Академический Документы
Профессиональный Документы
Культура Документы
Communications Systems
For a listing of recent titles in the
Artech House Space Technology and Applications Series,
turn to the back of this book.
Radio Frequency Interference in
Communications Systems
Bruce R. Elbert
Library of Congress Cataloging-in-Publication Data
A catalog record for this book is available from the U.S. Library of Congress.
ISBN-13: 978-1-60807-965-0
All rights reserved. Printed and bound in the United States of America. No part
of this book may be reproduced or utilized in any form or by any means, elec-
tronic or mechanical, including photocopying, recording, or by any information
storage and retrieval system, without permission in writing from the publisher.
All terms mentioned in this book that are known to be trademarks or service
marks have been appropriately capitalized. Artech House cannot attest to the
accuracy of this information. Use of a term in this book should not be regarded
as affecting the validity of any trademark or service mark.
10 9 8 7 6 5 4 3 2 1
Contents
Preface ix
CHAPTER 1
The What and Why of RFI in Radio and Wireless Communications 3
1.1 Engineering of Radiocommunication and Wireless Systems in
Light of RFI 5
1.1.1 Efficient Use of Radio Spectrum 8
1.1.2 Frequency Band Assignments 9
1.1.3 Efficient Use of the Geostationary Satellite Orbit and Other Space
Environments 10
1.2 Identifying RFI Modes and Consequences 12
1.3 Electromagnetic Compatibility and Spectrum Sharing 12
1.4 Addressing RFI When It Occurs 13
References 13
CHAPTER 2
The Radiocommunication System in an Interference-Limited Environment 15
2.1 Radiocommunication Requirements 15
2.1.1 Who: The Organization or Market 16
2.1.2 What: Network Topology and Frequencies 17
2.1.3 Where: Locations to Be Served 17
2.1.4 When: Timing for Fixed, Mobile, or Temporary Communications 18
2.1.5 How: Equipment and Other Resources 18
2.2 Transmitters and Receivers 19
2.2.1 Modem Properties 22
2.2.2 Frequency Conversion and Amplification 24
2.2.3 Land-Based Radiocommunication Concepts 26
2.2.4 Space-Based Radiocommunication Concepts 26
2.2.5 Communications in Space Research and Remote Sensing 28
2.3 Antennas in Radiocommunication Services 29
2.3.1 Nondirectional Antenna Radiation Patterns 29
2.3.2 Waveguide Horns 31
v
vi Contents
CHAPTER 3
Key Concepts for Evaluating RFI 47
3.1 Manmade RFI (Unintentional and Intentional) 47
3.2 RF Intermodulation and Harmonics 50
3.3 Interference from Radars and Other Pulsed Sources 54
3.4 Protection Ratio and Receiver Sensitivity 56
3.4.1 How to Determine Receiver Threshold 57
3.4.2 Receiver Desensitization 59
References 61
CHAPTER 4
Link Properties Under Free Space Conditions 65
4.1 Path Geometries 66
4.1.1 Ground to Ground 68
4.1.2 Air to Ground 73
4.1.3 Space to Earth 74
4.1.4 Space to Space 78
4.2 Line-of-Sight Propagation Characteristics in Atmosphere 80
4.2.1 Coupling to the Antenna 80
4.2.2 Alternative Measures of Signal Strength 83
4.2.3 Atmospheric Loss 84
4.2.4 Rain Attenuation 84
4.2.5 Tropospheric Scatter 88
References 89
CHAPTER 5
Propagation on Obstructed Paths 91
5.1 Path Profiles and Obstructions 91
5.2 Fading on Direct and Blocked Paths 96
5.2.1 Direct and Reflected Signals: Vector Addition 96
5.2.2 Ricean Multipath Fading 97
5.2.3 Rayleigh fading 98
5.3 Geographic Coverage Analysis 100
5.3.1 Path Computation Using Software 101
5.3.2 Measures of Radio Coverage 103
Contents vii
CHAPTER 6
Interference Protection Ratio (C/I) and Its Application 115
6.1 C/I Criteria 117
6.1.1 Calculation of Static Values 118
6.1.2 Antenna Characteristics and Isolation 122
6.2 Non-Steady State Propagation: Scatter, Rain-Induced Interference,
Ionospheric Conditions 123
6.3 Service Performance in the Presence of Interference 125
6.4 Interference Among Various Systems 126
6.4.1 Radars vs. Land Mobile Systems 126
6.4.2 Protection of GPS from Terrestrial Wireless Systems 133
6.5 Satellite Communications vs. Terrestrial Microwave Systems 135
References 138
CHAPTER 7
RFI Resolution Techniques 141
7.1 Spectrum Analysis and Monitoring 141
7.1.1 Spectrogram View and Adjustment 144
7.2 Spectrum Monitoring and Carrier Measurement 147
7.3 Interference Location and Radio Direction Finding 150
7.4 Transmitter ID 154
7.5 Additional Resources and Strategies 156
7.5.1 Planning and Coordination 156
7.5.2 Research and Testing 157
7.5.3 Remediation 161
References 163
CHAPTER 8
How to Identify, Prevent, and Fix Common RFI problems 165
8.1 Frequency Planning Techniques 165
8.1.1 Frequency Assignment 165
8.1.2 Transmitter Control and ID 167
8.1.3 Interleaving Spectra and Band Segmentation 168
8.1.4 Power Control 171
8.2 Avoiding Intermodulation 171
8.3 Interference Cancellation 173
8.3.1 Cancellation at RF (Antenna) 174
viii Contents
CHAPTER 9
Prospective for RFI Resolution in Future Radiocommunication Systems 185
9.1 New Approaches to Interference Management 185
9.2 Cognitive Radio 186
9.3 Spectrum Management Tools and Resources 194
9.3.1 Licensed Shared Access 194
9.3.2 Model Cities Demonstrations 195
9.3.3 Spectrum Monitoring and RFI Remediation 198
References 200
List of Acronyms 201
About the Author 205
Index 207
Preface
Radio frequency interference (RFI) is the kind of technical topic that creates anxi-
ety in those who use radio waves for telecommunications. We refer to this mode as
radiocommunication, a contraction adopted by the International Telecommunica-
tion Union (ITU) and which identifies our domain of interest. But, RFI is a unique
form of disruption to radiocommunication in that it enters through the receiving
antenna as do desired signals. Radio waves know no boundaries because of their
unguided nature and ability to cross geographic and political boundaries. For this
reason, the ITU long ago established radiocommunication as one of its primary
areas of activity, conducted through ITU-R, where the R stands for radio.
