.8 the QPSK Constellation Diagram.

• Quadrature Phase Shift Keying has twice the bandwidth efficiency of
BPSK since 2 bits are transmitted in a single modulation symbol
• The phase of the carrier takes on 1 of 4 equally spaced values, such as
where each value of phase corresponds to a unique pair of message bits.
The QPSK signal for this set of symbol states may be :

(3)
• Like BPSK, QPSK can also be differentially encoded to allow non-
coherent detection.
Four-level PSK (QPSK) as shown in Fig.7.9
 Each element represents two bits
 Phase shift in multiples of /4

89
Digital Communication systems, 2nd Ed

Fig. 7.9 QPSK constellation

 OQPSK: Introducing a time-delay

 Phase change less than /2
 Therefore, less interference

90
Digital Communication systems, 2nd Ed

Fig. 7.10 QPSK & OQPSK Diagram

7.7 QAM is a combination of ASK and PSK
Two different signals sent simultaneously on the same carrier frequency.
s(t ) = d1 (t )cos 2f ct + d2 (t )sin 2f ct (4)

Fig. 7.11 Quadrature Amplitude Modulation

QAM as depicted in Fig.7.11 is a popular analog signaling technique that is
used in some wireless standards. This modulation technique is a combination of
ASK and PSK. QAM can also be considered a logical extension of QPSK.
QAM takes advantage of the fact that it is possible to send two different signals
simultaneously on the same carrier frequency, by using two copies of the carrier
frequency, one shifted by 900 with respect to the other.

91
Digital Communication systems, 2nd Ed

Figure 6.10 shows the QAM modulation scheme in general terms. The input is a
stream of binary digits arriving at a rate of R bps. This stream is converted into
two separate bit streams of R/2 bps each, by taking alternate bits for the two
streams. In the diagram, the upper stream is ASK modulated on a carrier of
frequency Ie by multiplying the bit stream by the carrier. Thus, a binary zero is
represented by the absence of the carrier wave and a binary one is represented
by the presence of the carrier wave at a constant amplitude. This same carrier
wave is shifted by 900 and used for ASK modulation of the lower binary
stream. The two modulated signals are then added together and transmitted. The
transmitted signal can be expressed as follows:
If two-level ASK is used, then each of the two streams can be in one of two
states and the combined stream can be in one of 4 = 2 X 2 states. This is
essentially QPSK.
7.8 QPSK Demodulation
Is
process of removing the carrier signal to obtain the original signal
waveform
• Detection – extracts the symbols from the waveform
◼ Coherent detection
◼ Non-coherent detection
• Coherent Detection
- An estimate of the channel phase and attenuation is recovered. It is
then possible to reproduce the transmitted signal and demodulate.
- Requires a replica carrier wave of the same frequency and phase at the
receiver.
- Also known as synchronous detection (I.e. carrier recovery)
• Applicable to
- Phase Shift Keying (PSK)
- Frequency Shift Keying (FSK)
- Amplitude Shift Keying (ASK)
• Non-Coherent Detection
Requires no reference wave; does not exploit phase reference information
(envelope detection)
- Differential Phase Shift Keying (DPSK)
- Frequency Shift Keying (FSK)
- Amplitude Shift Keying (ASK)
- Non coherent detection is less complex than coherent detection
(easier to implement) but has worse performance.
7.9 Summary
This chapter has presented the different techniques of the digital modulation
which are widely used in digital communication systems. It highlights the
analysis and the implementation of ASK, FSK, MFSK, PSK and QPSK types. It

92
Digital Communication systems, 2nd Ed

also, presents short hints about the detection and the demodulation types which
are used in digital communication systems.

