Course Title: Advanced Wireless Communication

Wireless Communication
Wireless communication refers to the transmission of information over a distance without the use
of physical conductors or wires. It uses electromagnetic waves (such as radio waves,
microwaves, infrared) to enable the transfer of voice, data, and multimedia between devices.
Common systems include cellular networks, Wi-Fi, satellite links, and IoT networks.
Example: Your smartphone communicating with a 5G base station, a satellite TV dish receiving
signals from space, Bluetooth headphones connecting to your phone, Wi-Fi enabling internet
access on your laptop.
Necessity: Enables mobility, flexibility, and connectivity in scenarios where wired connections
are impractical, costly, or impossible. It's the backbone of modern mobile networks, IoT devices,
satellite communication, and many other applications.
Components:
• Transmitter
• Receiver
• Communication channel
Communication Channel
A communication channel is the medium through which a wireless signal propagates from the
transmitter to the receiver. In wireless systems, this includes free space, air, obstacles, and the
radio environment, which may introduce impairments like noise, fading, and delay spread.
Power Dissipation
Power dissipation refers to the conversion of useful electrical power into heat or other unwanted
forms of energy in the hardware of a wireless system. It typically occurs in transmitters,
receivers, amplifiers, and circuit components due to internal resistance or inefficiency.
Power Attenuation
Power attenuation is the reduction in signal power as the wave propagates through the
communication channel. It is caused by path loss, material absorption, diffraction, and scattering.
It affects the strength of the received signal, often requiring amplification or error correction
techniques.

Feature Power Dissipation Power Attenuation

Definition Energy lost as heat in a system Reduction in signal power over
distance
Cause Resistance, inefficiencies in Path loss, absorption, reflection,
components scattering
Where it Occurs Electrical circuits, resistors, power Wireless channels, optical fibers,
supplies coaxial cables
Measurement Watts (W) Decibels (dB)
Can be No (Irreversible) Yes (Amplification can compensate)
Recovered?

Page 1 of 36
Fading
Fading is the fluctuation in the amplitude, phase, or frequency of a received wireless signal over
time or space, primarily due to multipath propagation and mobility. Types include Rayleigh
fading, Rician fading, and shadow fading, and it can be fast/slow or flat/frequency-selective.
Cause:
• Multipath Propagation: The transmitted signal reaches the receiver via multiple paths due to
reflections, diffractions, and scattering from obstacles. These multiple copies arrive at the
receiver with different delays and phases, leading to constructive or destructive interference.
• Doppler Shift: Relative motion between the transmitter, receiver, and/or obstacles causes a
shift in the frequency of the received signal, which can lead to phase variations and affect the
received signal amplitude over time.
Effect: Rapid and unpredictable changes in signal strength at the receiver, which can cause deep
fades where the signal power drops significantly below the receiver sensitivity threshold, leading
to communication link failures.
Components: A phenomenon arising from the interaction of the transmitted signal with the
environment (obstacles, moving objects).
Necessity of Mitigation: Fading severely degrades the performance of wireless systems.
Techniques like diversity (space, frequency, time), equalization, and channel coding are
employed to combat its effects.
Relation to Other Concepts: Fading describes the dynamic variations in the received signal
strength, which is initially affected by path loss and attenuation. Multipath propagation, leading
to reflection, diffraction, and scattering, is a primary cause of fading.
Types of Fading:
• Small-Scale Fading (Fast Fading): Rapid fluctuations over short distances or time intervals,
caused by multipath and Doppler shift. Examples include Rayleigh fading (no dominant
line-of-sight path) and Rician fading (dominant line-of-sight path exists).
• Large-Scale Fading (Slow Fading): Slower variations in the average signal strength over
larger distances or time scales, influenced by path loss and shadowing (obstructions).
Reflection
Reflection occurs when a wireless signal encounters a large surface (like a wall or building) and
bounces back rather than passing through. This causes multiple copies of the signal to arrive at
the receiver from different paths, contributing to multipath fading.
Cause: Change in the electrical properties (permittivity, permeability, conductivity) of the
medium at the boundary. Smooth, conductive surfaces are strong reflectors.
Effect: Creates multiple paths for the signal to reach the receiver (multipath propagation),
contributing to fading. Reflected signals can interfere constructively or destructively with the
direct path signal.
Components: Occurs at surfaces like buildings, walls, the ground, and other large objects in the
environment.

Page 2 of 36
Relation to Other Concepts: Reflection is a key mechanism leading to multipath propagation,
which in turn causes fading. The strength of the reflected signal depends on the incident angle
and the reflective properties of the surface, which also contribute to signal attenuation (some
energy is absorbed during reflection).
Absorption
Absorption is the process by which part of the signal energy is absorbed by the material it passes
through, such as walls, trees, or the atmosphere. It reduces the overall signal strength,
contributing to attenuation.
Cause: Interaction of the electromagnetic field with the molecules of the medium, causing them
to vibrate and generate heat. The amount of absorption depends on the frequency of the wave
and the properties of the medium (e.g., water content in air, density of materials).
Effect: Reduces the power of the transmitted signal, contributing to signal attenuation.
Absorption is more significant at higher frequencies (e.g., millimeter waves are strongly
absorbed by atmospheric gases and rain).
Components: The medium itself (atmosphere, building materials, foliage).
Relation to Other Concepts: Absorption is a direct cause of power attenuation. It influences the
overall path loss experienced by the signal.
Diffraction
Diffraction is the bending of radio waves around sharp edges or obstacles. It allows signals to
propagate into regions that are not in the direct line-of-sight (LOS) of the transmitter, such as
behind buildings or terrain.
Cause: According to Huygens' principle, every point on a wavefront can be considered as a
source of secondary spherical wavelets. When a wavefront encounters an obstacle, the wavelets
from the unobstructed portions propagate, effectively bending the wave around the obstacle's
edges.
Effect: Allows wireless signals to propagate beyond the line-of-sight path, enabling
communication in scenarios where there are obstructions. However, the diffracted signal
typically has a lower power than the direct path signal.
Components: Edges and corners of obstacles (buildings, hills, etc.), apertures (windows,
doorways).
Relation to Other Concepts: Diffraction is another mechanism contributing to multipath
propagation, especially in urban environments. It helps in extending the coverage area beyond
direct line-of-sight but also contributes to signal attenuation.

Scattering
Scattering occurs when the signal strikes small irregular objects (e.g., foliage, rough surfaces,
rain droplets), causing the wave to scatter in many directions. This leads to rapid signal
fluctuations and affects signal phase and amplitude.

Page 3 of 36
Cause: Interaction of the wave with small particles or rough surfaces (e.g., dust particles,
raindrops, leaves on trees, rough terrain). Each particle or irregularity acts as a secondary
radiator, emitting waves in various directions.
Effect: Weakens the main signal propagating towards the receiver as energy is dispersed in other
directions. Contributes to signal attenuation and can also contribute to multipath fading if the
scattered waves reach the receiver with different delays.
Components: Small objects or irregularities in the environment.
Relation to Other Concepts: Scattering is another factor contributing to signal attenuation and
multipath propagation, especially at higher frequencies where smaller objects become
comparable to the wavelength.
Feature Diffraction Scattering
Definition Bending of waves around obstacles Dispersion of waves in multiple directions
Cause Large obstacles with sharp edges Small objects or rough surfaces
Effect on Allows signals to reach shadowed Causes multipath propagation and fading
Signal areas
Dominant in Lower frequencies Higher frequencies
Mathematical Knife-edge diffraction, Fresnel Rayleigh, Mie, or Raman scattering
Model zones
Example Signal bending around a building Signal dispersion due to rain or foliage

Path Loss
Path loss is the reduction in signal power as it travels through space from the transmitter to the
receiver, even in the absence of obstacles. It is a function of distance, frequency, and
environment. Path loss is often modeled using empirical or analytical models such as Free Space
Path Loss (FSPL), Okumura, Hata, or COST-231.
Term Definition Internal/External Related To Characteristics
Power Power lost as heat Internal Hardware design Device-related
Dissipation in circuits
Power Signal weakening External Path loss Channel-related
Attenuation in the channel
Fading Signal fluctuation External Reflection, Time-varying
due to multipath scattering
Reflection Signal bounces off External Causes multipath Specular path
surfaces
Absorption Signal energy External Attenuation Converts energy
absorbed by to heat
materials
Diffraction Signal bends External Path coverage Enables NLOS
around edges
Scattering Random External Fading Irregular
redirection by dispersion
small objects
Path Loss Total signal loss External Composite of all Predictable loss
over distance impairments

Page 4 of 36
Introduction to the Okumura Model
The Okumura Model is an empirical radio propagation model used to predict path loss in urban,
suburban, and rural environments. It is widely used for designing mobile communication
systems, especially in the 150 MHz to 1920 MHz frequency range with link distances of 1 km to
100 km.
Characteristics of the Okumura Model
• Empirical Model – Based on extensive field measurements conducted in Japan.
• Applicable for – Urban, suburban, and open rural environments.
• Covers Wide Frequency Range – 150 MHz to 1920 MHz (extended up to 3 GHz).
• Valid for Large Cell Systems – Used for macrocells with distances up to 100 km.
Okumura Model Equation
The path loss (Lp) in dB is given by:
Lp(dB)=Lf+Amu(f,d)−G(hte)−G(hre)−Garea

where:
• Lf = Free space path loss (dB)
Lf = 32.4+20log10(d)+20log10(f)

• Amu(f,d) = Median attenuation relative to free space (obtained from Okumura curves).
• G(hte) = Base station antenna height gain factor (dB).
• G(hre) = Mobile station antenna height gain factor (dB).
• Garea = Gain due to the type of environment (urban, suburban, rural).
Key Parameters
Parameter Definition
F Carrier frequency (MHz)
d Distance between transmitter and receiver (km)
hte Transmitter antenna height (m)
hre Receiver antenna height (m)
G(hte) Gain factor for base station height
G(hre) Gain factor for mobile station height

Okumura Model Corrections

Since the Okumura model is empirical, corrections are applied to improve accuracy.