This book is about the practice of identifying, resolving, and managing RFI
across the broad range of radiocommunication systems and applications. This is a
technical field which, like medicine, demonstrates how a science of maladies has a
direct impact on daily life. Radio developed over the years from the two most basic
modes of one-way broadcasting and two-way point-to-point communication, to
the point now that literally all forms of communication, including the Internet, rely
on it to reach the end user. In the latter case, the popular terms are “cellular” and
“wireless,” which seem to distinguish these uses from the classic term “radio.” But
those of us who work in and across these domains know full well that RFI is still a
common denominator of concern and even trouble. RFI has become a high-stakes
game where entire business plans and mission strategies can succeed or fail if RFI
is not resolved.
We have evolved from a time when static and background noise were the main
cause of reception problems and connection failures; now, RFI is often the de-
terminant of performance and, importantly, system capacity. Achieving the right
capacity becomes a financial issue in terms of revenue to a commercial wireless en-
terprise. We all recognize that the radio spectrum is a limited resource and must be
reused many times over across geography and direction. This book is dedicated to
proper consideration of RFI, wherein we can maintain the reliability and promote
greater use of radiocommunication going forward.
To understand RFI one needs to understand the principle of radio: modern
transmitters and receivers employing digital modulation and coding, directional
antennas to achieve coverage and reduce RFI in the environment, the ways that
radio waves propagate in space and across the landscape, and the mechanisms that
can produce RFI in the first place. From that, we address how it can be character-
ized and resolved into a workable problem with practical solutions. Many readers
ix
x Preface
�������
will come into the topic of RFI with a solid background in radio engineering or
practice. Others with less experience will still understand the importance of RFI
and need for its resolution, but want some of the background that supports the
necessary study and design (the “secret handshake” of radiocommunication art).
For that reason, we provide sufficient description of radiocommunication so that
all readers will be able to follow our exposition and apply its recommendations.
To that end, this book is organized into three broad topics: I) the requirements
for and engineering of radiocommunication systems and technology, II) radio wave
propagation, defining the direct and indirect paths on the Earth and between Earth
and space taken by the desired signal and the undesired RFI, and III) the details
of RFI analysis (particularly C/I assessment) and resolution strategies involving
investigation and mitigation of problems before, during and after RFI is identified
or experienced.
Part I includes Chapters 1, 2, and 3 and is a review of radiocommunication in
the context of the RFI question. There is enough technical material here to be a
basis for subsequent examination of RFI. Expert readers can scan this material and
then move on to the core topics in the remainder of the book. Chapter 1 points out
where the consideration of RFI fits within the overall field of electromagnetic com-
patibility (EMC) and why RFI problems require effective study and brainstorming.
The coverage of radiocommunication system design in Chapter 2 is at a fairly basic
level but is technically accurate. This is a good jumping off place for further study
using the World Wide Web or other resources referenced with the chapter. Chapter
3 delves specifically into the sources of electromagnetic disruption to radiocom-
munication and provides a good review of such key topics as intentional and unin-
tentional interference; intermodulation and spurious signal generation; interference
from pulse sources (e.g., radar); and the key topic of C/I, also called protection
ratio (addressed in greater detail in Chapter 6). These are considered throughout
part I) for land-based (terrestrial) systems as well as space-based repeaters and
signal sources.
Part II provides Chapters 4 and 5 to identify the various propagation modes
that radiocommunication signals and RFI take between transmitting emitter (an-
tenna or otherwise) and receiving antenna. This is a complex topic and depen-
dent on propagation data collected and analyzed over past decades and through
experience with radio systems. These extend from 100 kHz on the low end up to
100 GHz at the top of current millimeter wave development. Chapter 5 will be
familiar to many who have involvement with radio since it covers propagation in
free space. While this type of path may be unobstructed, there are key phenom-
ena in the earth’s atmosphere that both reduce (attenuate) signals and can allow
RFI to propagate well beyond the horizon (rain scatter and ducting, for example).