93
Digital Communication systems, 2nd Ed

Exercise-7

1- Explain the main processes to send the sequences {1,1,0,0,1,0} using the
FSK, PSK and DPSK principles
2- Draw the output wave forms in each case for the FSK and PSK
modulations
3- Draw the basic schemes of modulation and explain the basic operations to
send the pattern {11010} and state the output transmitted data ( consider
the reference data=’’1’’
4- With fc=250 kHz, fd=25 kHz and M=16 (L=4 bits),

f i = fc + (2i − 1 − M ) fd
- Determine the different FSK frequencies to represent a binary word
of 4 bits.
- Write down the Truth tables to express the 16 states and the
corresponding carrier frequency.
5- Explain the operation principle of QPSK modulation and show its
advantage and disadvantage
6- Explain how QPSK can transmit a constellation of the pattern {10,11}
7- Compare between the coherent and non-coherent detection of the PSK
and FSK modulated signals

94
Digital Communication systems, 2nd Ed

Chapter Eight

Spread Spectrum in Communication Systems

8.1 Introduction
The idea of spread spectrum (SS) is to spread the information signal over a
wider bandwidth to make jamming and interception more difficult.
Developed initially for military and intelligence and it becomes important
form of communications
Two types of SS:
• Frequency Hopping (FHSS)
• Direct Sequence (DSSS)

An increasingly important form of communications is known as spread

spectrum. It can be used to transmit either analog or digital data, using an
analog signal. The spread spectrum technique was developed initially for
military and intelligence requirements.
The first type of spread spectrum developed is known as frequency hopping.
A more recent type of spread spectrum is direct sequence.
8.2 Spread Spectrum Technique
Fig. 8.1 The Spread Spectrum Model in Digital Communication system is
shown in Fig. 8.1

Input is fed into a channel encoder and produces analog signal with narrow
bandwidth Signal is further modulated using sequence of digits (spreading
code) aka spreading code or spreading sequence or chip. Generated by pseudo-
noise, or pseudo-random number generator. Effect of modulation is to increase
bandwidth of signal to be transmitted
On receiving end, digit sequence (spreading code) is used to demodulate the
spread spectrum signal. Signal is fed into a channel decoder to recover data
Input is fed into a channel encoder that produces an analog signal with a
relatively narrow bandwidth around some centre frequency. This signal is
further modulated using a sequence of digits known as a spreading code or Sf
reading sequence. Typically, but not always, the spreading code is generated
by a pseudonoise, or pseudorandom number, generator. The effect of this
modulaion is to increase significantly the bandwidth (spread the spectrum) of
the signal to be transmitted.
On the receiving end, the same digit sequence is used to demodulate the
spread spectrum signal. Finally, the signal is fed into a channel decoder to
recover the data.

95
 Digital Communication systems, 2nd Ed

Input Channel Modulato- De- Channel Output D

Data Channel
encoder modulator decoder

Spreading Spreading
Code Code

Pseudonoise Pseudonoise
generator generator

Fig. 8.1 The Spread Spectrum Model in Digital Communication

system

8.2.1 waste of spectrum:

We can gain immunity from various kinds of noise and multipath distortion.
The earliest applications of spread spectrum were military: " where it was
used for its immunity to jamming. It can also be used for hiding and
encrypting signals. Only a recipient who knows the spreading code can
recover the encoded information. Several users can independently use the
same higher bandwidth wit 1 very little interference. This property is used in
cellular telephony applications.
8.3 Frequency Hoping Spread Spectrum (FHSS)
Signal is broadcast over seemingly random series of radio frequencies.
Several channels allocated for the FH signal (2^k channels). Width of each
channel corresponds to bandwidth of input signal. Channel sequence
dictated by spreading code. Signal hops from frequency to frequency at fixed
intervals. Transmitter operates in one channel at a time. Bits are transmitted
using some encoding scheme.
At each successive interval, a new carrier frequency is selected. Receiver,
hopping between frequencies in synchronization with transmitter, picks up
message. Both transmitter and receiver use the same spreading code.

