EC5.

203 Communication Theory (3-1-0-4):

Introductory Class

Instructor: Dr. Sachin Chaudhari

Email: [email protected]

Jan. 06, 2025

and
Jan. 09, 2025
Introduction and Motivation
 What is Communication?
• Communication is defined as transmission of
information from source to destination via a transmission
medium
Channel
Source Destination

• Examples:
– AM and FM Radios
– TV
– Phone call/Whatsapp Message/Internet
– Accessing Intranet over Ethernet
– Microprocessor and peripherals
Examples: Microprocessor and Peripherals
Communication: A slightly different perspective

Channel
Source Destination

• Information transfer can be across space or time

– Separated in space (telephony, web browsing,…)
– Separated in time ??
• Storage of information - CDs, DVDs, hard drives, cloud
 Popular Communication Systems
• Cellular Communication (2G/3G/4G/5G)
• LAN
– Ethernet
– WLAN
• Broadcasting: radio and TV
• Bluetooth
• Sensor networks
• Satellite communication
• Many more……..
 Why Wireless Communication?
• Wired communication will always be better than wireless
• Wireless provides “Anytime Anywhere communication”
• A necessity in today’s world
• Mobile phone has become an addiction
– Teenagers and technology: 'I'd rather give up my kidney than my
phone’
– Digital communication is not just prevalent in teenagers' lives. It
IS teenagers' lives
– https://www.theguardian.com/lifeandstyle/2010/jul/16/teenagers-mobiles-facebook-social-
networking
 Key Steps in Communication Link
• Insertion of information into a signal, termed the
transmitted signal, compatible with the physical medium
of interest.
• Propagation of the signal through the physical medium
(termed the channel) in space or time
• Extraction of information from the signal (termed the
received signal) obtained after propagation through the
medium.

Source Transmitter Channel Receiver Destination

message signals Estimate of message signals

Distortion and Noise
Questions?
 CHANNEL
The Main Resource and
the Main Challenge!
 Types of Channel
Classification based on Medium
• Guided channels
– Copper wire (twisted pair)
– Coaxial cable
– Optical fibre
– microwave guides

• Unguided channels
– wireless channel
– underwater acoustic channel
 Wired Medium

Unshielded Twisted Pair

Fiber Optics
https://en.wikipedia.org/wiki/Twisted_pair Coaxial Cable
https://en.wikipedia.org/wiki/Coaxial_cable http://www.slideshare.net/subrata11/
optical-fiber-55382806

Waveguide Waveguide
http://www.megaind.com/ https://en.wikipedia.org/wiki/Waveguide
 Channel: the challenges!
• Limited Spectrum
– Limited wireless spectrum: High frequencies high attenuation, at
low frequencies: large antenna sizes and low bandwidth
– Wired spectrum: the medium has inherent spectrum of bands it
will allow.
– Cannot transmit in all bands as it may not be allowed.

• Needs to be shared by many users

– Congestion, Collision, Call Drops

• Distorts the signal

– Attenuation at different frequencies
– Noise/Interference
– Multipath

• Main resource or infrastructure (such as road or pipe)

– Limits the amount of information that can be transmitted
 Wired: Limited Spectrum
• Channel is the most challenging part in communication
– Wired spectrum: the medium has inherent spectrum of bands it
will allow
 Wireless: Limited Spectrum

Radio Spectrum
3 KHz 300 GHz

Lower Valuable Radio Spectrum Higher

Frequencies 300KHz – 5GHz Frequencies

• Low radio frequencies: Large antenna heights

• High radio frequencies
– High attenuation
– Above 5 GHz only line of sight
 Channel: the challenges!
• Limited Spectrum
– Limited wireless spectrum: High frequencies high attenuation, at
low frequencies: large antenna sizes and low bandwidth
– Wired spectrum: the medium has inherent spectrum of bands it
will allow.
– Cannot transmit in all bands as it may not be allowed.

• Needs to be shared by many users

– Congestion, Collision, Call Drops

• Distorts the signal

– Attenuation at different frequencies
– Noise/Interference
– Multipath

• Main resource or infrastructure (such as road or pipe)

– Limits the amount of information that can be transmitted
 Wireless: Need for sharing spectrum
• Shared spectrum across different technologies

2G/ 2G/ LTE CableTV,

AM FM TV 4G 4G /4G
ISM
Satellite
(0.535-1.7) (88-108) (470-900) 900 1800 2100 2400 4000

Frequency in MHz

• Shared spectrum across same technology

– By the end of 2012, the number of mobile-connected devices
exceeded the number of people on earth (approx. 7 billion).
– Leads to congestion, collisions, call drops
 Channel: the challenges!
• Limited Spectrum
– Limited wireless spectrum: High frequencies high attenuation, at
low frequencies: large antenna sizes and low bandwidth
– Wired spectrum: the medium has inherent spectrum of bands it
will allow.
– Cannot transmit in all bands as it may not be allowed.

• Needs to be shared by many users

– Congestion, Collision, Call Drops

• Distorts the signal

– Attenuation at different frequencies
– Noise/Interference
– Multipath

• Main resource or infrastructure (such as road or pipe)

– Limits the amount of information that can be transmitted
 Channel Distortion: Noise
• Sources: Thermal noise at the receiver, Interference from
other sources, man-made EMI, powerline interference
• Modelled as AWGN (additive White Gaussian noise)
(t)

x(t) + y(t) = x(t) + (t)

 Channel Distortion: Multipath
• Guided Medium: Time Invariant
 Channel: Linear Time-Invariant filter
• Channels whose time/ frequency characteristics does
not change in time are modeled in this fashion.
• E.g: Wired Channels
(t)

x(t) h(t) +
y(t) = h(t)*x(t) + (t)
 Channel Distortion: Multipath
• Wireless Channel: Delay and Time Variance
 Channel: Linear Time-Variant filter
• Channels whose time/ frequency characteristics change
in time are modeled in this fashion.
• h( : t) is the time varying impulse response of the filter.
E.g: Wireless Channels with multi-path.

(t)

x(t) h(:t) +
y(t) = h(:t)*x(t) + (t)
 Channel: the challenges!
• Limited Spectrum
– Limited wireless spectrum: High frequencies high attenuation, at
low frequencies: large antenna sizes and low bandwidth
– Wired spectrum: the medium has inherent spectrum of bands it
will allow.
– Cannot transmit in all bands as it may not be allowed.