Obstructed paths are covered in Chapter 6 to address signals that may be blocked
but nevertheless reach the receiver through a variety of means. The methods of
knife-edge diffraction (Gourdet/Fresnel) and smooth earth diffraction (Bullington)
are reviewed with examples provided. When mobility is introduced, the individual
desired and RFI signals experience respective fading of their own that varies with
time and position. The discussion includes methodologies for calculating the levels
of signals for use in general assignment of spectrum to operators (which requires
consideration of how transmission by one organization can produce RFI into the
Preface xi
operation of another) and end users. Area coverage models such as Okamura and
Langley-Rice are discussed with regard to their applicability.
This brings us to Part III with Chapters 6, 7, and 8, addressing the subject
of how RFI can be evaluated properly and dealt with effectively. First, we see in
Chapter 6 how various forms of RFI affect the desired signals. It’s not enough to
know how much RFI is present; we must know and be able to quantify the impact
on a victim service from the presence of a given type and level of RFI. Much of the
effort in RFI resolution ends up going into this question, especially if parties to the
discussion must agree on what is an acceptable impact on service quality for a given
RFI condition. Fortunately, it can be addressed analytically using equations and
computer number crunching, and through actual laboratory tests using the types
of equipment to be fielded.
Chapter 6 gives some practical examples to provide a jumping off place for
detailed study in a specific RFI case. These include quantitative studies done of 4G
LTE sharing with radar and applicability of path analysis software available on
the commercial market. Some proven ways of addressing RFI once it is identified
and understood are covered in Chapters 7 and 8. There is never a final answer to
all cases of RFI but Chapter 7 provides a number of proven tools and methodolo-
gies, especially spectrum analysis and spectrum monitoring. Chapter 8 follows in
a similar vein by laying out practical spectrum planning techniques that can be
introduced in many cases. This listing grows with time, as we gain experience with
new systems and as technology advances in such areas as radio direction finding
(RDF) and RFI cancellation.
In looking ahead, we offer Chapter 9 as a prospective for RFI resolution in
future radiocommunication systems. Most of what we need is already known and
being applied on a piece-meal basis. But, new models of how spectrum could be al-
located and shared are of great interest in a world where more services are desired
by the global population. We address the topic of cognitive radio, where intelli-
gence within the system uses “white spaces” of spectrum that is currently lost due
to the rigid rules of allocation and limitations on radio network management. So
far, this topic has been of interest mainly to researchers in universities and regula-
tors wishing to grant more access. In time, many of these ideas will be tested in
limited trials (in the US and elsewhere), and some will be implemented to mitigate
the RFI that comes with increasing access to spectrum.
This book comes out of the author’s 50 years of experience with radiocommu-
nication in general and work involving RFI in particular. It’s not that one starts out
to be an RFI practitioner; it happens as you get involved with real radiocommu-
nication systems and have to solve these problems as they are recognized to exist
or, worse, when they raise their ugly heads. There is profound hope in our ability
to understand RFI and find those solutions that work and are economical. But, as
is always the case with systems engineering practices, you can pay for the solution
now, or pay a lot more later.
The author wishes to thank the organizations and individuals who contributed
the materials identified in the reference lists with each chapter. Specifically, I ac-
knowledge the guidance and learning in the past from very experienced engineers
I’ve had the pleasure knowing: Col. Jacques Deygout of the French Army; Jorge
Fuenzalida and Hans Weis, formerly of COMSAT; Dr. Klaus Johnsen and Chuck
Sanderson, fellow Hughes engineers; and Professor Egon Brenner, my first EE
xii Preface
�������
professor at CCNY. Also vital to the success of this project was the encouragement
from the staff at Artech, including the publisher, and the acquisition team who as-
sisted at each step. The technical reviewer, Ray Sperber, really took on this project
as a partner so that it represents the real world in all its dimensions and maintains
its validity and relevancy in the years ahead. The original idea for a book on RFI
arose in Indonesia in the late 1970s during the Palapa satellite project, with the
help and encouragement of R. Wikanto and Dr. Arifin Nugroho of the Indonesian
Public Telecom Corporation (PERUMTEL).
I want to express my heartfelt appreciation and love for my wife, Cathy, and
our two wonderful daughters, Sheri and Michelle. They are the ones who make a
book like this become a reality.
Part I
Radiocommunication Systems and the
RFI Environment
1
CHAPTER 1
Radio frequencies (RFs) lie in the broad spectral range of 100 kHz to 100 GHz and
employ the principles of radio wave propagation and electrical communications.
We concern ourselves here with how what emanates from a transmitting antenna
of one radiocommunication system causes radio frequency interference (RFI) that
enters the receiving antenna of another. By “what,” we mean the fact that one
radio signal can deteriorate proper reception of other radio signals by preventing
the associated receiver from properly detecting the information, or data, on that
particular signal. By “why,” we mean that the RF spectrum must be shared by many
users and many applications in point-to-point communications and broadcasting;
as a result, no user can expect to enjoy any particular part of the spectrum on a
dedicated basis. These principles are explored in detail in this book, but the idea is
relatively simple.