Antenna Height Gain Correction

For a base station antenna height (hte) between 30 m and 1000 m,
𝒉𝒕𝒆
G(hte) = 20log10(𝟐𝟎𝟎 ), 30m ≤ hte ≤ 1000m

For a mobile antenna height (hre) upto 10 m,

𝒉
G(hre) = 10log10( 𝟑𝒓𝒆 ), hre ≤ 3m
𝒉𝒓𝒆
G(hre) = 20log10( ), 3m ˂ hre ≤ 10m
𝟑

Page 5 of 36
Area Correction Factor (Garea)
• Urban Areas: No additional correction.
• Suburban Areas:
𝑭
Garea = −2(log10𝟐𝟖)2−5.4
• Rural Areas:
Garea = −4.78(log10F)2+18.33log10F−40.94

Mathematical Problem 01:

A 900 MHz mobile communication system operates in an urban environment. The base station is
located 50 km from the mobile receiver. The transmitter and receiver antenna heights are 100 m
and 3 m, respectively. Calculate the path loss.

Solution:
Given,
Carrier Frequency, F = 900 MHz
Distance between transmitter and receiver, d = 50 km
Transmitter antenna height, hte = 100 m
Receiver antenna height, hre =3m
Area Correction Factor, Garea = 0 dB

Let us assume,
Lf = Free Space Path Loss
Amu(f,d) = Median attenuation relative to free space (obtained from Okumura curves)
G(hte) = Gain factor for base station height
G(hre) = Gain factor for mobile station height
Lp = Path Loss

We know,
Lf = 32.4 + 20log10(d) + 20log10(f)
= 32.4 + 20log10(50) + 20log10(900)
= 32.4 + 20 × 1.7 + 20 × 2.95
= 32.4 + 34 + 59
= 125.4 dB

For 900 MHz and 50 km:

Amu(f,d) ≈ 43 dB (from Okumura graphs).

Page 6 of 36
Base station antenna gain (G(hte)):
ℎ𝑡𝑒
G(hte) = 20log10(200 )
100
= 20log10(200)
= 20log10(0.5)
= -6 dB

Mobile antenna gain (G(hre)):

ℎ
G(hre) = 10log10( 3𝑟𝑒 )
3
= 10log10( )
3
= 10log10(1)
= 0 dB

Path loss Lp:

Lp = Lf + Amu(f,d) − G(hte) − G(hre) − Garea
= 125.4 + 43 − (−6) − 0 − 0
= 125.4 + 43 + 6
= 174.4 dB

Answer: Lp = 174.4 dB

Page 7 of 36
Mathematical Problem 02:
A 900 MHz mobile communication system operates in a suburban environment. The base
station is located 50 km from the mobile receiver. The transmitter and receiver antenna heights
are 100 m and 10 m, respectively. Calculate the path loss.
Answer: 155.04 dB (Do it youself.)

Advantages and Disadvantages of the Okumura Model

Advantages
• Based on real-world measurements → More accurate for practical scenarios.
• Suitable for different terrains (urban, suburban, rural).
• Useful for large cell planning (1 km to 100 km).
• Simple and easy to use.
Disadvantages
• Empirical in nature → Lacks deep theoretical foundation.
• Limited frequency range → Can't be directly applied to frequencies beyond 3 GHz.
• Interpolation required → Values from Okumura curves must be manually extracted.
• Accuracy → Less accurate for rural or open areas compared to urban environments.

Applications of the Okumura Model

Mobile Network Planning – Used for designing cellular networks.
Urban Communication – Helps estimate path loss in cities with high buildings.
Rural Area Deployment – Determines signal coverage in open environments.
Satellite Communication – Can be modified for higher frequency applications.
Probability Density Function
A Probability Density Function (PDF) is a function that describes the relative likelihood of a
continuous random variable taking on a specific value. Unlike probabilities for discrete variables,
the probability of a continuous variable taking any exact value is zero; instead, we compute the
probability over an interval.
𝒃
𝑷(𝒂 ≤ 𝒙 ≤ 𝒃) = ∫ 𝒇(𝒙)𝒅𝒙
𝒂
Where:
• f(x) is the PDF of the continuous random variable X,
• a and b are real numbers.

Properties of PDF:
• Non-negativity: 𝒇(𝒙) ≥ 0
∞
• Total Area = 1: ∫−∞ 𝒇(𝒙)𝒅𝒙 = 1
This means the total probability over the entire range of values is 1.

Examples:
1. Uniform Distribution (Continuous): 𝒇(𝒙) = 𝟏
𝒃−𝒂
𝒇𝒐𝒓 𝒂≤𝒙≤𝒃
(𝒙−𝝁)𝟐
−
2. Normal Distribution: 𝒇(𝒙) = 𝟏
√𝟐𝝅𝝈𝟐
𝒆 𝟐𝝈𝟐 [Where μ = mean and 𝜎2= variance]
3. Exponential Distribution: 𝒇(𝒙) =λ𝑒 −λx [Where λ=Rate Parameter]

Page 8 of 36
Importance of PDF
Modeling uncertainty: PDFs are essential in describing uncertainty in natural processes (e.g.,
noise, signal fluctuations).
Machine Learning: Used in generative models, Bayesian inference, likelihood estimation, etc.
Physics & Engineering: In modeling variables like energy distribution, decay, or time between
events.

Introduction to Rayleigh Fading

Rayleigh fading is a statistical model used to describe the effect of multipath propagation in
wireless communication, where there is no direct Line-of-Sight (LOS) between the transmitter
and receiver. Instead, the received signal is a sum of many reflections arriving from different
paths with random phase and amplitude.

Key Characteristics of Rayleigh Fading:

• No LOS Component – Only scattered multipath components.
• Applicable in Urban Areas – Used in city environments with buildings and obstacles.
• Multipath Effect – Signal strength varies due to constructive and destructive interference.
• Power Distribution – Follows an exponential distribution.
• Amplitude Distribution – Follows a Rayleigh distribution.

Applications of Rayleigh Fading Model

• Urban Cellular Networks – Used to model signal strength variations in cities.
• Wi-Fi and Bluetooth – Signal fluctuations in indoor environments.
• Satellite Communications – When direct LOS is blocked by atmospheric conditions.
• Vehicle-to-Vehicle (V2V) Communication – Used for modeling wireless communication
between moving vehicles.

Let r be the envelope of the received signal. Then, the Rayleigh distribution for the signal is:
𝟐
𝒓 −𝒓𝟐 , 𝒓≥𝟎
𝑓𝑅𝑎𝑦𝑙𝑒𝑖𝑔ℎ (𝑟) = 𝒆 𝟐𝝈
𝝈𝟐
Where:
• 𝜎 2 is the variance of the in-phase and quadrature components
• r is the received signal envelope

The average power is: E[𝑟2] = 2 𝜎 2

Page 9 of 36
Problem: Given that the signal follows a Rayleigh distribution with σ=1, find:
a) The probability that the signal amplitude r is less than 1.5.
b) The average power of the signal.
Solution:
a) Compute P(r<1.5):
1.5
𝑟 − 𝑟22 1.5
𝑟 −𝑟 2 1.5
−
𝑟2
−
𝑟2
−
1.52
𝑃(𝑟 < 1.5) = ∫ 𝑒 2𝜎 𝑑𝑟 = ∫ 𝑒 2 𝑑𝑟 = ∫ 𝑟𝑒 2 𝑑𝑟 = 1 − 𝑒 2 = 1 − 𝑒 2 = 0.675
0 𝜎2 0 12 0
b) Average Power: E[𝑟2] = 2 𝜎 2 = 2 × 12 = 2

Introduction to Rician fading

Rician fading describes a type of radio signal propagation where a dominant Line-of-Sight
(LOS) component exists along with multiple reflected and scattered paths. It models situations
where the received signal consists of one strong direct wave and many weaker scattered waves
caused by obstacles like buildings, trees, or vehicles.