• Needs to be shared by many users

– Congestion, Collision, Call Drops

• Distorts the signal

– Attenuation at different frequencies
– Noise/Interference
– Multipath

• Main resource or infrastructure (such as road or pipe)

– Limits the amount of information that can be transmitted
Limited Capacity: Shannon’s Theorem

C = 1 Mbps B

10 Mbps

100 Mbps
 Typical Wired Media Bandwidth

Typical Media Max. Bandwidth Max. Speed Max. Physical

Distance
50-Ohm coaxial
Few MHz 10 Mbps 185m
cable (Thinnet)
75-Ohm coaxial
Few MHz 10 Mbps 500m
cable (thicknet)
CAT 5
100 Base-TX 100 MHz 100 Mbps 100m
Ethernet
CAT 5e
1000 Base-TX 100 MHz 1000 Mbps 100m
Ethernet
Fiber 100 Base-FX 2000m
Few GHz 100 Mbps
Ethernet
Fiber 1000 base LX 1000 Mbps (1
Few GHz 5000m
Ethernet Gbps)
 Typical Wireless Bandwidth

Typical Media Max. Theoretical Max. Speed

Bandwidth

WLAN 802.11a 20 MHz 54 Mbps

LTE 3G 20 MHz 100 Mbps

LTE-Advanced 4G 20-100 MHz 1 Gbps

WLAN 802.11n 20-40 MHz 600 Mbps

 Channel: the bright side!
• Without the issues and challenges in the channel, there
would not had been any ECE branch and several
billions dollar companies such as Qualcomm, AT&T, Bell
Labs, Nokia,
Questions?
Types of Communication Systems:
Analog and Digital
 Type of Communication Systems
• Depends on the message signals

Source Transmitter Channel Receiver Destination

message signals
Analog and Digital: S&S
 Analog Communications
• Message signal is analog
– Continuous time signal which takes continuum of values.
• Example: Audio signals, Speech,
• Transmitted signals over physical medium are also
analog
 Analog Communication Systems
• Examples
– AM (amplitude modulation) and FM (frequency modulation)
radios
– Analog television
– first generation cellular technology (AMPS),
– vinyl records, audio cassettes, and VHS
 Analog Communication Systems

• Modulator: Moves the signal to higher frequency for

transmission and multiplexing
• The “obvious” thing to do
– Message waveforms are analog
– Waveforms sent over the channel must be analog
• But not the right thing to do
– Not efficient
– Analog communication is rendered obsolete by digital
communication
 Digital Communications
• Message signal is digital
– Discrete time signal which takes only discrete level of values
– Any signal which can be converted to 0’s and 1’s
• Example
– Demographic data: Census data (Number of people in family,
Gender, Age, Income (rounded to some decimal),
– Text in any language (ASCII code: A 0041H, B0042H,…)
– Data generated or stored by computer and mobiles
 Digital Communications
• Analog signal can be converted to digital signal or
sequence by sampling and quantization
– Songs/Movies stored in CD/DVD, Data in hard-drive
Digital Communication Systems
 Digital Communications: Transmitter
• Source Encoder
– Finite number of messages to bits
– Source compression
• Obtains a digital representation of source signals using minimum binary
digits. This leads to source compression. Removes Redundancy
• Compression based on statistics: Example Huffman coding

– Avg. number of bits/sample = 2

– Avg. number of bits/sample = 0.8(1) + 0.1(2)+0.08(3) +0.02(3) = 1.3
• Compression based on redundancy For e.g.: ZIP, PNG, MPEG, JPEG
– 2000000000 (2 billion) can be represented by 2E9 (3 characters for 10)
– Information bit rate depends on the message and nature of
application
 Digital Communications: Transmitter
• Channel Encoder
– Introduces controlled redundancy to combat channel errors.
 Digital Communications: Transmitter
• Digital Modulator
– Converts the bit stream into a waveform suitable for transmission
over the channel
• For e.g: If the channel has band-pass response, then the modulator
up-converts the frequency band of the information source to the
necessary band.
– Power and bandwidth constraints
 Digital Communication: Receiver
• Digital Demodulator
– It obtains the corrupt waveform and converts it to a bit-stream.
For e.g. It down-coverts the frequency of the waveform and
converts it in bits

• Channel Decoder
– Obtains an estimate of the information bits. Uses the
“Redundancy” to combat channel variations

• Source Decoder
– It obtains an estimate of the actual information transmitted
– Decompress or unzip the signal
– Bits to message mapping
Questions?
 Digital Communication Systems
• Examples
– Cellular 2G/3G/4G/5G
– TV and Radio broadcasting (DVB-T and DRM)
– CD/DVD
– Hard drive
 Digital Communications
• In digital electronics, digital signal transmitted as a pulse
train in baseband

• Digital communication in passband

– Distinct finite number of analog waveforms

• Digital signal processing and storage

8 levels into 3 bits

 Transition from Analog to Digital
• Content is often analog (speech, image, video)
• Signals sent over physical channels are analog
– Currents, voltages, EM waves are continuous-valued,
continuous-time functions
• Many communication systems have shifted
– Analog cellular to digital cellular (CDMA, GSM, OFDM)
– Analog TV/radio to Digital TV/radio
– VR and Audio casettes to CDs, VHS to DVD
 Why Digital?
• Robust against distortion and noise

Source Destination

1 0 1 1 0 0 0 1

Transmitted Signal

Distorted Signal

Received Signal=
Distorted Signal with Noise
Recovered Signal
 Why Digital?
• Robust against distortion and noise
• Case of Analog

Source Destination
 Why Digital?
• Viability of regenerative repeaters for digital comm.

Source RR1 Destination

1 0 1 1 0 0 0 1

Transmitted Signal

Distorted Signal

Received Signal=
Distorted Signal with Noise
Regenerated Signal
 Why Digital?
• Case of Analog and AWGN channels

Source Amplifier Destination

 Why Digital?
• Robust as compared to analog communication
– Viability of regenerative repeaters for digital comm.
• Design of digital communication systems based on
source-channel separation principle
– Source-independent and channel-optimized: Huge Gains
– Not possible in Analog
• Digital hardware is flexible and allows the use of
microprocessors and VLSI circuits
– Scalability
• Digital hardware is much cheaper and more robust
• Digital signals can be coded for arbitrary low error rates
– Shannon’s theorem
• Possibility of encryption
• Digital signal storage is easy and inexpensive to store
• And Many more…
 Is there still a role for analog?
• Of course! The physical world is analog
• Signal transmission: Need to convert digital data to
analog signals that can be sent over the physical
channel
• Signal reception: Need to convert analog received
signals back into digital data
• RF signal processing still a challenge
Today’s Class
 Importance of this subject
• Core subject
– Form foundation for communication theory
– Prerequisite for higher level subjects
• Wireless Communications
• Error Correcting Codes
• Information Theory
• Machine Learning for Communications
Analog Communication Systems

CT
Digital Communication Systems

Error Correcting
Codes

CT
Digital Communication Systems
Information Theory

Wireless
Communications
Digital Communication Systems
Information Theory

Error Correcting
Codes

CT
Wireless
Communications
 Importance of this subject
• Core subject
– Form foundation for communication theory
– Prerequisite for higher level subjects
• Wireless Communications
• Error Correcting Codes
• Information Theory
• Machine Learning for Communications
• Industry and Academic
– Decades of ECE: 5G, 6G, IoT, ML for Comm., mmWave, Tactile
internet (Autonomous vehicles, remote surgery, AR/VR tours)
• Problem solving and analytical skills
• Opportunity to work in SPCRC
Questions?
Course Details/ Logistics
 Syllabus (Tentative)
• Representation of bandpass signals and systems
– lowpass equivalent of bandpass signals and systems
• Analog Communication Methods
– AM-DSB and SSB; FM-narrowband and wideband; Demodulation
of AM and PM/FM, Phased locked loop (PLL);
– Brief overview of Line Coding and PWM
• Digital Modulation
– Representation of Digitally Modulated Signals; Memoryless
modulation methods: PAM, PSK, QAM, Orthogonal Multi-
Dimensional Signals
• Random Processes
– Review of Correlation, ESP and PSD; Noise Modelling, Thermal
Noise, AWGN
• Performance of Digital methods in the presence of AWGN
– Hypothesis testing, Signal Space Concepts, Performance analysis
of ML reception, Bit error probability, Link budget analysis
 Resources
BOOKS
• U. Madhow: Introduction to Communication Systems
– http://www.ece.ucsb.edu/wcsl/Publications/intro_comm_systems_madhow
_jan2014b.pdf
• B.P.Lathi, “Modern Digital and Analog Communication Systems”,
3rd Edition, Oxford University Press, 2007.
• J.G.Proakis, M.Salehi, “Fundamentals of Communication Systems”,
Pearson Education 2006