RFI is something that occurs by the very nature of this means of communica-
tion; as spectrum use increases, so does RFI. It is analogous to vehicle traffic on
highways and streets. With little or no traffic (e.g., at 2 a.m.), traffic flows freely
at the posted speed; with high traffic (during rush hour in major cities), traffic is
intense and we all have to watch out carefully as we drive. That doesn’t mean that
traffic cannot be handled in heavy situations; just that there must be techniques and
tools to help address those places and periods of congestion. Similar analogies can
be drawn with respect to flying airplanes on air routes.
RFI is one category of a much broader field of electromagnetic practice
involving:
3
4 ������������������������������������������������������������
The What and Why of RFI in Radio and Wireless Communications
The mechanisms for EMI are illustrated in Figure 1.1 [1]. We see many paths for
EMI to take, beginning at the signal line with conducted coupling at the bottom and
multiple other potential paths crisscrossing among the principle elements of both
the interfering and interfered-with (victim) systems. At the top is the RFI path (of
primary interest in this book), indicating radiated coupling between the antennas
(the dashed arrow). Keep in mind that we are speaking of two normally uncon-
nected systems and that the radiated coupling is undesired. It should be possible
to eliminate the non-RFI paths through proper design, construction, and testing
of each of the operating systems. Often, one of these paths ends up the cause of a
specific interference issue. When RFI arises in an operating system, we often must
consider all paths for the interference to take, which makes this illustration particu-
larly valuable.
The definitions of EMI and EMC are meant to distinguish these related (and
important) practices from our focus on RFI. The radiocommunication systems op-
erating on licensed or otherwise authorized frequencies can meet all of the relevant
EMC requirements and still cause RFI to one another. This is because one person’s
RFI is another person’s communication, and vice versa. It is up to the designers
and operators of these radiocommunication systems to minimize how RFI impacts
their respective operations. To address and control RFI (i.e., the “how”), we use
the same principles of radio engineering as are applied to the design of the systems
Figure 1.1 Coupling of EMI between two different radio systems [1]. RFI is illustrated as “radiated
coupling” between the antennas; all other paths are unintentional and avoidable (although still pos-
sible in a working system).
1.1 Engineering of Radiocommunication and Wireless Systems in Light of RFI 5
RFI appears in the receiver of a radiocommunication system like the thermal noise
from the atmosphere and produced within active devices like amplifiers and mixers,
and any type of passive matter that produces signal attenuation. From a technical
standpoint, RFI may be equivalent to an amount of thermal noise, or if consider-
ably different, as a unique interfering signal. Many radiocommunication systems,
such as cellular mobile radio and satellite transmissions, have internal forms of RFI
that are evaluated as part of the operation of that particular system. Engineering in
this context consists of techniques for analysis and means to take corrective action,
and the methodologies for accomplishing this within the constraints of radiocom-
munication system operation.
Figure 1.2 illustrates the generic nature of a radiocommunication system, pro-
viding a point-to-point link. The transmitter contains a microphone that represents
the information source, a modulator to transfer the information signal into a radio
carrier suitable for transmission, and a transmitting antenna. From there, the car-
rier is propagated over the medium (to be discussed in detail) and received by the
receiving antenna and the distant receiver. Electromagnetic energy is converted by
the antenna into an electrical signal that is transferred to the demodulator. There is
a bandpass filter between the receiving antenna transmission line and the demodu-
lator for the purpose of restricting the bandwidth to only that of the desired carrier.
As such, it is the principle element that protects the receiver from unwanted signals
that couple through the receiving antenna. The recovered information is amplified
before it is applied to a loudspeaker. Modern radiocommunication systems transfer
information in digital form and employ modulation with digital coding to maxi-
mize the capacity and reduce the effects of distortion and interference that result
from transmission and propagation.
Any system that radiates from antennas in the typical frequency bands is po-
tentially a source (or receiver) of RFI from other radiocommunication systems.
Examples of the former include radars, RF identification (RFID) systems, and navi-
gations systems using Global Positioning Satellite (GPS) system, and other radio
techniques. Within the radio spectrum, we see allocations for all of the applications
that employ RF and wireless schemes. Any and all can become part of a particular
RFI issue.
All this may appear to be uncontrollable and beyond precise investigation;
however, the basic physics behind all RFI issues and concerns are well known and
quantifiable. This book provides a review of the relevant principles involving an-
tennas and RF devices, mathematics of electromagnetism, radio wave propaga-
tion in free space and on obstructed paths, communications technologies involving
signal processing and detection, and engineering principles such as budgeting and
design specifications, for considering RFI in modern radiocommunication systems
on land, sea, in the air, and in space.
RFI may not obey national borders, but it can be constrained by geography.
Propagation in free space is governed by the inverse square law, wherein the signal
strength decreases by a factor of 4 (6 dB) for each doubling of distance. According
to the range equation, also called the Friis transmission equation, the attenuation
due to a free space path of distance R and at a wavelength λ can be expressed as
2
4 πR
A0 =
λ
Figure 1.3 Free space attenuation as a function of range for two frequencies: 1 GHz and 10 GHz.