Key Characteristics of Rician Fading:

• LOS Component – Present and dominant
• Statistical Model – Includes both Rayleigh-like scattered paths and a strong direct path
• Envelope Distribution – Rician distribution
• PDF Shape – Skewed due to dominant signal path
• K-Factor – Determines the strength of the LOS path
• Diversity Requirement – Less severe than Rayleigh, but still benefits from diversity
techniques
• Fading Severity – Less severe than Rayleigh fading

Applications of Rician Fading Model:

• Mobile networks in open or semi-open areas (e.g., roads, rural areas with LOS)
• Millimeter-wave (mmWave) systems, where LOS links dominate
• Satellite-to-ground links, especially in urban areas where some direct path is maintained
• Wi-Fi or Bluetooth in large indoor environments (airports, stations) where LOS is likely
• Vehicle-to-Infrastructure (V2I) Communication where base stations are visible to moving
vehicle
The Rician Probability Density Function (PDF) for the received signal envelope r is given by:
2 2
𝑟 (−𝑟 +𝐴2 ) 𝐴𝑟
𝑓𝑅𝑖𝑐𝑖𝑎𝑛 (𝑟) = 2 𝑒 2𝝈 𝐼0 ( 𝟐 ) , 𝑟≥0
𝝈 𝝈
Where:
A = Amplitude of the direct LOS path
2
σ = Variance of the multipath components
I0(⋅) = Modified Bessel function of the first kind, order zero
r = Signal envelope

Page 10 of 36
 𝐴2
Rician K-Factor: 𝐾 = 2𝝈2
• Measures the ratio of power in the direct path (LOS) to the scattered/multipath components
• K=0 → Rayleigh fading
• Higher K → Stronger LOS, closer to non-fading or AWGN channel

Problem:
Given a Rician fading channel with A=2, σ=1, compute:
a) The probability density at r=2
b) The Rician K-factor
Solution:
a) Compute PDF at r=2:
2 2
2 (−2 +22 ) 2×2
𝑓𝑅𝑖𝑐𝑖𝑎𝑛 (2) = 2 e 2×𝟏 𝐼0 ( 𝟐 ) = 2𝑒 −4 𝐼0 (4) = 2 × 0.0183 × 11.3019 = 0.414
𝟏 𝟏
b) Compute K:
𝐴2 4
K= = =2
2𝜎 2 2 × 12

Digital data in its original form is a baseband signal. For wireless transmission, which typically
occurs at higher frequencies, baseband signals need to be modulated onto a carrier signal to shift
their frequency spectrum to the desired transmission band called as passband.
Digital modulation is the process of converting digital data (bits) into a suitable analog signal
for transmission over a physical communication channel. The choice of modulation scheme
significantly impacts the performance of a wireless communication system in terms of bandwidth
efficiency, power efficiency, complexity, and robustness to channel impairments.

Page 11 of 36
Amplitude Shift Keying (ASK), Frequency Shift Keying (FSK), and Phase Shift Keying (PSK)
are fundamental digital modulation techniques where the information bits are encoded by
varying the amplitude, frequency, or phase of a carrier signal, respectively. These schemes form
the basis for understanding more complex modulation techniques and are still relevant in various
communication systems.

1. Amplitude Shift Keying (ASK):

• Principle: In ASK, the amplitude of the carrier signal is varied to represent the digital
data. Typically, one amplitude level represents a binary '1', and another (usually zero
amplitude) represents a binary '0' (On-Off Keying - OOK). M-ary ASK uses more than
two amplitude levels to transmit multiple bits per symbol.
• Binary ASK (OOK):
o Transmits a carrier signal 𝐴𝑐𝑜𝑠(2𝜋𝑓𝑐 𝑡) for a bit '1' and no signal for a bit '0'.
o Mathematical Representation:
𝐴𝑐𝑜𝑠(2𝜋𝑓𝑐 𝑡), 𝑓𝑜𝑟 𝑏𝑖𝑛𝑎𝑟𝑦 ′1′
𝑠(𝑡) = {
0, 𝑓𝑜𝑟 𝑏𝑖𝑛𝑎𝑟𝑦 ′0′
o Signal Constellation: Two points on the real axis, at 0 and A.
o Bandwidth: Approximately equal to the bit rate (B≈fb).
o Demodulation: Can be demodulated coherently (requires a synchronized local
oscillator) or non-coherently (using envelope detection). Non-coherent detection is
simpler but suffers a performance penalty.

Page 12 of 36
 • M-ary ASK:
o Uses M=2m different amplitude levels to transmit m bits per symbol.
o Signal Constellation: M points equally spaced along the real axis.
o Bandwidth: Approximately equal to the symbol rate (B≈fs=fb/m). Bandwidth
efficiency increases with m.
o Power Efficiency: Decreases with increasing M as the distance between adjacent
amplitude levels reduces.
• Advantages: Simple to implement, especially for binary ASK with non-coherent
detection.
• Disadvantages: Least power-efficient among ASK, FSK, and PSK, highly susceptible to
amplitude variations and fading in wireless channels.

2. Frequency Shift Keying (FSK):

• Principle: In FSK, the frequency of the carrier signal is varied to represent the digital
data. Binary FSK uses two distinct frequencies, f1and f2, to represent binary '0' and '1'. M-
ary FSK uses M different frequencies.
• Binary FSK:
o Transmits one of two sinusoidal signals with different frequencies.
o Mathematical Representation:
𝐴𝑐𝑜𝑠(2𝜋𝑓1 𝑡), 𝑓𝑜𝑟 𝑏𝑖𝑛𝑎𝑟𝑦 ′1′
𝑠(𝑡) = {
𝐴𝑐𝑜𝑠(2𝜋𝑓2 𝑡), 𝑓𝑜𝑟 𝑏𝑖𝑛𝑎𝑟𝑦 ′0′
o Signal Constellation: Not typically represented by a simple constellation diagram as
the frequency is the distinguishing feature.
o Bandwidth: Approximately ∣f1−f2∣+2fb, where fb is the bit rate. The required
bandwidth depends on the frequency separation. Minimum Shift Keying (MSK) is a
spectrally efficient form of binary FSK.
o Demodulation: Can be demodulated coherently or non-coherently. Non-coherent
detection using bandpass filters followed by envelope detectors is common.
• M-ary FSK:
o Uses M different frequencies to transmit m bits per symbol. The frequencies are
1
typically separated by 2𝑇 to maintain orthogonality, where Ts is the symbol duration.
𝑠
o Bandwidth: Increases linearly with M. Bandwidth efficiency decreases with
increasing M.
o Power Efficiency: Generally better than ASK and can be comparable to PSK for
lower values of M. Constant envelope property makes it less susceptible to non-
linearities.
• Advantages: More robust to amplitude variations and fading compared to ASK due to
the constant envelope. Non-coherent detection is relatively simple.
• Disadvantages: Bandwidth efficiency is generally lower compared to PSK and QAM.

Page 13 of 36
3. Phase Shift Keying (PSK):
• Principle: In PSK, the phase of the carrier signal is varied to represent the digital data.
The amplitude and frequency remain constant.
• Binary PSK (BPSK):
o Two phases, typically 0∘ and 180∘, represent binary '0' and '1'.
o Mathematical Representation:
𝐴𝑐𝑜𝑠(2𝜋𝑓𝑐 𝑡), 𝑓𝑜𝑟 𝑏𝑖𝑛𝑎𝑟𝑦 ′1′
𝑠(𝑡) = {
𝐴𝑐𝑜𝑠(2𝜋𝑓𝑐 𝑡 + 𝜋) = −𝐴𝑐𝑜𝑠(2𝜋𝑓𝑐 𝑡), 𝑓𝑜𝑟 𝑏𝑖𝑛𝑎𝑟𝑦 ′0′
o Signal Constellation: Two points on the real axis, at +A and −A.
o Bandwidth: Approximately equal to the bit rate (B≈fb).
o Demodulation: Requires coherent detection as the information is in the phase.
• Quadrature Phase Shift Keying (QPSK):
o Four phases, typically 45∘, 135∘, 225∘, and 315∘, represent two bits per symbol.
o Signal Constellation: Four points equally spaced on a circle in the complex plane.
o Bandwidth: Approximately half the bit rate (B≈fb/2), offering better bandwidth
efficiency than BPSK.
o Demodulation: Requires coherent detection.
• M-ary PSK (M-PSK):
o M=2m phases, equally spaced between 0 and 2π, represent m bits per symbol.
o Signal Constellation: M points equally spaced on a circle in the complex plane.
o Bandwidth: Approximately equal to the symbol rate (B≈fs=fb/m). Bandwidth
efficiency increases with m.
o Power Efficiency: Decreases with increasing M as the phase difference between
adjacent symbols reduces.
• Advantages: More power-efficient than ASK. For higher orders (M-PSK), it offers better
bandwidth efficiency than M-FSK. Constant envelope is advantageous for non-linear
power amplifiers.
• Disadvantages: Requires coherent detection, which is more complex than non-coherent
detection. Performance degrades rapidly in the presence of phase noise.