VIDEOS
• National Programme on Technology Enhanced Learning (NPTEL)
– Analog Communication http://nptel.ac.in/courses/117102059/
– Digital Communication: http://nptel.ac.in/courses/117101051/

VIRTUAL LAB
• Virtual lab on Systems, Communications and Control, IIT Guwahati
http://iitg.vlab.co.in/?sub=59&brch=163
 Course Portal
MOODLE: https://courses.iiit.ac.in/
Under Spring 2025
If you are not already enroled, email me.

• Assignments
• News
• Discussion Forum
 Approach
• Slides + Blackboard
• Active participation in class
• Regular Attendance
• Expected: Read-Attend-Revise-Assignments

Guidelines
• No disturbance in class
• No mobile phones/tablets in class
• No recording
 Exams and Evaluation
• Mark Distribution
– Quiz 1 (10)
– Quiz 2 (10)
– MidSem (20)
– Assignments + Quiz (20)
– Final Exam (40-50)

• Grading: TBD
– Mostly Gaussian with fixed cut/off for A (>=85) and F(<40)
grades)
 Assignments
• 6-8 handwritten assignments + 1-2 Matlab
• Due in one week
• Firm Deadlines
– One late homework assignment allowed without penalty
– 2 marks will be deducted on other late assignments
• Quiz based on Assignment!
– There are 2-2.5 marks for each assignment
– These marks will be given only if the assignment is completed
and quiz answers are correct.
• Cooperative learning is encouraged not copying!!!
– Discussion of concepts in homework is encouraged
– Sharing of homework or code is not permitted and will be
penalized
 Tutorials
• Time: Tuesdays 2-3.30 pm
• Venue: H204
• Attendance: Mandatory if informed in advance!
• Quizes, Problem Solving, Queries
• Use of discussion forum
• TAs: Naganjani, Nikhil, Manasa and Vishnu Priya
EC5.203 Communication Theory I (3-1-0-4):

Lecture 2
Overview of Signals And Systems

Instructor: Dr. Sachin Chaudhari

Email: [email protected]

Jan. 9, 2025
 Reference
• Chapter 2 (Madhow)
– Sec. 2.1-2.5 : Overview of Signals and Systems, Notations
– Sec. 2.6-2.8: Energy Spectral Density, Bandwidth, Structure of
passband signal
Quick Review of Signals and Systems
and Notations
Indicator Function
Sinusoidal Signal

Polar Form
Sinusoidal Signal

Polar Form

Rectangular form

Complex number
Complex Exponential
Inner Product
Energy and Norm
Example 1: Solve!
Power
Time average
Example 2
Convolution
Fourier Series
Fourier Transform

Inverse Fourier
Transform

Fourier Transform
Properties of Fourier Transform
Properties of Fourier Transform
Properties of Fourier Transform
Properties of Fourier Transform
Rectangular pulse and Sinc function: S&S
Rectangular pulse and Sinc function
Summary of Properties: S&S
Summary of Properties: S&S
EC5.203 Communication Theory I (3-1-0-4):

Lecture 3
Baseband and Passband Representations

Instructor: Dr. Sachin Chaudhari

Email: [email protected]

Jan. 16, 2025

and Jan. 20, 2025
 Reference
• Chapter 2 (Madhow)
– Sec. 2.6-2.8: Energy Spectral Density, Bandwidth, Structure of
passband signal
Recap
Indicator Function
Sinusoidal Signal

Polar Form

Rectangular form

Complex number
Complex Exponential
Inner Product
Energy and Norm
Power
Time average
Fourier Series
Fourier Transform

Inverse Fourier
Transform

Fourier Transform
Properties of Fourier Transform
Properties of Fourier Transform
Properties of Fourier Transform
Properties of Fourier Transform
Rectangular pulse and Sinc function: S&S
Rectangular pulse and Sinc function
Today’s Class
Energy Spectral Density
and Bandwidth
 Motivation
• Shared physical medium

• Example of Time Division Multiple Access (TDMA)

11 22 33 44 5 6
Time

• Example of Frequency Division Multiple Access (FDMA)

One band per user

1 2 3 4 5

Frequency
Motivation
Energy Spectral Density
Energy Spectral Density
Bandwidth for bandlimited signals

W
Bandwidth for bandlimited signal: Real Signal
Bandwidth for bandlimited signals: complex

W
Bandwidth for not-bandlimited signal

W
Bandwidth for not-bandlimited signal

W
 Baseband and Passband
The Complex Baseband Representation
Baseband Signal
Examples of baseband signals
Passband Signal
Example of Passband
Example of Passband in Wired Systems

https://en.wikibooks.org/wiki/Communication_Networks/DSL
Differences between Baseband and Passband
Baseband Channel
 Passband Channel

2G/ 2G/ LTE CableTV,

AM FM TV 4G 4G /4G
ISM
Satellite
(0.535-1.7) (88-108) (470-900) 900 1800 2100 2400 4000

Frequency in MHz
Baseband and Passband Signals/Channels
Questions?
Structure of Passband Signal
Motivation
Modulation or Upconversion
Modulation or Upconversion
Modulation: I and Q components

In-phase or Quadrature or
I component Q component
Modulation: I and Q components

Carry all the Carry NO

information information
Orthogonality of I and Q channels
Demodulation or Downconversion
Upconversion and Downconversion: Block Diagrams
Envelope and Phase
Complex Envelope
Relationship between three representations

Information resides in complex baseband

Time domain expressions for a passband signal

Starting from one representation, can derive the

rest based on the relations depicted in the figure
Effect of Frequency and Phase Offset
Coherent Detection: Synchronization
Summary of Properties: S&S
Frequency Domain Relationship
Construct Passband from Baseband
Reconstruct Baseband from Passband
I and Q components in frequency domain
 Passband filtering = complex baseband filtering

EQUIVALENT TO u(t) h(t) y(t)

up(t) hp(t) yp(t)
(except for a scale factor of 2)
What is complex baseband representation of passband filtering?
Filtering in complex baseband
Energy and Power
Correlation between two signals
Modern transceiver architectures based on complex baseband