Figure 1.4 Path attenuation for propagation between two locations taking account of free space and
flat Earth. This assumes that the transmitter and receiver are at heights of 20m and 2m, respectively.
or 12 dB. That is 6 dB more than free space propagation would indicate. This extra
isolation helps with the reassignment of the same spectrum to a separate geographi-
cal region. This figure indicates that at a distance of 10 km, the attenuation of the
path over the Earth experiences 20 dB more attenuation than that in free space.
Having an obstacle, such as a hill or building, adds significant attenuation
through the principle of diffraction and blockage. The amount of attenuation de-
pends on the frequency, distance, and degree to which the obstacle penetrates the
line-of-sight path. This will be addressed in detail in Chapter 5. As an example,
Figure 1.5 is the profile of a path between an interfering transmitter and a receiving
8 ������������������������������������������������������������
The What and Why of RFI in Radio and Wireless Communications
Figure 1.5 Example of an obstructed path between a potential source of RFI (on the left) and a
receiver on a hill top (on the right), courtesy of SoftWright.
point approximately 20 km away. There are several tall hills in the way, reducing
the received RFI by in excess of 80 dB. Thus, near-complete isolation is afforded. In
other cases, the diffraction loss from an obstacle will range between 6 and 20 dB,
necessitating precise calculation using the principles reviewed in Chapter 5.
hundreds of spectrum allocations across the RF range, broken down among three
ITU regions: (1) Europe and Africa, (2) North and South America, and (3) Asia and
the Pacific. There are three major communication application areas: Fixed Service,
providing communication between points on the ground, Mobile Service, among
mobile platforms on ground, on water and in the air, and Broadcasting Service, the
one-to-many application for distributing information and programming. There are
comparable satellite-specific services in the three domains: Fixed Satellite Service
(FSS), Mobile Satellite Service (MSS), and Broadcasting Satellite Service (BSS).
1.1.3 Efficient Use of the Geostationary Satellite Orbit and Other Space
Environments
There is value to having spectrum applied to the purposes of highest importance or
economic value. Sometimes, regulations end up holding spectrum resources back
from their best use in terms of allowing more access and application. We see this
play out in licensed wireless spectrum where terrestrial operators pay billions of dol-
lars to acquire the spectrum they need to meet demand and remain competitive. In
the satellite domain, most usable Geostationary Earth Orbit (GEO) orbit positions
are occupied by operators established decades ago. This resulted from a general
presumption that an operator may replace a satellite in an existing orbit location
with another of equal or better performance when the prior satellite either reaches
end of life or experiences a failure. To economize on spacecraft vehicles, some op-
erators have joined together on shared communications satellites. The overall RFI
environment, discussed later in this book, is well understood and reasonably well
contained within the GEO domain.
An aspirant operator with capital and operational resources will find it dif-
ficult, but not impossible, to obtain an appropriate GEO orbit longitude location
through the process defined in the ITU Radio Regulations in conjunction with the
system of licensing from a particular host country. The licensing bodies have vary-
ing degrees of experience with the process; the regulators in the United States, the
Russian Federation, Canada, the United Kingdom, France, Germany, Luxembourg,
Hong Kong, Japan, Brazil, Malaysia, and Indonesia having been at the game the
longest. Completing the ITU and national regulatory process will take from two
the five years, and there is risk that the assignment may end up being contested.
The process is illustrated in Figure 1.6.
There are cases where holders of licenses and prior registrations of the ITU
have sold or otherwise transferred their right to a location to a new operator or an
Figure 1.6 Basic structure applied to the process of obtaining a new orbit location through the
auspices of the ITU and a domestic regulatory body (administration).
1.1 Engineering of Radiocommunication and Wireless Systems in Light of RFI 11
existing operator needing expansion capacity. For example, the Kingdom of Tonga
registered many orbit locations in order to rent them to prospective satellite opera-
tors. Intentional RFI has occurred on a few occasions as disputes boil over. The
Isle of Man has become a popular jurisdiction for satellite companies wishing to
obtain orbit locations, although the actual jurisdiction is the UK. There have also
been outright purchases of operating satellites by operators in different countries
and the associated licenses and ITU registrations have crossed borders. While the
process is legal, it nevertheless can result in disagreements involving countries and
their associated satellite operators.
The basic mechanisms of orbital RFI can be seen in Figure 1.7, which illustrates
the geometry between the satellites in GEO and the transmitting and receiving
Earth stations. The case of uplink interference, shown in the left, results not from
the adjacent satellite but from the Earth station transmitting to that satellite. We
see how energy radiated in off angles from the Earth station antenna can reach the
satellite on the left, potentially interfering with the carrier coming up from the asso-
ciated earth station on the left. Downlink interference, illustrated on the right side
of the figure, is what we would normally associate with orbital RFI. The adjacent
satellite (on the right) radiates toward both receiving Earth stations, including the
one on the left that is not intended as a point of reception. Interference enters that
Earth station antenna through its off angle and thus can interfere with the desired
signal coming from the left-hand satellite.