96
Digital Communication systems, 2nd Ed

Advantages
• Eavesdroppers hear only unintelligible blips
• Attempts to jam signal on one frequency succeed only at
knocking out a few bits

chipping

Fig. 8.2 Frequency Hopping Example

8.4 Direct Sequence Spread Spectrum (DSSS)

 Each bit in original signal is represented by multiple bits in the

transmitted signal as explained in Fig.8.3
 Spreading code spreads signal across a wider frequency band
 Spread is in direct proportion to number of bits used

97
Digital Communication systems, 2nd Ed

 One technique combines digital information stream with the spreading

code bit stream using XOR

Fig.8.3 DSSS signals

For direct sequence spread spectrum (DSSS), each bit in the original signal is
represented by multiple bits in the transmitted signal, using a spreading code. The
spreading code spreads the signal across a wider frequency band in din :ct
proportion to the number of bits used. Therefore, a 10-bit spreading code spread
s the signal across a frequency band that is 10 times greater than a 1-bit spreading
code. One technique for direct sequence spread spectrum is to combine the digital
information stream with the spreading code bit stream using an e {elusive-OR
(XOR).
The XOR obeys the following rules:
O XOR O=O
O XOR 1=1
1 XOR O=1
1 XOR 1=0
Figure 8.4 shows an example. Note that an information bit of 01.e inverts the
spreading code bits in the combination, while an information bit of zeros causes
the spreading code bits to be transmitted without inversion. The combination
stream has the data rate of the original spreading code sequence, so it has a wide]
bandwidth than the
information stream. In this example, the spreading code bit StH am is clocked at
four times the information rate.

98
Digital Communication systems, 2nd Ed

A= C XOR B
Fig.8.4 Example of DSSS

Fig. 8.5 DSSS System

Hybrid system: DS/FFH

One data bit is divided over frequency-hop channels (carrier frequencies).

In each frequency-hop channel one complete PN-code of length is added to
the data signal.

8.5 Code Division Multi-Access (CDMA)

CDMA is a multiplexing technique used with spread spectrum.' [be scheme works
in the following manner. We start with a data signal with rate D, which we call
the bit data rate. We break each bit into k chips according to a fixed pattern that

99
Digital Communication systems, 2nd Ed

is specific to each user, called the user's code. The new channel has a chip data
rate of kD chips per second.
Start with data signal rate D Called bit data rate. Break each bit into k chips
according to fixed pattern specific to each user : User’s code
New channel has chip data rate kD chips per second E.g. k=6, three users
(A,B,C) communicating with base receiver R
Code for A = <1,-1,-1,1,-1,1>
Code for B = <1,1,-1,-1,1,1>
Code for C = <1,1,-1,1,1,-1>
As an illustration we consider a simple example with k = 6. It is simplest to
characterize a code as a sequence of I’s and -I’s. Figure 8.6 shows the codes
for three users, A, B, and C, each of which is communicating with the same
base station receiver, R. Thus, the code for user A is CA = <1, -1, -1,1, -1,1>.
Similarly, user B has code CB = <1,1,1, -1, 1, 1>, and user C has CC = <1,1,
-1,1, 1, -1>.
We now consider the case of user A communicating with the base station. The
base station is assumed to know A's code. For simplicity, we assume that
communication is already synchronized so that the base station knows when
to look for codes. If A wants to send a 1 bit, A transmits its code as a (chip
pattern <1, -1, -1,1, -1,1>. If a 0 bit is to be sent, A transmits the complement
(Is and -Is reversed) of its code, <-1,1,1, -1, 1, -1>. At the base station, the
receiver decodes the chip patterns. In our simple version, if the receiver
receives a chip pattern d = < 𝑑1, 𝑑2, 𝑑3, 𝑑4, 𝑑5, 𝑑6 > and the receiver is
seeking to communicate with a user u so that it has at hand u's code, <
𝑐1, 𝑐2, 𝑐3, 𝑐4, 𝑐5, 𝑐6 >, the receiver performs electronically the following
decoding function:
Send) = (𝑑1𝑥𝑐1) +(𝑑2 𝑥𝑐2) + (𝑑3𝑥𝑐3) + (𝑑4𝑥𝑐4) + (𝑑5𝑥𝑐5) (𝑑6 𝑥𝑐6)
 User A code = <1, –1, –1, 1, –1, 1>
 To send a 1 bit = <1, –1, –1, 1, –1, 1>
 To send a 0 bit = <–1, 1, 1, –1, 1, –1>
 User B code = <1, 1, –1, – 1, 1, 1>
 To send a 1 bit = <1, 1, –1, –1, 1, 1>
 Receiver receiving with A’s code
(A’s code) x (received chip pattern
 User A ‘1’ bit: 6 → 1
 User A ‘0’ bit: -6 → 0
 User B ‘1’ bit: 0 → unwanted signal ignored
If A sends a 1 bit that corresponds to d = <1, -1,-1, 1,-1, ->
we get: SA=[1x1+-1x-1+-1x-1+1x1+-1x-1+1x1=6 ->”1”
If A sends a 0 bit that corresponds to d = <-1,1,1, -1,1, ->,
we get: SA=
(-l, 1, 1, -1,1, -1) = [-1 X 1] + [1 x (-1)] + [1 x (-1)] + [(-1) x 1]
+ [1 x (-1)] + [1 x (-1)] + [(-1) x( 1] = -6 ->”0”
100
Digital Communication systems, 2nd Ed