All the action is in complex baseband for a typical wireless transceiver

RF signal Filtering
Channel Encoding Synchronization
Modulation Upconversion Dnconversion
Demodulation
Spectral Shaping Decoding
Complex Complex
DSP-based envelope envelope DSP-based
processing processing
TRANSMITTER RECEIVER
Complex baseband representation: summary
EC5.203 Communication Theory I (3-1-0-4):

Lecture 5
Analog Communication Techniques:
Amplitude Modulation

Instructor: Dr. Sachin Chaudhari

Email: [email protected]

Jan. 27, 2025

 References
• Chap. 3 (Madhow)
Analog comm techniques: Motivation
Terminology and Notations
Terminology and Notations
Terminology and Notations
 Key Concepts

Amplitude Frequency Phase

Modulation Modulation Modulation
Key Concepts
Amplitude Modulation
Example: sinusoidal message
AM: Double Sideband Suppressed Carrier
DSB-SC signal for sinusoidal message
Example 2
DSB-SC spectrum for Example 2

Bandwidth of passband
Demodulation of DSB-SC
 Recap: Chapter 2
Effect of Frequency and Phase Offset
Example: Phase Offset
Causes of Phase Offset
Need of Coherent Detection
Conventional AM
Conventional AM
Conventional AM: spectrum

Strong carrier component

Envelope and its importance
What does the envelope tell us?
Sidestepping sync requirement
Constraint for recovering message from envelope

Example of sinusoidal message

Envelope = message + DC Envelope = message

Message info not preserved

in envelope
 Example of Sinusoidal Message

Envelope = message + DC Envelope = message + DC

Message info not

preserved in envelope
Modulation Index
AM signal in terms of modulation index
 Effect of modulation index
Example of sinusoidal message

amod = 0.5
amod =1

 Envelope = message + DC Envelope = message



Message info not preserved

in envelope
amod =1.5
 Envelope Detectors

Positive carrier cycle ➔ capacitor charges up (reaches value of envelope)

Negative carrier cycle ➔ capacitor discharges with RC time constant
 Envelope detector operation

Positive carrier cycle ➔ capacitor charges up (reaches value of envelope)

Negative carrier cycle ➔ capacitor discharges with RC time constant
Should not discharge too fast during negative cycle
Should react fast enough to follow variations in envelope (which depend on
message bandwidth B)
EC5.203 Communication Theory I (3-1-0-4):

Lecture 6:
Analog Communication Techniques:
Amplitude Modulation

Feb. 03, 2025

Recap
 Key Concepts

Amplitude Frequency Phase

Modulation Modulation Modulation
AM: Double Sideband Suppressed Carrier
DSB-SC signal for sinusoidal message
Example 2
DSB-SC spectrum for Example 2

Bandwidth of passband
Demodulation of DSB-SC
Causes of Phase Offset
Need of Coherent Detection
Conventional AM
Conventional AM: spectrum

Strong carrier component

Envelope and its importance
What does the envelope tell us?
Sidestepping sync requirement
Constraint for recovering message from envelope

Example of sinusoidal message

Envelope = message + DC Envelope = message

Message info not preserved

in envelope
 Example of Sinusoidal Message

Envelope = message + DC Envelope = message

Message info not

preserved in envelope
Modulation Index
AM signal in terms of modulation index
 Effect of modulation index
Example of sinusoidal message

amod = 0.5
amod =1

 Envelope = message + DC Envelope = message



Message info not preserved

in envelope
amod =1.5
 Envelope Detectors

Positive carrier cycle ➔ capacitor charges up (reaches value of envelope)

Negative carrier cycle ➔ capacitor discharges with RC time constant
 Envelope detector operation

Positive carrier cycle ➔ capacitor charges up (reaches value of envelope)

Negative carrier cycle ➔ capacitor discharges with RC time constant

Should not discharge too fast during negative cycle

Should react fast enough to follow variations in envelope (which depend on
message bandwidth B)
Todays’ Class
 References
• Chap. 3 (Madhow)
• BP Lathi
 Conventional AM Modulators

B.P.Lathi pages: 155-159

How do we do conventional AM modulation?
Non-linear Modulators
Switching Modulators

DSB-SC
Switching Modulators

Conventional
DSB

Suppressed by BPF
 Conventional AM Demodulators

B.P.Lathi pages: 166-169

Rectifier Detector

Lathi
 Envelope detector operation

Madhow

Positive carrier cycle ➔ capacitor charges up (reaches value of envelope)

Negative carrier cycle ➔ capacitor discharges with RC time constant
Should not discharge too fast during negative cycle
Should react fast enough to follow variations in envelope (which depend on message
bandwidth B)
Power efficiency of conventional AM
Power efficiency of conventional AM

Extra Non-information carrying component

Comments on Conventional AM
Example on power efficiency computation

Tutorial
Example: Tutorial!
Questions?
EC5.203 Communication Theory I (3-1-0-4):

Lecture 7:
Analog Communication Techniques:
Amplitude Modulation - 3

Feb. 06, 2025

Recap
 Key Concepts

Amplitude Frequency Phase

Modulation Modulation Modulation
AM: Double Sideband Suppressed Carrier
DSB-SC signal for sinusoidal message
Example 2
DSB-SC spectrum for Example 2

Bandwidth of passband
Demodulation of DSB-SC
Need of Coherent Detection
Conventional AM
Conventional AM: spectrum

Strong carrier component

Sidestepping sync requirement
Modulation Index
AM signal in terms of modulation index
 Effect of modulation index
Example of sinusoidal message

amod = 0.5
amod =1

 Envelope = message + DC Envelope = message



Message info not preserved

in envelope
amod =1.5
 Envelope Detectors

Positive carrier cycle ➔ capacitor charges up (reaches value of envelope)

Negative carrier cycle ➔ capacitor discharges with RC time constant
 Conventional AM modulation

B.P.Lathi pages: 155-159

Power efficiency of conventional AM

Extra Non-information carrying component

Todays’ Class
Amplitude Modulation:
Single Side Band
 SSB: Motivation

Message signal

m(t) is complex envelope of DSB

 DSB ➔ USB (SSB)

DSB signal

Message signal USB signal

 DSB ➔ LSB (SSB)

Message signal DSB signal

 Recover Signal from USB Signal? Solve!

Upper Side Band (USB) Signal

Message signal
 Solution!

Real part of
original SSB

Right Shifted
SSB by fC

Left Shifted
SSB by fC

Sum of the FILTER

two
Corresponding Time Domain Equations

Simply Extracts I
component
Message signal is I component of filter o/p
 Recover Signal from USB Signal!

Upper Sideband Signal

Message signal
 Recover Signal from LSB Signal?

Lower Sideband (LSB) Signal

Message signal
Use of Hilbert Transform for
SSB Generation
Motivation: Requirement of ideal filters
Hilbert transform
SSB in baseband using Hilbert Transform
 USB passband signal is real!