Non-GEO orbit constellations like Iridium and O3b involved a somewhat dif-
ferent process because only one such constellation can operate in a particular fre-
quency band at a time. Iridium engaged the FCC and the entire membership of the
ITU to obtain its ability to launch something that had not been done before. O3b
rode on the coattails of a previous effort called Teledesic that did not itself proceed.
It is therefore a different process to gain access to non-GEO orbital spectrum re-
sources and more time could be required than for the typical GEO satellite. One of
the reasons for this is that RFI would occur as satellites move through the sky and
at an altitude considerably below GEO. Thus, terrestrial radiocommunication and
GEO satellite systems may also be subject to a new source of RFI. The approach
usually taken is to segment the particular frequency band so that the non-GEO
Figure 1.7 Basic geometry for RFI in GEO satellite operations; uplink interference is shown on the
left and downlink interference is shown on the right.
12 The
������������������������������������������������������������
What and Why of RFI in Radio and Wireless Communications
system has its own, noninterfering bandwidth to occupy. See Section 8.1.3 for a
discussion of band segmentation.
We could call this RFI management because it encompasses the vital functions of
reducing the problem to something that is manageable within the capabilities of
the organization or industry. It can be broken up into the familiar discussion of
Who, What, Why, When, Where, and How. We have to perform these studies and
investigations well before the radiocommunication system is designed and brought
into use, as well as after it is in operation when the RFI is actually experienced (at
any time). Doing this within the system itself may be the easiest part of the problem
because the characteristics of the radio stations and their locations are presumably
known. The analysis can be performed according to the procedures in this book.
What can be much more difficult is taking into account radiocommunication sys-
tems and radio emitters that are external. Some of these will exist prior to operation
of the potentially interfered with system (the RFI going both ways), and some will
come later when the undesired source (or receptor) of RFI appears on the scene.
The impact could be anticipated if the new radio emitters must notify existing us-
ers through an established regulatory process, and the impacts determined before it
goes live. Otherwise, it would be particularly troublesome because its characteris-
tics would not normally be known ahead of time.
What we are trying to achieve when a new radio emitting system initiates operation
is compatibility with all incumbent radiocommunication operations. Recall that
all spectrum is shared among multiple user types and applications, using principles
of RFI management. A review of the Table of Frequency Allocations of the ITU
Radio Regulations makes it clear that spectrum sharing is in the DNA of all radio
frequency assignment procedures. The Radio Regulations attempt to address this
through technical rules governing how the frequencies are to be used with each
service. If these rules are followed (and many are quite complex), spectrum sharing
should work reasonably well. Actual enforcement of the rules is up to the member
nations themselves and their regulatory agencies. Most of these entities lack criti-
cal resources for thorough review of license applications and monitoring the use
of spectrum to identify violators. Companies like Aerosystems, Inc., of Canada,
sell spectrum-monitoring systems that automate the scanning and verification of
spectrum use in critical bands used by wireless and satellite operators. There are
software tools that employ propagation models and mapping systems to predict
RFI conditions and how one operator would have an impact on another. In addi-
tion, the operators themselves have a big stake in spectrum sharing and manage-
ment, both within their own bounds and on the borders where RFI can exceed an
acceptable level.
1.4 Addressing RFI When It Occurs 13
RFI incidents represent some of the more serious problems that beset radiocommu-
nication systems when we are dealing with an unknown assailant. The cause can
be due to an equipment malfunction in our own system or in one within our field
of view. It could come from the introduction of a new radio emitter or system of
emitters of which we were unaware. Last, RFI can come from an intentional jam-
mer who wishes to disrupt the operation of a properly licensed or authorized user.
There are some tried-and-true techniques used in the troubleshooting of RFI
incidents. These are can summarized as follows. More detail on measurement and
mitigation techniques is presented in Chapter 7.
•• Is it from your own radio facility or system? Carefully review current opera-
tion and eliminate the possibility that the RFI is internally caused. All poten-
tial sources must be reviewed and eliminated from consideration.
•• Are there other “usual suspects”? These are RFI sources that have appeared
in the past. Build a database of RFI incidents and corrective actions taken.
•• How can I gather facts to allow deductive reasoning? Use instrumentation
such as a portable spectrum analyzer (discussed in Chapter 7), along with
detailed reports from credible sources; from this, brainstorm what the pos-
sible RFI sources could be and how to gather additional data for verification
and corrective action.
•• How can I get the attention and cooperation of possible sources? RFI identi-
fication and resolution, especially in the satellite communications and micro-
wave fields, involves getting the word out to all potential sources. Often, the
RFI results from an issue with otherwise properly working equipment. The
perpetrator has a stake in finding the source because it may accompany a loss
of service on their end. Involve those entities that can assist with the process,
including the high-level service operator (e.g., the satellite operator or domes-
tic telecom operator) as well as other users of the same or adjoining spectrum.
These are practical guidelines to get started in any RFI investigation, but the
devil is in the details. Take advantage of all of the brains and thinking that come
from good research as well as failure modes and effects analysis. Often, the process
itself has its way of resolving the RFI issue even if the actual cause is never identi-
fied. More information and guidelines are provided in Chapter 8. A very serious
RFI situation that prevents service may require its treatment as an incident in the
context of emergency management is discussed in Section 8.4.