Comparative Discussion:
The choice between ASK, FSK, and PSK depends on the specific requirements of the
communication system. Here's a comparison based on key parameters:

Feature ASK FSK PSK

Low (Binary ASK) to Low (Binary FSK) to Moderate (BPSK,
Bandwidth
Moderate (M-ary Very Low (M-ary QPSK) to High (M-
Efficiency
ASK) FSK) PSK)
Simple (non-coherent Simple (non-coherent Moderate (BPSK,
Complexity binary ASK) to binary FSK) to QPSK) to
(Transceiver) Moderate (M-ary Moderate (M-ary Moderate/High (M-
ASK, coherent) FSK, coherent) PSK, coherent)

Page 14 of 36
 Power Efficiency Lowest Moderate Highest
Robustness to
Robust (constant Robust (constant
Amplitude Highly Susceptible
envelope) envelope)
Variations/Fading
Highest (for the same
Robustness to Noise Lowest Moderate bit error rate and
bandwidth)
No (except for OOK
Constant Envelope being effectively Yes Yes
on/off)
Low to medium data
Simple low-data-rate Satellite
rate systems,
systems, optical fiber communication,
Applications telemetry, some
communication wireless LANs, digital
wireless sensor
(OOK) cellular systems
networks

Detailed Comparative Points:

• Power Efficiency: For achieving the same bit error rate, PSK generally requires the least
transmitted power, followed by FSK, and then ASK. This is because information in PSK
and FSK is conveyed by changes in phase or frequency (constant envelope), making them
less susceptible to amplitude noise.
• Bandwidth Efficiency: For binary modulation, ASK, FSK, and PSK have similar
bandwidth requirements. However, for M-ary signaling (M>2), PSK and ASK achieve
better bandwidth efficiency than FSK (bits/symbol/Hz). QPSK is particularly attractive as
it doubles the bandwidth efficiency compared to BPSK.
• Complexity: Non-coherent detection of binary ASK and FSK is relatively simple.
Coherent detection, required for PSK and higher-order ASK/FSK for optimal
performance, adds complexity to the receiver.
• Channel Conditions: In channels experiencing significant amplitude variations or fading
(common in wireless environments), FSK and PSK are preferred over ASK due to their
constant envelope property. However, PSK is more susceptible to phase noise and
requires accurate phase synchronization.
• Applications: The choice of modulation scheme is dictated by the specific application
requirements. For example, simplicity might be prioritized in low-cost sensor networks
(potentially favoring non-coherent ASK or FSK), while bandwidth efficiency is crucial in
high-data-rate wireless systems (favoring QPSK or higher-order PSK/QAM).

Inter Symbol Interference (ISI): In frequency-selective fading channels, the time-dispersed

versions of a transmitted symbol can interfere with subsequent symbols, making it difficult for
the receiver to correctly decode the intended information. The delay spread of the channel is a
key factor determining the severity of ISI.

Page 15 of 36
In the above figure, the transmitter transmits a single symbol. This symbol is propagated along
different paths (A, B and C), and eventually reaching the receiver at multiple time instants, and
therefore with multiple “replication.”

Equalization Techniques:
Equalization is a signal processing technique employed at the receiver to mitigate the effects of
channel-induced distortion, primarily ISI caused by frequency-selective fading. The goal of an
equalizer is to estimate the inverse channel characteristics and apply it to the received signal to
recover the transmitted symbols.
• Linear Equalizers: These equalizers use linear filters to process the received signal.
o Zero-Forcing (ZF) Equalizer: The ZF equalizer aims to completely eliminate
ISI by forcing the channel impulse response convolved with the equalizer's
impulse response to be a Dirac delta function at the desired sampling instant.
However, it can significantly amplify noise, especially when the channel has deep
fades.
▪ Mathematical Representation: Let H(f) be the channel frequency
response and E(f) be the equalizer frequency response. The ZF condition
is H(f)E(f)=1, or E(f)=1/H(f).
o Minimum Mean Square Error (MMSE) Equalizer: The MMSE equalizer aims
to minimize the mean square error between the transmitted and the equalized
received signal. It considers the presence of noise and achieves a better trade-off
between ISI reduction and noise amplification compared to the ZF equalizer.
▪ Mathematical Representation: The MMSE equalizer's frequency
𝐻 ∗(𝑓)
response is given by 𝐸𝑀𝑀𝑆𝐸 (𝑓) = 𝑁 , where H∗(f) is the conjugate
|𝐻(𝑓)|2 + 0
𝐸𝑠

of H(f), N0 is the noise power spectral density, and Es is the transmitted

symbol energy.

Page 16 of 36
 • Non-Linear Equalizers: These equalizers employ non-linear processing to combat ISI,
often offering better performance than linear equalizers in severely distorted channels.
o Decision Feedback Equalizer (DFE): The DFE uses past decisions on the
transmitted symbols to estimate and cancel the ISI affecting the current symbol. It
consists of a feed-forward filter and a feedback filter. The feedback filter utilizes
the detected symbols to reconstruct the ISI they cause on subsequent symbols and
subtracts this estimate from the received signal before the current symbol is
detected. The performance of the DFE heavily relies on the accuracy of the past
decisions; error propagation can be a significant issue.
o Maximum Likelihood Sequence Estimation (MLSE): MLSE is an optimal
equalization technique that considers all possible transmitted symbol sequences
and chooses the sequence that is most likely to have produced the received signal.
It typically uses the Viterbi algorithm for efficient implementation. While offering
the best performance, MLSE has a computational complexity that grows
exponentially with the length of the channel impulse response, making it
impractical for channels with long delay spreads and large signal constellations.
• Adaptive Equalization: In time-varying wireless channels, the channel characteristics
change over time. Adaptive equalizers continuously estimate the channel and adjust their
filter coefficients to track these changes. Algorithms like Least Mean Squares (LMS) and
Recursive Least Squares (RLS) are commonly used for adaptive equalization.

Diversity Techniques:
Diversity techniques exploit the statistical independence of multiple versions of the transmitted
signal received through different paths or using different resources to improve the reliability of
communication. The fundamental idea is that if one version of the signal experiences a deep
fade, other independently faded versions are likely to be strong.
• Space Diversity: Multiple antennas are used at the transmitter, receiver, or both.
o Receive Diversity: Multiple receive antennas are used to receive multiple faded
replicas of the transmitted signal. These replicas are then combined to improve the
signal-to-noise ratio (SNR). Common combining techniques include:
▪ Selection Combining (SC): The receiver selects the branch with the
highest instantaneous SNR.
▪ Maximal Ratio Combining (MRC): The signals from all branches are
co-phased and weighted proportionally to their SNR before being
summed. MRC provides the optimal SNR performance.
▪ Equal Gain Combining (EGC): The signals from all branches are co-
phased and then summed with equal weights. EGC offers performance
close to MRC with simpler implementation.
o Transmit Diversity: Multiple transmit antennas are used to transmit multiple
versions of the signal. Techniques include:

Page 17 of 36
 ▪ Space-Time Coding (STC): Encodes the data stream across multiple
antennas and time slots to provide both diversity gain and coding gain.
Examples include Alamouti coding.
▪ Spatial Multiplexing: Transmits independent data streams from different
antennas to increase the data rate. While not strictly a diversity technique,
it leverages multiple antennas.
• Frequency Diversity: The signal is transmitted on multiple carrier frequencies that are
sufficiently separated so that they experience independent fading. Techniques include:
o Frequency Hopping Spread Spectrum (FHSS): The carrier frequency is rapidly
changed according to a pseudorandom sequence.
o Multicarrier Modulation (MCM): The data stream is divided into multiple
lower-rate substreams, each modulating a different subcarrier. If the subcarrier
spacing is appropriately chosen, each subcarrier experiences flat fading, even if
the overall channel is frequency-selective. Orthogonal Frequency Division
Multiplexing (OFDM) is a prominent example of MCM.

• Time Diversity: The signal is transmitted at different time instants separated by more
than the coherence time of the channel. Techniques include:
o Repetition Coding: The same information is transmitted multiple times.
o Interleaving and Error Correction Coding: Interleaving spreads the effects of
burst errors caused by deep fades over multiple codewords, allowing the error
correction code to effectively correct them.
• Polarization Diversity: Transmitting and receiving signals with different polarizations
(e.g., vertical and horizontal) can provide independent fading paths.