Upper Sideband Signal (Passband)

Baseband Message signal

m(t) will be real!

m(t) is complex envelope of DSB

In SSB discussion, u(t) will be complex envelope of USB
 Complex envelope for USB signal

USB passband

In SSB discussion, u(t) will be complex envelope of USB

I and Q components for SSB
Sign Function
Message signal is I component of filter o/p
Hilbert transform
SSB and Hilbert Transform
Complex envelope for SSB signal
Complex envelope for SSB signal

Complex baseband

Real USB passband

Real in time
Even and Odd
Symmetric in frequency domain
 Implementing SSB in baseband

Implementing Hilbert transform in baseband avoids need

for sharp filtering at passband
SSB for sinusoidal message
SSB demodulation: Coherent

Attenuation Interference

Assignment!
SSB demodulation: Noncoherent
Questions?
Quadrature Amplitude Modulation
QAM
 QAM

USB LSB LSB USB

Need both sidebands!

Effect of Frequency and Phase Offset
Coherent Detection: Synchronization
Questions?
EC5.203 Communication Theory I (3-1-0-4):

Lecture 8:
Analog Communication Techniques:
Frequency Modulation - 1

Feb. 10, 2025

Recap
 Key Concepts

Amplitude Frequency Phase

Modulation Modulation Modulation
AM: Double Sideband Suppressed Carrier
Conventional AM
Power efficiency of conventional AM

Extra Non-information carrying component

 DSB ➔ USB (SSB)

DSB signal

Message signal USB signal

SSB in baseband using Hilbert Transform
Todays’ Class
 Amplitude Modulation:
Vestigial Side Band (VSB)
VSB signaling
How to choose VSB filter?
Quadrature Amplitude Modulation
QAM
 QAM

USB LSB LSB USB

Need both sidebands!

Effect of Frequency and Phase Offset
Coherent Detection: Synchronization
Questions?
Frequency Modulation
 Recap: Different Modulations

Amplitude Frequency Phase

Modulation Modulation Modulation
Frequency Modulation
Example of FM Wave
Example 2 of FM Wave
Phase Modulation
Example of PM Wave
Generalized Model: Angle Modulation
Equivalence of PM and FM: FM using PM

FM

PM
Equivalence of FM and PM: FM using PM
Equivalence of PM and FM: PM using FM

PM

FM
Poll
Non-linearity of Angle or Phase Modulation
 PM versus FM

PM
(PSK for digital
messages)

Digital message
Discontinuous phase ➔ vulnerable to nonlinearities,
poorer frequency containment

Continuous phase ➔ more robust to nonlinearities,

better frequency containment

FM
 PM versus FM in practice

Focus on FM in this chapter. PSK studied in Chapter 4 and beyond.

Frequency Modulation
Example of FM Wave
FM Modulation: Effect on Phase
FM Modulation Index
Modulation
 Example of VCO

Message Passband Sinusoid with

signal varying freq.

Source: Globe Collector - Own work, CC0,

https://commons.wikimedia.org/w/index.php?curid=18036362
FM Demodulation
 Limiter Discriminator

Enforces constant envelope

Converts FM to Amplitude changes
A crude discriminator

Envelope DC
d/dt
Detector Block
Approximate differentiation
Questions
EC5.203 Communication Theory I (3-1-0-4):

Lecture 9:
Analog Communication Techniques:
Frequency Modulation - 2

Feb. 13, 2025

Recap
 Key Concepts

Amplitude Frequency Phase

Modulation Modulation Modulation
Frequency Modulation
Example of FM Wave
Example 2 of FM Wave
Phase Modulation
Example of PM Wave
Generalized Model: Angle Modulation
Equivalence of PM and FM: FM using PM

FM

PM
Equivalence of FM and PM: FM using PM
Equivalence of PM and FM: PM using FM

PM

FM
 PM versus FM

PM
(PSK for digital
messages)

Digital message
Discontinuous phase ➔ vulnerable to nonlinearities,
poorer frequency containment

Continuous phase ➔ more robust to nonlinearities,

better frequency containment

FM
 PM versus FM in practice

Focus on FM in this chapter. PSK studied in Chapter 4 and beyond.

Frequency Modulation
FM Modulation Index
Modulation
FM Demodulation
 Limiter Discriminator

Enforces constant envelope

Converts FM to Amplitude changes
 Today’s Class
Ref Books: U. Madhow and B. P. Lathi
FM Spectrum
FM spectrum
 Recap: Time domain expressions for a passband signal

Starting from one representation, can derive the

rest based on the relations depicted in the figure
Narrowband FM
Narrowband FM: Example
Questions?
Wideband FM
Carson’s rule
Bandwidth of Angle Modulated Waveforms

Ref: B.P.Lathi
FM spectrum for periodic messages
(Approximate) FM spectrum
Exact FM spectrum for sinusoidal message

Complex-valued integral

Real Valued
Bessel function plots
Bessel function properties

This is only an approximation

Modulation index and Power in Sidebands

https://en.wikipedia.org/wiki/Frequency_modulation#Bessel_functions
Fourier coefficients for complex envelope
Fractional power containment BW
Carson’s rule
Example 3.3.1: Tutorial
Features of Angle Modulated Non-linearities

Ref: B.P.Lathi
EC5.203 Communication Theory (3-1-0-4):

Lecture 10:
Analog Communication Techniques:
Superheterodyne Receiver

Feb. 22, 2025

 References
• Chap. 3 (Madhow)
 Recap:
FM Spectrum
FM spectrum
Wideband FM
Carson’s rule
Bandwidth of Angle Modulated Waveforms

Ref: B.P.Lathi
FM spectrum for periodic messages
(Approximate) FM spectrum
Exact FM spectrum for sinusoidal message

Complex-valued integral

Real Valued
Fractional power containment BW
Features of Angle Modulated Non-linearities

Ref: B.P.Lathi
 Today’s Class

Superheterodyne Receiver:
Applicable to AM/FM/PM
 Downconversion: How it is done?

Interference Signal

Desired Signal
Different Methods for Receiver
Superheterodyne Receiver

Source:
https://en.wikipedia.org/wiki/Superheterodyne_receiver
Motivation for Superheterodyne Receivers
 Vacuum Tube Triode

Symbol: http://www.allaboutcircuits.com/textbook/se
miconductors/chpt-13/the-triode/

http://www.oneillselectronicmuseum.com/lar
gephotos/tubes/yel/yel23.jpg

Amplifier Circuit: http://www.daenotes.com/electronics/devices-circuits/vacuum-tube-triode

 Transistors

Old Germanium Modern (Silicon)

http://ibm-1401.info/VintageGermaniumComputerTransistorsDavid%20Laws.pdf

http://www.talkingelectronics.com/projects/BasicElectronics-1A/BasicElectronics-1A_Page2.html
Example of Superheterodyne Transistor Receiver

Pocket transistor radio, late 1960s to

1970s, with 5 germanium transistors, that
received long wave and medium wave
(AM broadcast) bands. View with back
open, showing parts. Powered by 2
button cells supplying 3V

By Pocket_radio_open.jpg: Ulfbastelderivative work: Chetvorno (talk) - Pocket_radio_open.jpg, Public

Domain, https://commons.wikimedia.org/w/index.php?curid=12819355
Superheterodyne receiver: Principle
Radio Frequency Intermediate Frequency Baseband
 Mixer

https://en.wikipedia.org/wiki/Frequency_mixer
Mixer: Use of Non-linearity
Example: superhet for AM radio