Please note that it is always the case that -6 < SA(d) < 6 no matter what
sequence of -Is and Is comprise d, and that the only values of d resulting in
the extreme values of 6 and -6 are A's code and its complement, respectively.
So if SA produces a +6, we say that we have received a 1 bit from A; if SA
produces a -6, we say that we have received a 0 bit from user A; otherwise,
we assure that someone else is sending information or there is an error. So
why go through all this? The
reason becomes clear if we see what happens if user B is sending and we try
to receive it with SA, that is, we are decoding with the wrong code, A's. If B
sends a 1 bit, then d = <1,1, -1, -1, 1, 1>. Then SA(l,l,-l,-l,l,l) = [1 x 1] + [1 x
(-1)] + [(-1) x (-1): + [(-1) xl] + [1 x (-1)] + [1 Xl] = 0

1 1 -1 1 1 -1
Fig.8.6 CDMA Example
Code A =1 -1 -1 1 -1 1
Code B= 1 1 -1 -1 1 1
Code C= 1 1 -1 1 1 -1
Fig. 8.6 CDMA codes and transmitted bit streams

As was mentioned, the spreading sequence, c(t), is a sequence of binary digits

shared by transmitter and receiver. Spreading consists of multiplying, (XOR) the
input data by the spreading sequence, where the bit rate of the spreading sequence
is higher than that of the input data. When the signal is received, the spreading is
removed by multiplying with the same spreading code, exactly synchronized with
the received signal.
8.6 CDMA: Decoding
• If k = 6 and code is a sequence of 1s and -1s
o For a ‘1’ bit, A sends code as chip pattern
▪ <c1, c2, c3, c4, c5, c6>

101
Digital Communication systems, 2nd Ed

o For a ‘0’ bit, A sends complement of code

▪ <-c1, -c2, -c3, -c4, -c5, -c6>
• Receiver knows sender’s code and performs electronic decode function
Su (d ) = d1 c1 + d 2  c2 + d 3  c3 + d 4  c4 + d 5  c5 + d 6  c6
o <d1, d2, d3, d4, d5, d6> = received chip pattern
o <c1, c2, c3, c4, c5, c6> = sender’s code
CDMA Example
User A code = <1, –1, –1, 1, –1, 1>
To send a 1 bit = <1, –1, –1, 1, –1, 1>
To send a 0 bit = <–1, 1, 1, –1, 1, –1>
User B code = <1, 1, –1, – 1, 1, 1>
To send a 1 bit = <1, 1, –1, –1, 1, 1>
Receiver receiving with A’s code
(A’s code) x (received chip pattern)
User A ‘1’ bit: 6 → 1
User A ‘0’ bit: -6 → 0