References
[1] Malaric, K., EMI Protection in Communication Systems, Norwood, MA: Artech House,
2010.
[2] ITU Radio Regulations, Vol 1, Chapter 2, Geneva, Switzerland, 2012.
[3] Elbert, Bruce R., The Satellite Communication Applications Handbook, Second Edition,
Norwood, MA: Artech House, 2004.
CHAPTER 2
RFI management is but one aspect of the design and operation of a radiocommuni-
cation system serving some population or set of users. We recognize that the limited
spectrum considered in Chapter 1 is what forces engineers to take RFI seriously in
the development of a system in the first place. The overall RF design of the system
will be driven by the communications requirements, discussed next. How RFI af-
fects this design can only be assessed as the architecture comes together and choices
are made for the radio terminals and repeaters, and as locations are identified and
exploited. This is an ongoing process, and so RFI planning needs to be integrated
throughout the system lifetime (planning, building, and running). It will impact
the particular part of the spectrum employed, as different allocations bring with
them different bandwidth constraints, propagation characteristics, degree of prior
occupancy by incumbents, potential for new entrants, and physical characteristics
such as antenna designs, heights, application of shielding, choice of modulation
and coding methods, and so on. This also goes for properties of orbiting satellites
if these are included in the mix.
15
16 The Radiocommunication System in an Interference-Limited Environment
��������������������������������������������������������������������
One question that comes to mind is whether these services are provided over a
common network (which is effectively what we have with the Internet) and if they
are delivered to a standard and common terminal design (like the smartphone or
notepad computer). Larger devices can have larger antennas and transmit more
power; smaller devices can be taken almost anywhere and used in very difficult
environments.
Each of these has associated with it a basic requirement for data throughput
in bits per second (bps), time duration and frequency of use (use statistics), service
reliability (whether data delivery is guaranteed or “best efforts”), latency or delay,
and so forth. Voice services are tolerant of variable quality and dropped bits, but
latency needs to be kept to less than 1 second, round trip. Video is a bandwidth
hog but may not be sensitive to delay. Machine-to-machine communications covers
the gambit of commercial, industrial, and military uses, where lost bits or excessive
delay could result in disaster.
The “who” step in the systems engineering process is often the most difficult
because it forces participants and stakeholders to make choices before anything is
cast in concrete. Many have a very hard time being specific about their needs. In a
commercial market, potential customers or subscribers can’t imagine what the new
2.1 Radiocommunication Requirements 17
service will do for them. Demonstrations and trials go a long way, but until people
live and breathe with the terminal device and service they are at a loss to know how
it will impact their work and their lives. The same goes for how RFI would impair
or otherwise impact acceptance and ultimate success. An RFI incident can produce
dropped calls, reduced throughput, capacity bottlenecks, and failure to connect at
all. The system designer must think in many domains: like a user considering where
and how to use the service, like a terminal device with its software, like a spectrum
manager who considers where the bandwidth will come from and how RFI finds
its way through the system, like a businessperson considering the ultimate capital
and operating costs, and like a marketing person finding that marketing mix that
will achieve the revenue requirements.
terminal is receiving the weakest level of desired signal and at the same time subject
to the largest RFI level. Then, the logic goes, service performance everywhere else
will be better. This is a highly conservative approach that can burden the overall
system with high cost or even render it impractical. The second approach involves
a lot more work as it is based in evaluating RFI impact at every potential point in
the overall radiocommunication network. Then, we can calculate levels of service
impact, possibly on a percentage of quality basis at various points of coverage. Ra-
dio link performance can be enhanced by dynamic control when and where needed.
Some of the computer and software tools to do this are on the commercial market;
newer systems would require the development of custom software tools for the task
at hand. Generally speaking, network operators are pursuing the latter approach as
they tend to have substantial resources and a lot at stake. This is discussed further
in Chapter 9.
Figure 2.1 General engineering design process for a new radiocommunication system.
steady nature. RFI that is pulsing or sweeps across the band will interfere when it is
within the passband of the victim receiver and may have no impact at other times.
Also to be considered are out-of-band signals that can overload preamplifiers and
other active devices in the receive chain. This is of particular concern with regard to
legacy systems that might have been designed to more relaxed specifications than
for modern equipment in RFI-dominated environments.
The link budget is the primary tool for quantifying the key parameters that
determine whether the service is possible or not. Figure 2.2 presents a basic illus-
tration to graphically display the progression through a link budget for a point-to-
point link. The carrier power level is indicated on the scale on the left in dBm. The
particular path is free space, which is the ideal case discussed in Chapter 1. There
will be additional losses on the path from the atmosphere, obstacles, multipath,
and absorption from trees. These losses are resolved through higher transmit power
and antenna gain on transmit and receive ends of the path. The receiver will experi-
ence both thermal noise and RFI that appears at the distant end. The widening of
downward-sloping link power across the path is meant to represent how received
power varies due to factors such as multipath and presence of obstacles.
Diagrams of the type shown in Figures 2.1 and 2.2 are created for the project
and system at hand and will no doubt need updating from time to time. Documen-
tation such as the above needs to be maintained to define the “as built” state and
to track issues and failures, as well as their remedies.