AM Radio frequencies: 535-1605 KHz

Audio Message is limited to 5 KHz
Bandwidth of conventional AM is 10 KHz
This makes 106 possible bands
Center frequency can range between 540-1600 KHz
 Example: superhet for FM radio
http://www.daenotes.com/electronics/communication-system/superheterodyne-fm-receiver

FM Radio frequencies: 88-108 MHz

Bandwidth of 150 KHz + 25 KHz guard band
200KHz spacing makes 100 stations
Center frequency assigned starting at 88.1 MHz
Superhet: freq domain operations
Disadvantages of SHR
Direct Conversion Receiver

Source: Sachin Chaudhari, PhD Thesis; "Sensing for Cognitive Radios: Algorithms, Performance, and Limitations",
Aalto University School of Electrical Engineering, Finland, 2012
Disadvantages of DCR
Receiver for mmWave communication?
Phase Locked Loop (PLL)
PLL intro
High-level view of PLL
Mixer as Phase Detector
Mixers can extract phase differences
Mixers can extract phase differences
Mixer-based phase detector

Phase difference term Double frequency term

driving feedback loop (filtered out)
Mixed signal phase detection
 XOR-based phase detector

Average of XOR output is related to phase difference

XOR-based phase detector output
PLL Applications:
FM Demodulation
and
Frequency Synthesis
PLL for FM demodulation
 PLL for frequency synthesis

40 MHz VCO can produce 2.4 GHz

but not exact value
ECE335 Communication Theory I (3-1-0-4):

Lecture 11:
Analog Communication Techniques:
Phase Locked Loop (PLL)

Feb. 22, 2025

 References
• Chap. 3 (Madhow)
PLL intro
High-level view of PLL
Mixer as Phase Detector
Mixers can extract phase differences
Mixers can extract phase differences
Mixer-based phase detector

Phase difference term Double frequency term

driving feedback loop (filtered out)
Mixed signal phase detection
 XOR-based phase detector

Average of XOR output is related to phase difference

XOR-based phase detector output
 Today’s Class

PLL Applications:
FM Demodulation and Frequency Synthesis
PLL for FM demodulation
 PLL for frequency synthesis

40 MHz VCO can produce 2.4 GHz

but not exact value
Mathematical Model for PLL
For mixer-based phase detector
For mixer-based phase detector
 For mixer-based phase detector

Replace sine by triangular wave

for XOR-based phase detector
 For mixer-based phase detector

This is difficult to analyze as system is non-linear

Linearized Model
Linearized analysis
First order PLL
1st order PLL: phase step response
1st order PLL: freq step response


Second order PLL
2nd order PLL: freq step response
Choosing PLL order
Example
Actual nonlinear PLL model
Original nonlinear model
Phase plane plot
 PLL locking

Unstable
Stable equilibrium
equilibrium
Nonlinear vs linearized PLL analysis
EC5.203 Communication Theory (3-1-0-4):

Lecture 12:
Digital Modulation

03 March 2025
 References
• Chap. 4 (Madhow)
Digital modulation
 Recap: Digital Communications
• Analog signal can be converted to digital signal or
sequence by sampling and quantization
– Songs/Movies stored in CD/DVD, Data in hard-drive
Recap: Digital Communication Systems
 Digital modulation: baseband example

Bit-to-symbol
Map Pulse
..1001011… Modulation
 Digital modulation: passband example
Bit-to-symbol
Map Pulse
..1001011… Modulation
BPSK (also called 2-PAM)
Baseband

Passband
Signal Constellation for BPSK
Baseband
BPSK (also called 2-PAM)

Baseband
Can we load more bits per symbol?
Another way for 2 bits/symbol: QPSK

01 11

00 10
 QPSK baseband
Pulse
1st bit
Modulation
Take
2 bits
Pulse
2nd bit
Modulation
QPSK: Passband
QPSK: Passband
Can we load even more bits?

COMMON CONSTELLATIONS
Bandwidth Occupancy
 Motivation

CableTV,
AM FM TV 2G 2G LTE ISM
Satellite
(0.535-1.7) (88-108) (470-900) 900 1800 2100 2400 4000

Frequency in MHz

FM1 FM 2 FM N

88 108 MHz
Frequency in MHz

Ideal: No interference between different bands

Practical: Some interference between different bands

Modeling bandwidth occupancy
Power Spectral Density (PSD)
Periodogram-based PSD estimation
PSD of Linearly Modulated Signal
EC5.203 Communication Theory (3-1-0-4):

Lecture 13:
Digital Modulation - 2

06 March 2025
 References
• Chap. 4 (Madhow)
Digital modulation
 Digital modulation: baseband example

Bit-to-symbol
Map Pulse
..1001011… Modulation
 Digital modulation: passband example
Bit-to-symbol
Map Pulse
..1001011… Modulation
Can we load more bits per symbol?
Another way for 2 bits/symbol: QPSK

01 11

00 10
 QPSK baseband
Pulse
1st bit
Modulation
Take
2 bits
Pulse
2nd bit
Modulation
QPSK: Passband
Can we load even more bits?

COMMON CONSTELLATIONS
 Motivation for Bandwidth Occupancy

CableTV,
AM FM TV 2G 2G LTE ISM
Satellite
(0.535-1.7) (88-108) (470-900) 900 1800 2100 2400 4000

Frequency in MHz

FM1 FM 2 FM N

88 108 MHz
Frequency in MHz

Ideal: No interference between different bands

Practical: Some interference between different bands

Modeling bandwidth occupancy
Power Spectral Density (PSD)
Periodogram-based PSD estimation
PSD of Linearly Modulated Signal
Today’s Class
 Bandwidth based on PSD

B/2
B

We will focus mostly on this!

Time Frequency Normalization: Time Domain

t in secs t0 = t/T Time Unit = T secs

Time Frequency Normalization: Freq. Domain

normalized

B1/T Hz
B1

f Frequency unit = 1 cycle/sec f0 = fT Frequency unit = 1/T

Bandwidth computation for Rectangular Pulse

Real and Positive

Parseval

Symmetry of PSD
Bandwidth Computation: Example
Poll!
Bandwidth computation for Sine Pulse
Design for Bandlimited Channels:
Nyquist Sampling Criteria

Recap of S&S Oppenheim Chapter 7

[Not in syllabus]
Relation between CT and DT signals

Sampling
Relation between CT and DT signals

Interpolation
 Importance of Sampling

C/D D/C
DT System
Conversion Conversion
Issue with sampling
Sampling Theorem
S&S Recap: Impulse Train Sampling (Time)
S&S Recap: Impulse Train Sampling (Freq.)
Reconstruction of signal: Frequency Domain
S&S Recap: Frequency Domain
S&S Recap: Reconst. of signal (Freq. Domain)
S&S Recap: Reconst. of signal (Time Domain)
Questions?
Design for Bandlimited Channels:
Nyquist Sampling Criteria
[Madhow]
Nyquist Sampling Theorem (Book Notations!!!)
Problem With Sinc Pulse

Time-offset=0.3
Design for Bandlimited Channels:
Nyquist Criteria for pulse shaping!
Bandlimited Channels
 Motivation

CableTV,
AM FM TV 2G 2G LTE ISM
Satellite
(0.535-1.7) (88-108) (470-900) 900 1800 2100 2400 4000

Frequency in MHz

FM1 FM 2 FM N

88 108 MHz
Frequency in MHz

Ideal: No interference between different bands

Practical: Some interference between different bands

 Design for Bandlimited Channels

Samples to
Pulse
Sampling Bits to
Modulation
Symbols

Minimum bandwidth dictated by Sampling

W=1/T
Bandwidth occupancy can be designed based
on P(f) and independent of S(f) (FT of s(t)).