User B ‘1’ bit: 0 → unwanted signal ignored

8.7 PN Sequences
PN generator produces periodic sequence that appears to be random. PN
Sequences Generated by an algorithm using initial seed Sequence isn’t
statistically random but will pass many test of randomness. Sequences referred
to as pseudorandom numbers or pseudonoise sequences.
Unless algorithm and seed are known, the sequence is impractical to predict
An ideal spreading sequence would be a random sequence of binary ones and
zeros. However, because it is required that transmitter and receiver must have
a copy of the random bit stream, a predictable way is needed to generate the
same bit stream at transmitter and receiver and yet retain the desirable
properties of a 1andom bit stream. This requirement is met by a PN generator.
A PN generator will produce a periodic sequence that eventually repeats but
that appears to be random. The period of a sequence is the length of the
sequence before it starts repeating. PN sequences are generated by an
algorithm using some initially called the seed. The algorithm is deterministic
and therefore produces sequences (If numbers that are not statistically random.
However, if the algorithm is good, the resulting sequences will pass many
reasonable tests of randomness. Such number s are often referred to as
pseudorandom numbers, or pseudonoise sequences. An important point is
that unless you know the algorithm and the seed, it is impractical to predict
the sequence. Hence, only a receiver that shares this information with a
transmitter will be able to decode the signal successfully.

102
Digital Communication systems, 2nd Ed

PN sequences find a number of uses in computers and communications, and

the principals involved are well developed. We begin with a general
description of desirable properties of PNs and then look at the generation
method tYl1ically used for spread spectrum applications.
PN Properties Two important properties for PNs are randomness and
unpredictability. Traditionally, the concern in the generation of a sequence of
illegally random numbers has been that the sequence of numbers be random
in some well-defined statistical sense. The following two criteria are used to
validate that a sequence of numbers is random:
Two important properties of PN:
Randomness: the sequence of numbers be random
Unpredictability.
Two criteria used to validate that a sequence is random:
Uniform distribution of numbers (0/1): the frequency of occurrence
of 1 should be approximately the same as occurrence of 0.
Independence: no value in the sequence can be inferred from the
others.
Uniform distribution: The distribution of numbers in the sequence should be
uniform; that is, the frequency of occurrence of each of the numbers should
be approximately the same. For a stream of binary digits, we need to expand
on this definition because we are dealing with only 2 numbers (0 and 1).
Independence: No one value in the sequence can be inferred from the others.
Although there are well-defined tests for determining that a sequence of
numbers matches a particular distribution, such as the uniform distribution,
there is no such test to "prove" independence. Rather, a number of tests can be
applied to demonstrate that a sequence does not exhibit independence. The
general strategy is to apply a number of such tests until the confidence that
independent exists is sufficiently strong.

8.8 Noise in Communication Systems

The term noise refers to unwanted electrical signals that are always present in
electrical systems; e.g spark-plug ignition noise, switching transients, and
other radiating electromagnetic signals.
Can describe thermal noise as a zero-mean Gaussian random process.
A Gaussian process n(t) is a random function whose amplitude at any arbitrary
time t is statistically characterized by the Gaussian probability density
function

1  1 n  
2

p ( n) = exp  −   
 2  2     (1)

103
Digital Communication systems, 2nd Ed

The normalized or standardized Gaussian density function of a zero-mean

process is obtained by assuming unit variance.
8.2 White Noise
The primary spectral characteristic of thermal noise is that its power spectral
density is the same for all frequencies of interest in most communication
systems
Power spectral density Gn(f )
N0
Gn ( f ) = watts / hertz
2 (2)
Autocorrelation function of white noise is

N0
Rn ( ) =  {Gn ( f )} =  ( )
−1

2 (3)
The average power Pn of white noise is infinite

N0
p ( n) =
−
 2
df = 
(4)