Advanced Equalization and Diversity Techniques:

• MIMO (Multiple-Input Multiple-Output) Systems: Employ multiple antennas at both
the transmitter and the receiver. MIMO can offer significant improvements in capacity
and reliability through spatial multiplexing and spatial diversity techniques. Equalization
in MIMO systems is more complex due to the multiple parallel channels. Techniques like
zero-forcing (ZF), minimum mean square error (MMSE), and successive interference
cancellation (SIC) are used for equalization in MIMO systems.
• OFDM with Equalization and Diversity: OFDM inherently converts a frequency-
selective fading channel into multiple parallel flat-fading subchannels, simplifying
equalization. Simple one-tap equalizers can be used for each subcarrier. OFDM can also
be combined with various diversity techniques like space-time coding and frequency
diversity.
• Cognitive Radio and Adaptive Techniques: In cognitive radio systems, the transceiver
can sense the channel conditions and dynamically adapt its transmission parameters,
including the choice of equalization and diversity techniques, to optimize performance.

Page 18 of 36
 • Machine Learning for Channel Equalization and Prediction: Recent research
explores the use of machine learning algorithms for channel estimation, equalization, and
prediction in highly dynamic wireless environments. Neural networks and other machine
learning models can learn complex channel characteristics and potentially outperform
traditional equalization techniques in certain scenarios.

Telecommunication Network
One can envision a telecommunications network as a kind of medium interconnecting a
collection of devices that is equipped to exchange information between them. The most common
information sources are voice, data, and video. Traditional communication networks have
evolved around applications using one of these specific information sources. The future of
communications networks, however, is directed toward multimedia services supporting
applications using a combination of all these information sources. The entire communications
network consists of two types of elements, the communicating devices and the network
infrastructure. The communicating devices provide an interface between the user application and
the network infrastructure. These devices are usually referred to as communication stations,
terminals, or hosts. The network infrastructure is a collection of point-to- point wired or wireless
links and a number of switches and routers interconnecting several of these communication
terminals in geographically separated locations. Traditional communication devices are
connected to the communications network through a fixed connection point. The geographic
location of the terminal and its connection to the infrastructure remains fixed. Figure 1 shows a
simple traditional wired network with examples of its elements.

Figure 1: Traditional Wired Networks

Page 19 of 36
The future direction of communications networks is toward wireless mobile connections to the
infrastructure that shall support continual connection though the geographical location of the
terminal is constantly changing due to the mobility of the terminal.

In order to modify the traditional fixed network infrastructures to support wireless connections, a
new wireless infrastructure is needed as an interface between the backbone wired network
infrastructure and the mobile communication terminals.

The mobile communication terminals need to be equipped with wireless front-ends to

communicate with the wired backbone through the new wireless infrastructure. Figure 2 shows
the positioning of the wireless infrastructure in relation to the wired infrastructure.

In addition to switches, routers, and point-to-point links, the wireless network infrastructure also
needs wireless transceivers to communicate with the wireless communication terminals and act
as points of access to the fixed part of the wire- less network infrastructure. These transceivers
are referred to as base stations (BSS) or access points (APs). Any wireless base station has a
limited coverage area. If the coverage area is less than the desirable coverage area for the
wireless service, we need multiple base stations to cover the service area. In the case of multiple
base stations operating in an area, the wireless network infrastructure needs to co- ordinate the
continuity of a wireless connection as the mobile communication terminal moves through the
coverage areas of different base stations.

Figure 2: Positioning of the wireless network infrastructure in relation to the wired network infrastructure.

Wireless Network Topologies

We refer to wireless network topology as the configuration in which a mobile terminal communicates
with another. There are two fundamental types of topologies used in wireless networks. They are
infrastructure, centralized, or hub-and-spoke topology, and the ad hoc or distributed topology.

Page 20 of 36
Infrastructure Network Topology

In the infrastructure topology, there is a fixed (wired) infrastructure that supports communication between
mobile terminals and between mobile and fixed terminals. The infrastructure networks are often designed
for large coverage areas and multiple base station or access point operations. Most of the discussion in
this section is around this type of operation. Figure 3 shows the basic operation of an infrastructure
network with a single BS/AP. The BS/AP serves as the hub of the network, and the mobile terminals are
located at the ends of the spokes. Any communication from one wireless user station to another, that is,
between peers, has to be sent through the BS/AP. The hub station usually controls the mobile stations and
monitors what each station is transmitting. Thus, the hub station is involved in managing user access to
the network.

Figure 3: Basic operation of an infrastructure network topology

Example 1: Systems That Employ Infrastructure Network Topology

All standardized cellular mobile telephone and wireless data systems use an infrastructure
network topology to serve mobile terminals operating within the coverage area of any BS. The
IEEE 802.11 standard and most of the wireless LAN products support infrastructure operation.

Ad Hoc Network Topology

Ad hoc or distributed network topology applies to reconfigurable networks that can operate
without the need for a fixed infrastructure. These networks are primarily used by the military and
also in a few commercial applications for voice and data transmission. Such a topology is
suitable for rapid deployment of a wireless network in a mobile or fixed environment. Figure 4
shows two variations of the ad hoc network topology. Figure 4(a) is a single-hop ad hoc network
where, as the name implies, every user terminal has the functional capability of communicating
directly with any of the other user terminals.

Page 21 of 36
Example 2: Systems That Support Single-Hop Ad Hoc Network Topology

The IEEE 802.11 wireless LAN standard supports single-hop peer-to-peer topology for ad hoc
networking. When a terminal is turned on, it first searches for a beacon signal from an AP or
another terminal announcing the existence of an ad hoc network. If a beacon is not detected, the
terminal takes the responsibility of announcing the existence of an ad hoc network. Several PCS
services, such as PHS and NEXTEL satellite, support peer-to-peer walkie-talkie type
communication among voice terminals.

Figure 4: Ad hoc networking: (a) single hop peer to peer topology and (b) multi hop ad hoc networking topology

In some ad hoc networking applications, where users may be distributed over a wide area, a
given user terminal may be able to reach only a portion of the other users in the network due to
transmitter signal power limitations. In this situation, user terminals will have to cooperate in
carrying messages across the network between widely separated stations. Networks designed to
function this way are called multi-hop ad hoc networks and are illustrated in Figure 4(b).

In an ad hoc multi-hop network, each terminal should be aware of the neighboring terminals in
its coverage range. The multi-hop network configuration was originally used in military tactical
networks, where providing reliable communications under unpredictable propagation conditions
and over widely varying geographic areas was important.

Example 3: Systems That Support Multi-hop Ad Hoc Network Topology

The early packet radio networks studied for military applications in 1970s were employing
multi-hop ad hoc network topology. ETSI BRAN's HIPERLAN standard for wireless LANs that
was developed during the mid-1990s, supports multi-hop ad hoc networking for commercial
applications.

Page 22 of 36
Comparison between Infrastructure and Ad hoc network topologies

Controllability: To coordinate proper operation of a radio network, we need to centrally control

certain features such as time synchronization, transmitted power of the mobile stations operating
in a certain area, and so on. In an infrastructure network, all these features are naturally
implemented in the BS or AP. In an ad hoc network, implementation of these features requires
more complicated structures demanding changes in all terminals.

Scalability: In peer-to-peer single-hop networks, expansion is always limited to the coverage of

the radio transmitter and receiver, and there is no simple way to scale up the network coverage or
capacity (wireless traffic) that can be supported by the network. In multi-hop ad hoc networks, as
the number of terminals increases the potential coverage of the network is increased. However,
the traffic handling capacity of the network remains the same. To connect an ad hoc network to
the backbone wired network, one needs to use a proxy server with a wireless connection as a
member of the ad hoc network. In practice all terminals supporting ad hoc networking operate in
a dual mode that also supports infrastructure operation. Wireless infrastructure networks are
inherently scalable. To scale up a wireless infrastructure network, the number of BSs or APs is
increased to expand the cover- age area or to increase the capacity while using the same available
spectrum. Therefore, for wide area coverage and for applications with variable traffic loads,
infrastructure networks are always used.

Flexibility: Operation of infrastructure networks requires deployment of a network infrastructure

which is very often time-consuming and expensive. Ad hoc networks are inherently flexible and
can be set up instantly. Therefore, ad hoc networks are always used for temporary applications
where flexibility is of prime importance.

Reliability: Another issue of concern in small-scale WLAN operation and military applications
in battlefields is resistance to failure. Infrastructure networks are "single failure point" networks.
If the AP or BS fails, the entire communications network is destroyed. This problem does not
exist in ad hoc peer-to-peer configurations.