C/D D/C
DT System
Conversion Conversion
Sum of Sinc Pulses: Ex. 1

• Normalized W=1=1/T
• p(t) is sinc function
Signal as Sum of Sinc Pulses: Ex. 2

• Normalized W=1=1/T
• p(t) is sinc function
 What is Inter Symbol Interference (ISI)?

Time Domain p(t) Freq. Domain P(f)

Sinc
Rectangular Pulse
No ISI=> SmallTiming offset
does not cause issues

Freq. Domain P(f)

Rect. Pulse
Sinc Pulse
ISI=> SmallTiming offset
does not cause issues

No ISI at sampling instances though!

Nyquist Criterion for ISI avoidance

DT Fourier Transform Pair

Nyquist Criterion for ISI avoidance
S&S Recap: Impulse Train Sampling (Time)
S&S Recap: Impulse Train Sampling (Freq.)
Theorem 4.5.1: Sampling
EC5.203 Communication Theory (3-1-0-4):

Lecture 14:
Digital Modulation - 3

10 March 2025
 References
• Chap. 4 (Madhow)
Recap: Bandwidth Occupancy
 Motivation

CableTV,
AM FM TV 2G 2G LTE ISM
Satellite
(0.535-1.7) (88-108) (470-900) 900 1800 2100 2400 4000

Frequency in MHz

FM1 FM 2 FM N

88 108 MHz
Frequency in MHz

Ideal: No interference between different bands

Practical: Some interference between different bands

Modeling bandwidth occupancy
PSD of Linearly Modulated Signal
 Bandwidth based on PSD

B/2
B

We will focus mostly on this!

Time Frequency Normalization: Time Domain

t in secs t0 = t/T Time Unit = T secs

Time Frequency Normalization: Freq. Domain

normalized

B1/T Hz
B1

f Frequency unit = 1 cycle/sec f0 = fT Frequency unit = 1/T

Bandwidth computation for Rectangular Pulse

Real and Positive

Parseval

Symmetry of PSD
Design for Bandlimited Channels:
Nyquist Criteria for pulse shaping!
Nyquist Sampling Theorem (Book Notations!!!)
 Design for Bandlimited Channels

Samples to
Pulse
Sampling Bits to
Modulation
Symbols

Minimum bandwidth dictated by Sampling

W=1/T
Bandwidth occupancy can be designed based
on P(f) and independent of S(f) (FT of s(t)).

C/D D/C
DT System
Conversion Conversion
Sum of Sinc Pulses: Ex. 1

• Normalized W=1=1/T
• p(t) is sinc function
Signal as Sum of Sinc Pulses: Ex. 2

• Normalized W=1=1/T
• p(t) is sinc function
 What is Inter Symbol Interference (ISI)?

Time Domain p(t) Freq. Domain P(f)

Sinc
Rectangular Pulse
No ISI=> SmallTiming offset
does not cause issues

Freq. Domain P(f)

Rect. Pulse
Sinc Pulse
ISI=> SmallTiming offset
does cause issues

No ISI at sampling instances though!

Nyquist Criterion for ISI avoidance

DT Fourier Transform Pair

Nyquist Criterion for ISI avoidance
S&S Recap: Impulse Train Sampling (Time)
S&S Recap: Impulse Train Sampling (Freq.)
Today’s Class
Theorem 4.5.1: Sampling
Significance of Nyquist Criterion for ISI avoidance
 Nyquist Pulses?
Corresponds to sinc in time domain
Nyquist pulse with
Not Nyquist minimum bandwidth 1/T

Nyquist
Problem With Sinc Pulse: ISI

Time-offset=0.3
 Problem with Sinc Pulse

Time-offset=0.3
Data: All ones were sent for -3<n<3 and 0 otherwise.

Here interference considered only from neighbouring 4 pulses

Problems with Sinc Pulse: Slow Decay
Nyquist Pulses: Excess Bandwidth

Not Nyquist Nyquist pulse with minimum bandwidth

Nyquist pulse with excess bandwidth

Need of Excess Bandwidth!
Trapezoidal Pulse
Questions?
Trapezoidal Pulse: Normalized Freq.
Trapezoidal Pulse: Non-normalized
Trapezoidal Pulse: General Expressions with scaling
Some Interesting Properties of Nyquist Pulses
Sinc Pulse and Nyquist Rate

; T=1 sec

1/a = 2
Example: Tutorial
Trapezoidal Pulse: Example 4.1
Trapezoidal Pulse: Example 4.1

Min. Nyquist bandwidth = 1/T

Raised-cosine pulse: Freq. Domain
Raised-cosine pulse: Time Domain
Example: Tutorial

Sampling at
rate T
Example continues: Tutorial
Bandwidth Efficiency
Questions?
EC5.203 Communication Theory (3-1-0-4):

Lecture 15:
Digital Modulation - 4

13 March 2025
 References
• Chap. 4 (Madhow)
Recap: Design for Bandlimited Channels:
Nyquist Criteria for pulse shaping!
 Motivation

CableTV,
AM FM TV 2G 2G LTE ISM
Satellite
(0.535-1.7) (88-108) (470-900) 900 1800 2100 2400 4000

Frequency in MHz

FM1 FM 2 FM N

88 108 MHz
Frequency in MHz

Ideal: No interference between different bands

Practical: Some interference between different bands

 Design for Bandlimited Channels

Samples to
Pulse
Sampling Bits to
Modulation
Symbols

Minimum bandwidth dictated by Sampling

W=1/T
Bandwidth occupancy can be designed based
on P(f) and independent of S(f) (FT of s(t)).

C/D D/C
DT System
Conversion Conversion
 What is Inter Symbol Interference (ISI)?

Time Domain p(t) Freq. Domain P(f)

Sinc
Rectangular Pulse
No ISI=> SmallTiming offset
does not cause issues

Freq. Domain P(f)

Rect. Pulse
Sinc Pulse
ISI=> SmallTiming offset
does cause issues

No ISI at sampling instances though!