The effect on the detection process of a channel with additive white gaussian
noise (AWGN) is that the noise affects each transmitted symbol
independently. Such a channel is called a memoryless channel.
The term “additive” means that the noise is simply superimposed or added to
the signal
Exercise-8
104
Digital Communication systems, 2nd Ed

1- Explain the importance of the SSS techniques in digital communication

systems
2- Compare between the FHSSS and DSSS types from point of views:
- Schematic Diagrams
- Principle of operation
3- Explain the main concepts of the CDMA techniques operation and code
patterns
4- A 3 users with code sequences as:
A={1,1,-1,1,-1}
B={1,-1,-1,1,1}
C={-1,-1,1,1,-1}
If a “1” and “0” bits are wanted to send to the destination through the
communication channels.
The Receiver at the destination received the pattern {1,-1,-1,1,1} and {1,-
1,1,1,-1}
Explain how the receiver can identify the sent user and check the
correctness of the received data.

105
Digital Communication systems, 2nd Ed

References
[1] Dennis Derickson and Marcus Müller, “Fundamentals of Digital
Communications Systems”, Mar 14, 2008
[2] K. Sam Shanmugam,” Digital and Analog Communication Systems”, Wiley,
January 2012.
[3] Heath “Digital Communication”, April 2004, North Carolina State
University
[4] Fuyun Ling, “An Overview of Digital Communication Systems”
DOI: https://doi.org/10.1017/9781316335444.003, Publisher: Cambridge
University Press.
[5] Simon Haykin, Michael Moher, “Communication Systems”, 5th Edition,
ISBN: 978-0-471-69790-9 March 2009 448 Pages.
[6] Simon Haykin and Michael Moher”An Introduction to Analog and Digital
Communications” 2nd Edition,
[7] B.P. Lathi and Zhi Ding ,” Modern Digital and Analog Communication”,
The Oxford Series in Electrical and Computer Engineering 5th Edition.
[8] Simon Haykin,” Digital Communication Systems”,
https://www.amazon.com/Digital-Communication-Systems-Simon-
Haykin/dp/0471647357

106
Digital Communication systems, 2nd Ed

Appendices

Project-1

Name : DELTA Modulation Coding Technique

Aim of the project:
The aim of this Matlab project is to enhance the student’s understanding of,
and ability to simulate, digital communication systems. It should also help the
student better understand various design aspects of communication systems .
The project
This project aims to learn the student how to design and implement the
DELTA modulation Scheme shown in Figure 3 using the MATLab
Programming
Procedures:
1. Assume the analog input signal is represented by:
𝑥(𝑡) = 5 sin(2𝜋𝑓𝑡 )
2. Signal frequency=2kHz
3. Sampling frequency 10kHz and choose 1000 samples.
4. Write the MATLab Codes to implement the different modules of the Delta
modulation scheme
5. Display and Plot the digitized transmitted output wave forms.
6- Write a report of at least 5 Page in word including the MATlab codes and
the displayed results and Scheme.
7- Illustrate your Comments
8- Deliver To Prof: Elsayed Soleit directly, the project at the week 8 after
MID term.

ID:…………………………………..
Name:………………………………..

107
Digital Communication systems, 2nd Ed

Project-2

Objective
The aim of this Matlab project is to enhance the student’s understanding
of,
and ability to simulate, digital communication systems.
It should also help the student better understand various design aspects of
communication systems.

Project
The aim of this project is to illustrate the effect of dispersive channel.

Procedures:

1- Design and Implement the FSK and PSK digital modulation scheme.
2- The digital data to be transmitted pattern is “110100101”
3- Take the base frequency carrier 10kHz
4- Choose the binary PSK and QPSK
5- Write the MATlab programing
6- Write the Final report including the MATlab codes, the results and the
output wave forms in Word.
7- Deliver the report to Prof: Elsayed Soleit Directly

ID:…………………………………..
Name:………………………………..

108