Coverage: In WLANs, coverage of the network is an issue of concern because it has an effect on
the selection of the topology. In peer-to-peer single hop network topology, the maximum distance
between two terminals is the range of coverage of the wireless interface used in the terminal. In
an infrastructure network, two wireless terminals communicate through an AP or a BS. The
maximum distance between two terminals is thus twice the range of coverage of a single wireless
modem because the communicating terminals may be located at the edge of the coverage area of
the BS or AP. In practice often APs or BSs are fixed in opportunistic locations using elevated
mountings that increase the coverage of the wire- less modem. This usually results in a
maximum coverage distance between two terminals that is greater than twice the coverage
distance of the same modem in an ad hoc configuration.

Page 23 of 36
Store and Forward Delay and Media Usage Efficiency: In peer-to-peer single-hop networks,
information is transmitted only once, and there is no store and forward procedure. In the
infrastructure topology, we have transmission of data twice, once from the source to the BS/AP
and once from the BS/AP to the destination. The BS/AP also should store the message and
forward it later. This adds to the delay encountered by the data packets. Multi-hop ad hoc
networks may have several transmissions and several store and forward delays that depend on
the instantaneous topology and number of hops required to send the data from the source to the
destination.

Cellular Concepts

By cellular radio, we mean deploying a large number of low-power base stations for
transmission, each having a limited coverage area. In this fashion, the available capacity is
multiplied each time a new base station or transmitter is set up because the same spectrum is
being reused several times in a given area.

The fundamental principle of the cellular concept is to divide the coverage area into a number of
contiguous smaller areas which are each served by its own radio base station. Radio channels are
allocated to these smaller areas in an intelligent way so as to minimize the interference, provide
an adequate performance and cater to the traffic loads in these areas. Each of these smaller areas
is called a cell. Cells are grouped into clusters. Each cluster utilizes the entire available radio
spectrum. The reason for clustering is that adjacent cells cannot use the same frequency spectrum
because of interference. So, the frequency bands have to be split into chunks and distributed
among the cells of a cluster.

The spatial distribution of chunks of radio spectrum (which are called sub-bands) within a cluster
has to be done in a manner such that the desired performance can be obtained. This forms an
important part of network planning in cellular radio. The number of cells in a cluster is called
cluster size or frequency reuse factor.

Two types of interference are important in such a cellular architecture.

1. The interference due to using the same frequencies in cells of different clusters is referred
to as cochannel interference. The cells that use the same set of frequencies or channels
are called cochannel cells.
2. The interference from different frequency channels used within a cluster whose side-
lobes overlap is called adjacent channel interference.

The allocation of channels within the cluster and between clusters must be done so as to
minimize both of these. The cellular concept can increase the number of customers that can be
sup- ported in the available frequency spectrum as illustrated by the following examples by
deploying several low power radio transmitters.

Page 24 of 36
Example 5.4: Cellular Concept

Consider a single high-power transmitter (see Fig. 5) that can support 35 voice channels over an
area of 100 square km with the available spectrum. If seven lower power transmitters are used so
that they support 30 percent of the channels over an area of 14.3 square km each, a total of ≈80
voice channels are now available in this area instead of 35. In reality, channels will have to be
allocated to base stations in such a way as to prevent interference between one base station and
another. In Figure 5, base stations 1 and 4 could use the same channels, as their coverage areas
are sufficiently far apart and so also are base stations 3 and 6. Suppose the cells labeled 1, 2, 5, 6,
and 7 use disjoint frequency bands and the channels used in 1 and 6 are reused in 3 and 4. The
set of cells {1, 2, 5, 6, 7} forms a cluster. Cells 3 and 4 form part of another cluster.

Figure 5: The cellular concept

Example 5: Importance of Cellular Topology

We want to provide a radio communication service to a city. The total bandwidth available is 25
MHz, and each user requires 30 KHz of bandwidth for voice communication. If we use one
antenna to cover the entire town, we can only support 25 MHz/30 KHz=833 simultaneous users.
Now let us employ a cellular topology where 20 lower power antennas are opportunistically
located to minimize both kinds of interference. We divide our frequency band into four sets and
assign one set to each cell. Each cell has a spectrum of 25 MHz/4=6.25 MHz allocated to it. We
have a cluster of four cells in this example. The number of simultaneous users supported per cell
is 6.25 MHz/30 KHz=208. The number of users per cluster is 4×208=832. The total number of
simultaneous users is now 832×5=4160, because we have five clusters of four cells each. The
new capacity is roughly five times the capacity with a single antenna.

Page 25 of 36
Examples 4 and 5 illustrate the main benefits and elements of a cellular network planning by
relating the bandwidth, number of cells, frequency reuse factor, and capacity of the network. If
W is the total available spectrum, B is the bandwidth needed per user, N the frequency reuse
factor and m is the number of cells required to cover an area, the number of simultaneous users is
given by:

𝑚(𝑊/𝑁)
𝑛=
𝐵
In particular we observe that the capacity of the network can be increased by:

a. Increasing m, such as the number of cells (by reducing the size of each cell) and
b. Decreasing the frequency reuse factor.

A cellular topology reduces the coverage requirements of both the mobile terminal and the BS.
The reduction of the size of coverage lowers the required transmitted power by the mobile
terminal because mobile terminals are located closer to the base stations and they require less
power to communicate with the network. This increases the battery lifetime and reduces the size
of a terminal. These issues are extremely important to the user of a handheld terminal. Therefore,
the larger the number of cells, the larger the capacity and the smaller the size of the handheld
terminal. However, we need a fixed network infrastructure to interconnect the cells and ensure
that the entire system works in a coordinated manner. As we increase the number of cells, the
cost and the time for deploying the network increases. In addition, as the cell size becomes
smaller, the number of handoffs increases. Therefore, a reduction in the size of the cells increases
the complexity of the design and deployment of the network, as well as the signaling load in the
fixed part of the infrastructure. Another important factor in deployment of wireless cellular
networks is provision for expansion. The main investment of a wireless service provider is
toward the cost of the fixed infrastructure, which includes the BS and connections between them.
When a service provider starts an operation, that person needs to minimize the cost of the
infrastructure while continuously increasing the number of subscribers. As the number of
subscribers increases, new income is generated and the service provider can afford to expand the
network by increasing the complexity of its infrastructure to support a further larger population
of subscribers. Therefore, there is a need for a plan to take into account the growth of the
subscriber base and thus the entire wireless network.

In summary, we need to address the following technical issues for planning a cellular network:

• Selection of a frequency reuse pattern for different radio transmission techniques

• Physical deployment and radio coverage modeling
• Plans to account for the growth of the network
• Analysis of the relationship between the capacity, cell size and the cost of the
infrastructure

Page 26 of 36
Cell Sizes

In a modern deployment of a cellular network, a number of cell sizes are used to provide a
comprehensive coverage supporting traffic fluctuations in different geographic areas and
supporting a variety of applications. One way of dividing the cells into a hierarchy is to define
the following cell sizes:

• Femtocells: These are the smallest unit of the cellular hierarchy used for connection of
personal equipment such as laptops, notepads, and cellular telephones. These cells need
to cover only a few meters where all these devices are in physical range of the user.
• Picocells: These are small cells inside a building that support local indoor networks such
as wireless LANs. The size of these networks is in the range of a few tens of meters.
• Microcells: These cells cover the inside of streets with antennas mounted at heights lower
than the rooftop of the buildings along the streets. They cover a range of hundreds of
meters and are used in urban areas to support PCS.
• Macrocells: Macrocells cover metropolitan areas, and they are the traditional cells
installed during the early phases of the cellular telephony. These cells cover areas on the
order of several kilometers, and their antennas are mounted above the rooftop of typical
buildings in the coverage area.
• Megacells: Megacells cover nationwide areas with ranges of hundreds of kilo-meters and
are mainly used with satellites.

Co-channel Reuse Ratio

In order to investigate the effects of interference, which changes with distance, there is a need to
come up with an elegant way of determining distances and identifying cells. Fortunately, it is
possible to do this easily in the case of hexagonal cells. In order to maximize the capacity,
cochannel cells must be placed as far apart as possible for a given cluster size. It can be shown
that there are only six cochannel cells for a given reference cell at this distance. The relationship
among the distance between cochannel cells DL, the cluster size N and the cell radius, RL, is
given by:

𝐷𝐿
= √3𝑁
𝑅𝐿

This quantity is also referred to as the cochannel reuse ratio. Values for N can only take on values
of the form i2+ij+j2 where i and j are integers.

Example 6: Cluster Size of N=7

As described earlier, i and j can only take integer values. If we take i=2 and j=1, we see that
N=4+2+1=7. Selecting a cell A, we can determine its cochannel cell by moving two units along
one face of the hexagon and one unit in a direction 60∘ or 120∘ to this direction.