Nyquist Criterion for ISI avoidance

DT Fourier Transform Pair

Nyquist Criterion for ISI avoidance
Theorem 4.5.1: Sampling
Significance of Nyquist Criterion for ISI avoidance
Nyquist Pulses: Excess Bandwidth

Not Nyquist Nyquist pulse with minimum bandwidth

Nyquist pulse with excess bandwidth

Need of Excess Bandwidth!
Trapezoidal Pulse: General Expressions with scaling
Some Interesting Properties of Nyquist Pulses
Sinc Pulse and Nyquist Rate

; T=1 sec

1/a = 2
 Today’s Class
Orthogonal Modulation
Orthogonal Modulation
 Frequency Shift Keying (FSK)

01 10 00 00 11

M=2 M=4
FSK
FSK: coherent demodulation (passband)
FSK: Issue with coherent demodulation
Noncoherent Reception
Summarizing the concept of Orthogonality
Biorthogonal Modulation
EC5.203 Communication Theory (3-1-0-4)

Lectures 16 and 17:

Noise Modelling,
And
Linear Operations on Random Process

17 and 20 March 2025

 References
• Chap. 5 (Madhow)
– Sections 5.1-5.7: Probability theory and Random Process (Self-
Study, Not part of syllabus)
– Section 5.8: Noise Modeling
– Section 5.9: Linear operation on Random Processes
Noise Modeling
 Noise in Communication Systems

(t)
x(t) +
y(t) = x(t) + (t)
 Thermal Noise

https://www.st-andrews.ac.uk/~www_pa/Scots_Guide/iandm/part3/page1.html
Thermal Noise..
Modeling Noise: Passband model
Real-valued
Modeling Noise: Baseband model
Real-valued
Unified Model: White Noise

implicitly assumes that later receiver

processing will limit the bandwidth
Modeling Noise as Gaussian
White Gaussian Noise (WGN)
 Example computation: Tutorial

5 GHz WLAN with receiver bandwidth 20 MHz and receiver noise figure 6 dB.
What is the noise power?

convert to dBm

Can do this computation directly in the dB domain.

Filtering
Filtered Random Processes
Filtered Random Processes
WSS Filtered Random Processes
Wide Sense Stationarity
Example
Example: Tutorial
Correlation
Definition and Motivation
Recap: FSK coherent demodulation
SNR
Theorem 5.9.1
Filter as Correlator
Matched Filter
EC5.203 Communication Theory I (3-1-0-4):

Lectures 18-23:
Optimal Demodulation

24 March 2025 to 21 April 2025

 References
• Chap. 6 (Madhow)
Optimal Demodulation
Received Data at SNR = 30 dB
Received Data at SNR = 20 dB
Received Data at SNR=10dB
Optimal Demodulation

AWGN
S&S Recap: Signal Energy
S&S Recap: Signal Power
S&S Recap: Note on Dimension of Energy and Power Definitions
Example 5.6.3
(Bit) Error Probability vs SNR for Example
Ingredients of Hypothesis Testing Framework

Example of Decision regions for M=4

Error Probabilities
Ingredients of Hypothesis Testing Framework
MAP rule
Optimality of MAP (or MPE) rule
ML rule
Binary Hypothesis Testing Problem
Example of exponential distribution
Likelihood Ratio
Likelihood Ratio: MAP
Signal Space Concepts
Receiver design as hypothesis testing
Signal space concepts: an example
Example (continued)
Another example: 2D modulation
Another example: 2D modulation…
Basis expansion  constellation
From Signals to Vectors
 Gram-Schmit Orthogonalization Process
Orthogonal Vectors Orthonormal Vectors

This operator projects the vector v orthogonally onto the

line spanned by vector u.

Source: Wikipedia http://en.wikipedia.org/wiki/Gram%E2%80%93Schmidt_process

 Gram-Schmit Orthogonalization Process

Source: Wikipedia http://en.wikipedia.org/wiki/Gram%E2%80%93Schmidt_process

Gram-Schmidt Orthogonalization Process
Example: Gram-Schmidt Orthogonalization
 From signals to vectors

We can express each signal as a vector of basis coefficients

 si [0] 
 
 i s [1] 
n −1
 . 
si (t) =  si [k] k (t), where si [k] = si , k  si =  
 . 
l =0
 . 
 
 si [n −1]


 Inner products are preserved

Inner products are preserved from CT to DT and vice versa

➔ norms, energies, distances are preserved
Modeling WGN in signal space

Noise

Dimension of signal space

Recap: FSK coherent demodulation
Noise projection on Signal Subspace
Noise projection on Signal Subspace
Geometric interpretation of WGN
Hypothesis testing in signal space
Irrelevance of orthogonal-noise component
Optimal Reception in AWGN

This is minimum distance rule!

Min distance rule between 2 signals
Geometry of min distance rule

Too far away from

s1 to affect its
decision region
 2 Dimensional Examples

Solve: Find decision regions for these constellations for ML demodulation

ML decisions for common 2D constellations

ML regions are
sectors
ML boundaries
parallel to I/Q axes,
decision regions are
the four quadrants

ML boundaries
parallel to I/Q axes
MAP Demodulation In CT

DT(Finite Dimensional)

CT(Infinite Dimensional)
Recap: Receiver design as hypothesis testing
ML Demodulation In CT

DT(Finite Dimensional)

CT(Infinite Dimensional)

Infinite Power
Implementation using correlation

DT(Finite Dimensional)

CT(Infinite Dimensional)
Equivalent Match Filter Implementation

DT(Finite Dimensional)

CT(Infinite Dimensional)

Passband implementation
Implementation in Complex Baseband
Performance Analysis of ML Reception
Performance Analysis of ML Reception
 The Geometry of Errors
Basic building block:
Send signal s.
Noise N gets added.
P[crossing a given boundary]?

Noise component parallel to

boundary cannot cause crossing,
no matter how big it is.
In n-dimensional space, this Boundary crossing determined by length and sign of
component is (n-1)-dimensional noise component perpendicular to boundary. 1-
dimensional, regardless of the signal space dimension
 Geometry for binary signaling

(starting from either signal)

Important Note
Performance with binary signaling: Algebraic
Importance of Scale Invariance
Some Standard Measures
ML Binary Signaling Performance..
 Performance for different binary schemes

d =1 d =2 d= 2
E b = (0 2 +12 ) /2 = 1/2 E b = ((−1) 2 +12 ) /2 = 1 E b = (12 +12 ) /2 = 1
P = d 2 / E b = 2 P = d 2 / E b = 4 P = d 2 / E b = 2
  E   E    E   2E    E   E 
Pe,ML = Q P b
 = Q b
 Pe,ML = Q P b
 = Q b
 Pe,ML = Q P b
 = Q b

 2N 0   N 0   2N 0   N 0   2N 0   N 0 

 
Performance for different binary schemes..
Performance analysis for larger
constellations
M-ary signaling in AWGN
Some important parameters
Exact Error Probability of QPSK
Performance analysis for
larger constellations:
Union Bounds and Variants
Exact analysis might be difficult!
Approximation using Union Bound
Applying the union bound

Simple evaluation using Q-functions

Exact Analysis for higher dimensionals difficult!

Can’t draw pictures like this in a 20-dimensional signal space!

General Union Bound for M-ary Signal
Applying the general union bound

Intelligent union bound

Intelligent Union Bound
Example: QPSK
Example: QPSK ..
Nearest Neighbor Approximation
Example: QPSK ...
Performance Analysis of 16 QAM
16QAM performance…
 QPSK vs 16QAM

About 4 dB

Note: Intelligent union bound is very close to the exact error prob in both cases
Performance Analysis of M-ary
orthogonal modulation

Included in course

Not included in course

Symbol Error Probability for M-ary Orthogonal
Power-Bandwidth Tradeoff for M-ary orthogonal!