Page 27 of 36
Clusters of size N=7 can be created as shown in Figure 6. A value of N=7 is employed in the
United States in the advanced mobile phone service (AMPS).

Figure 6: Hexagonal cellular architecture with a cluster size on N=7

Tradeoff between the system capacity and performance

The number of cells in a cluster N determines the amount of cochannel interference and also the
number of frequency channels available per cell. Suppose there are Nc channels available for the
entire system. Each cluster uses all the Nc channels. With fixed channel allocation, each cell is
allocated Nc/N channels. It is desirable to maximize the number of channels allocated to a cell.
This means that N should be made as small as possible. However, reducing N increases the
signal-to-interference ratio (as discussed in the following section). There is thus a trade-off
between the system capacity and performance.

Reasons to use a hierarchical cellular infrastructure

There are three reasons to use a hierarchical cellular infrastructure supporting cells of different
sizes.

One is to extend the coverage to the areas that are difficult to cover by a large cell. For example,
cells designed to cover suburban areas have antennas on tall towers and cover a large area.
Signals from these antennas, however, cannot propagate sufficiently into urban canyons or
indoor environments. For urban canyons we need to install antennas at lower heights and in
indoor areas we may mount the antennas on walls to provide a comprehensive coverage.
Antennas mounted in these locations are of low power and cover a smaller area, resulting in the
creation of a smaller sized cell.

Page 28 of 36
The second reason to have a cellular hierarchy is to increase the capacity of the network for those
areas that have a higher density of users. Imagine the number of cellular phone users in the
downtown area of a large city and compare it with the number of mobile users on an interstate
highway. To support the larger subscriber demand and higher traffic in smaller areas, we need to
increase the number of cells by reducing their sizes.

The third reason is that sometime an application needs certain coverage. Consider the increasing
number of wireless devices that we are carrying in our bags these days and the increasing need
for communication between these devices. This necessitates extremely small sized cells that
provide a wireless network for connecting laptops or notepads to cellular phones.

SIR Calculation

A cellular architecture was essential in order to reuse the available spectrum while reducing
interference caused by reusing the frequency spectrum. In this section, we look in detail at the
performance measures that are useful in system design, in particular the signal-to-interference
ratio and its relationship with the path loss and the grade of service. In general, the signal-to-
interference ratio can be written as follows:

𝑃𝑑𝑒𝑠𝑖𝑟𝑒𝑑
𝑆𝑟 =
∑𝑖 𝑃𝑖𝑛𝑡𝑒𝑟𝑓𝑒𝑟𝑒𝑛𝑐𝑒,𝑖

Here 𝑃𝑑𝑒𝑠𝑖𝑟𝑒𝑑 is the signal strength from the desired BS and 𝑃𝑖𝑛𝑡𝑒𝑟𝑓𝑒𝑟𝑒𝑛𝑐𝑒 , i is the signal strength
from the ith interfering BS. The signal strength falls as some power of α the distance called the
power-distance gradient or path-loss gradient. That is, if the transmitted power is Pt, after a
distance d in meters, the signal strength of a radio signal will be proportional to Ptd−α. In its most
simple case, the signal strength falls as the square of the distance in free space (α=2). Suppose
there are two base station transmitters BS1 and BS2 located in an area with the same transmit
power Pt and a mobile terminal is at a distance of d1 from the first and d2 from the second. If the
mobile terminal is trying to communicate with the first BS, the signal from the second BS is
interference. The signal-to-interference ratio for this mobile terminal will be:

𝐾𝑃𝑡 𝑑1−α 𝑑2 α
𝑆𝑟 = =( )
𝐾𝑃𝑡 𝑑2−α 𝑑1
𝑑2
The larger the ratio is, the greater is 𝑆𝑟 and the better is the performance. The objective in a
𝑑1
cellular radio system is to allocate frequencies or channels to cells within a cluster so that the
distance between interfering cells (cochannel or adjacent channel) is as large as possible. For
urban land mobile radio, the distance power gradient increases from two (in the case of free
space) to roughly four so that the received signal strength falls as the fourth power of the
distance. This further improves the signal-to-interference ratio.

Page 29 of 36
If there are 𝐽𝑠 interfering BS surrounding a given BS, the general form of the signal-to-noise ratio
will be:

𝑑0α
𝑆𝑟 =
∑𝐽𝑛=1
𝑠
𝑑𝑛α

Where the distance of the mobile from the given base station is 𝑑0 and its distance from the nth
base station is 𝑑𝑛 .

Example 7: Sr in a Hexagonal Cellular Architecture

Recalling that there are exactly six cochannel cells with a hexagonal cellular structure, it is clear
that they will all cause similar levels of interference to a mobile terminal in the given cell. So 𝐽𝑠
= 6 here. Also, the distance at which the cochannel cells are located depends on the size of the
cluster from the equation below:

𝐷𝐿
= √3𝑁
𝑅𝐿

The farthest distance a mobile terminal can be from the base station of a given cell is the cell
radius R. The approximate distance of the mobile terminal from the base stations of each of the
cochannel cells is DL. For land mobile radio, if only the six cochannel cells that make up the first
tier of interferers are considered, 𝐽𝑠 = 6 and the signal-to-interference ratio can be approximated
as:

𝑅 −4 𝑅 −4 1 𝐷𝐿 4 3 2
𝑆𝑟 ≈ = = ( ) = 𝑁
𝐽𝑠 𝐷𝐿−4 6𝐷𝐿−4 6 𝑅 2

In terms of dB, we can write the signal-to-interference ratio as:

1 𝐷𝐿 4 𝐷𝐿 3
𝑆𝑟 = 10 𝑙𝑜𝑔 [ × ( ) ] = −7.78 + 40𝑙𝑜𝑔 ( ) = 10 𝑙𝑜𝑔 [ × 𝑁 2 ] = 1.76 + 20𝑙𝑜𝑔𝑁
6 𝑅 𝑅𝐿 2

Figure 7 shows how the signal-

to-interference ratio given by
the equation above varies with
the cluster size N. This
equation is commonly used to
determine the cluster size for
an adequate performance. Note
that the signal-to-interference
ratio is influenced by the
cochannel reuse ratio DL/RL in
that a given DL/RL has to be
Figure 7: Signal-to-interference ratio Sr as a function of frequency reuse factor N.
maintained for a particular 𝑆𝑟 .

Page 30 of 36
Example 8: Cellular Architecture of AMPS

As an example of cellular architecture, we consider the very first cellular radio telephone system
in the United States, called AMPS based on an analog FM modulation scheme. Each voice
channel in AMPS occupies 30 kHz of bandwidth and uses FM. Figure 8 shows the spectrum
allocations for AMPS.

Figure 8: Frequency spectrum allocation for AMPS

A bandwidth of 25 MHz is allocated for both the uplink and downlink so that transmission is full
duplex. The 25 MHz of spectrum is divided into two blocks of 12.5 MHz each. Block A is
allocated to carriers who are not traditional telephone service providers. Block B is allocated to
traditional telephone service providers. Each 12.5 MHz of spectrum can support 416 channels
each of which is 30 kHz wide. Of these, 395 are dedicated channels for voice, and 21 are
dedicated for call control. Based on subjective voice quality tests, it was determined that a
signal-to- interference ratio of 18 dB can be tolerated while providing a good voice quality to the
user. This means that the cluster size has to be N=7. Figure 6 shows the cellular architecture with
this cluster size. Cells with the same label use the same frequency spectrum. They are separated
by a distance DL=4.58 RL in this case which ensures that the signal-to-interference ratio is around
18 dB.

Let the 395 voice channels available for a service provider be numbered from 1 to 395. For
example, on the downlink, 869-869.030 corresponds to channel 1, 869.030-869.060 to channel 2,
and so on. Channels 1, 8, 15, ... are allocated to cells labeled A. Channels labeled 2, 9, 16, ... are
allocated to cells labeled B and so on. This ensures that there is a sufficient separation between
channels used within a cell so that adjacent channel interference is minimized. In practice, the
numbering scheme is different because the entire 25 MHz of bandwidth was not initially
available for AMPS.

Page 31 of 36
However, a separation of seven adjacent channels is maintained between channels used within a
cell. It was also found in some cases that, because cells actually do not have a hexagonal shape
and because the assumptions made in coming up with the value for N are not satisfied, a cluster
size of N=12 has to be employed for good voice quality.

Main investment in deploying a cellular network

The main investment in deploying a cellular network is the cost of the infrastructure, which
includes the cost of BS and switching equipment, property (land for setting up the cell sites),
installation, and links connecting the BSs. This cost is proportional to the number of BS sites.
The income of the service is directly proportional to the number of subscribers. As the number of
subscribers grows, a cellular service provider must develop a reasonable deployment plan with a
sound financial structure to account for these aspects. All service providers begin operations with
the minimum number of cell sites to cover a service area, requiring the least initial investment.
As the number of subscribers increases, it generates a source of income for the service provider.
At this point, they can increase investment in the infrastructure to improve service and increase
network capacity to support additional subscribers. Therefore, a number of methodologies have
evolved to facilitate the expansion of cellular telephone networks.

Capacity expansion techniques

There are basically four methods to expand the capacity of a cellular network.

The simplest method is to obtain additional spectrum for new subscribers, which is a very simple
but expensive approach. For example, the PCS bands were sold in the United States for around
$20 billion. If each new subscriber generates a profit of approximately $1,000 per year, 20
million additional subscribers would be needed to recover this amount in a year. Due to fierce
competition to provide the lowest cost to the customer, this approach has proven fatal. As an
example, the top three companies that purchased the PCS bands have already filed for
bankruptcy. However, this does not mean it is not an acceptable method. This pessimistic
scenario accentuates the vital importance of other alternatives to expand capacity in addition to
simply acquiring additional spectrum.

The second method to expand the capacity of a cellular network is to change the cellular
architecture. Architectural approaches include cell splitting, cell sectoring using directional
antennas, microcell zone technique and using multiple reuse factors (called reuse partitioning).
These techniques, described in detail later in this section, increase capacity by changing the size
and shape of cell coverage through adding cell sites or modifying antenna characteristics. These
techniques do not require additional spectrum or major changes in the wireless modem or access
technique of the system that would necessitate users purchasing new terminals. These features
distinguish architectural approaches as one of the more practical and less expensive solutions for
expanding network capacity.

Page 32 of 36
The third method for capacity expansion is to change the frequency allocation methodology.
Instead of distributing existing channels equally among all cells, it is possible to use a
nonuniform distribution of frequency bands among different cells according to their traffic needs.
The traffic load of each cell changes dynamically based on the geography of the service area and
with time. Most downtown areas experience the largest traffic loads during rush hours and
relatively light traffic loads in the evening and on weekends. This situation is reversed in
residential areas. Dynamically allocating channels to different cells can increase the overall
capacity of the network. These techniques do not require any changes in the terminal or physical
architecture of the system, and they are implemented within the computational devices used for
network control and management.

The fourth and most effective method to expand network capacity is to change the modem and
access technology. The cellular industry began with analog technology using FM modulation and
has evolved toward TDMA and then CDMA air interface using digital modems. Digital
technology increases network capacity and provides a fertile environment for integrating voice
and data services. However, this migration requires users to purchase new terminals and service
providers to install new components in the infrastructure.

Architectural methods for capacity expansion - Cell Splitting

As the number of subscribers increases within a given area, the number of channels allocated to a
cell may become insufficient to support the subscriber demand. It then becomes necessary to
allocate more channels to the area covered by this cell. This can be achieved by splitting cells
into smaller cells and allowing additional channels in these smaller cells. Consider Figure 9,
which shows a cellular architecture employing a cluster size of seven. When the traffic load
increases, a smaller cell is introduced, having half the area of the larger cells. This will ultimately
increase the capacity fourfold (because area is proportional to the square of the radius). In
practice, however, only a single small cell will be introduced midway between two cochannel
cells. In this case, these are the larger cells labeled A. It is logical to reuse the channels allocated
to these cells in the smaller cell to minimize interference.

Figure 9:Illustration of cell splitting for capacity expansion.

Page 33 of 36
This approach presents some problems. Let us assume, the radius of the smaller split cell
(labeled 'a') is R/2 and the transmit power of its base station is the same as that of the larger cells.
As far as the smaller cell is concerned, the signal-to-interference ratio is maintained because the
maximum distance a mobile can be from the BS in this cell is R/2. So, even though the distance
between this cell and the cochannel cells A is reduced by half, the value of Sr remains the same.
On the other hand, this is not the case for cells labeled A because their cochannel reuse ratio with
respect to the smaller split cell is now DL/2R.

To maintain the same level of interference, the transmit power of the base station in the smaller
cell should be reduced. However, this would increase the interference observed by mobiles in the
smaller cell. The alternative is to divide the channels allocated to cells labeled A into two parts:
those used by 'a' and those not used by 'a'. The channels used by 'a' will only be used in the larger
cells within a radius of R/2 from the center of the cell to maintain the cochannel reuse ratio for
these channels. This is known as the overlaid cell concept, where a larger macrocell coexists
with a smaller microcell.

The drawback of this approach is that the capacity of the larger cells is reduced, which will
eventually lead to introducing split cells in their area, resulting in a chain reaction where the
entire area is served by smaller radius cells. Additionally, the BS in cells labeled A will become
more complex, and handoffs between the overlays will be necessary.

Cell Sectoring using Directional Antennas

The simplest and most popular scheme for expanding the capacity of cellular systems is cell
sectoring using directional antennas. This technique aims to reduce the signal-to-interference
ratio, thereby reducing the cluster size and increasing capacity. The idea behind using directional
antennas is to reduce cochannel interference by focusing radio propagation only in the required
direction. To achieve this, the coverage of a base station antenna is restricted to a part of a cell,
called a sector, by making the antenna directional. In implementing this technique, cell site
locations remain unchanged; only the antennas used at the site are changed. The main objective
is to increase the signal-to-interference ratio to a level that allows for a lower frequency reuse
factor. A lower frequency reuse factor permits a larger number of channels per cell, increasing
the overall capacity of the cellular network. As discussed earlier, the signal-to-interference ratio
is given by:

1 𝐷𝐿 4 9 2
𝑆𝑟 = ( ) = 𝑁
𝐽𝑠 𝑅 𝐽𝑠

Here, 𝐽𝑠 is the number of interfering cell sites. Using a sector antenna reduces the factor 𝐽𝑠 , which
decreases interference and increases 𝑆𝑟 . The most popular directional antennas employed in
cellular systems are 120° directional antennas, and in some cases, 60° directional antennas are
also used. The following examples evaluate the impact of these antennas, which allows the reuse
factor to be reduced from N=7 to N=4 and N=3, respectively. In some instances, even with a
reduced reuse factor, sectorization yields a better 𝑆𝑟 than the required 18 dB value.

Page 34 of 36
Example 9: Three-Sector Cells and a Reuse Factor of N=7.

Consider a seven-cell cluster scheme with 120° directional antennas, as shown in Figure 10.
Channels allocated to a cell are further divided into three parts, with each part used in one sector
of a cell. As the figure illustrates, the number of cochannel interfering cells is reduced from six to
two. This results in an improvement in the signal-to-interference ratio. For omnidirectional
antennas, the 𝑆𝑟 value for a cluster size of N=7 is 18.66 dB. In this case, the signal-to-
interference ratio is given by:

1 𝐷𝐿 4 1 𝐷𝐿 4 9 2
𝑆𝑟 = ( ) = ( ) = 𝑁
𝐽𝑠 𝑅 2 𝑅 2

Figure 10: Seven cell reuse with 120° directional antennas (3-sector cells)

For N=7, this will result in 𝑆𝑟 =23.43 dB. To understand the importance of this gain, note that the
required signal-to-noise ratio for AMPS systems is 18 dB, which suggests N=7. However, a
larger 𝑆𝑟 is needed due to nonideal situations.

Page 35 of 36
Example 10: Three-Sector Cells and a Reuse Factor of N=4.

With omnidirectional antennas, 𝐽𝑠 =6, and for N=4, the result is 𝑆𝑟 =13.8 dB, which is inadequate
for AMPS. It can be observed that the signal-to-interference ratio with three-sector cells is
substantially better compared to omnidirectional antennas and no cell sectoring. With N=4, the
signal-to-interference ratio is 19.9 dB. This value exceeds the 18 dB requirement based on
subjective mean-opinion-score tests of voice quality.

Figure 11: Four cell reuse with 120° directional antennas (3-sector cells)

Example 11: Six-Sector Cells and a Reuse Factor of N=4 and N=3.

With 60° directional antennas, there are six sectors within a cell. The number of interfering
cochannel cells reduces to one and the signal-to-interference ratio can be written as:

1 𝐷𝐿 4 1 𝐷𝐿 4
𝑆𝑟 = ( ) = ( ) = 9𝑁 2
𝐽𝑠 𝑅 1 𝑅

So, when N=4 the signal-to-interference ratio:

𝑆𝑟 = 10 𝑙𝑜𝑔[9 × 𝑁 2 ] = 10 𝑙𝑜𝑔[9 × 42 ] = 21.58 𝑑𝐵

So, when N=3 the signal-to-interference ratio:

𝑆𝑟 = 10 𝑙𝑜𝑔[9 × 𝑁 2 ] = 10 𝑙𝑜𝑔[9 × 32 ] = 19.08 𝑑𝐵

Page 36 of 36