We will now discuss transmitters and receivers to give an overview for all further
investigation of RFI later in this book. Radiocommunication stations are all pretty
much the same as they must contain elements that transmit as well as receive (for
20 The Radiocommunication System in an Interference-Limited Environment
��������������������������������������������������������������������
Figure 2.2 Graphical illustration of power level, in dBm, across a radiocommunication link.
the moment we ignore receive-only terminals and buoys that are used for transmit-
ting only). Figure 2.3 shows the major blocks of the transmitter and receiver, with
the channel in the middle. The upper diagram is for analog transmission, generally
obsolete in today’s environment. That said, it indicates the basics of how a user
information source is processed, modulated, and applied to an RF stage (the actual
transmitter). The receiving side is basically the reverse direction. The lower diagram
Figure 2.3 Typical transmission chains for analog and digital communications.
2.2 Transmitters and Receivers 21
is for a digital chain that consists of more elements because the user information
must be converted from analog form and then processed to reduce the bit rate
(through compression) and encoded to improve the error-rate performance over the
channel to the right. There has been a trend for the digitization and compression to
be performed within the source device, such as a digital telephone or camera. The
baseband segments of the transmit and receive chains are shown within the dotted
boxes.
We will examine the digital transmitter and receiver in detail in the remaining
chapters. To do this, we must consider how the RF stages on transmit and receive
sides of the station interact with potential and existing RFI that is prevalent over
the channel. The latter consists of all aspects of radiowave propagation as well as
the detailed properties of the transmitting and receiving antennas. Through a de-
tailed knowledge of the RFI “entries” in the channel we are in a position to assess
its impact on radiocommunication. The transmitters themselves also are capable
of causing RFI to the receiving equipment at this and other stations through inter-
modulation and harmonic generation.
The receiver and transmitter baseband sections, within the dotted boxes, are
nevertheless essential to the evaluation of RFI. Compression and coding systems
can make the link more susceptible to RFI, depending on the nature of the interfer-
ence. Noiselike RFI would be easy to evaluate because coding systems are intended
to correct errors in the presence of thermal, also called Gaussian, noise. Swept or
pulsing interference could be a different matter where the effect will depend on the
duty cycle for the presence of interference within the passband of the receiver. This
can be seen in the comparison between continuous and pulsed RFI in Figure 2.4.
Various types of pulsed interference are suggested, indicating a broad range of po-
tential sources. Continuous forms of RFI result from digital carriers including code
division multiple access (CDMA), as well as unmodulated carriers used as reference
signals (the pilot carrier on terrestrial digital TV signals being a good example).
Figure 2.4 Classification of RFI according to whether it is pulsed or continuous. Different RFI sources
are suggested for each category. Time waveforms are shown at the bottom.
22 The Radiocommunication System in an Interference-Limited Environment
��������������������������������������������������������������������
Figure 2.5 Elements of the modulator and demodulator (together, the modem) in a digital radio-
communication system. These are for the most basic form of modulation called binary phase shift
keying (BPSK).
2.2 Transmitters and Receivers 23
over 80% or 90% of its occupied bandwidth, with a symmetrical rolloff to zero.
From an RFI perspective, the signal has a rectangular shape and the total power
is the simple product of the bit rate times the constant spectral density previously
mentioned.
The demodulator section of the station is its principal receptor of RFI, along
with the noise from the received path. After the RF equipment, the carrier enters
the IF amplifier. Identified in the diagram is the bandpass filter (BPF), which sets
the input bandwidth for reception. It determines the amount of thermal noise that
the demodulator must accept and rejects out of band RFI. Any RFI and noise
within the bandwidth of the BPF passes through to the demodulator as it cannot
be eliminated through filtering. This then sets the carrier to interference plus noise
ratio of the link, the key determinant of bit error rate (BER). Following the BPF is a
level adjust or automatic gain control (AGC) that maintains constant carrier power
level into the demodulation circuitry to follow. If the level is too low, the remain-
ing components cannot extract carrier and data, and if it is too high, the devices
saturate and cannot function property.
A three-way power divider delivers the same carrier to three critical operations
in the demodulator. Actual demodulation requires a pure sinewave on precisely the
same frequency to multiply against (mix with) the modulated carrier, present on
the top arm of the power divider. The unmodulated carrier is produced by a phase-
locked oscillator in the carrier recovery circuit; the frequency reference for that
oscillator comes from the received carrier, on the middle arm of the power divider.
Modern demodulators employ direct digital synthesis (DDS) to produce an ideally
synchronized sinewave [2]. The actual carrier frequency is subject to modification
along the link: it begins within the modulator but is modified by the RF equipment
of the transmitting and receiving station, along with Doppler on the path and any
other frequency changes that might be produced in a repeater or other equipment.
Thus, the carrier recovery circuit tracks the incoming frequency to remove error
or jitter. The ability of the demodulator to recover the original baseband is related
again to the ratio of carrier to noise power; RFI will increase this degradation until
carrier recovery is not possible.
The last step in demodulation is the actual detection of the data in terms of
ones and zeros. This is accomplished by the bit decision circuit shown at the top
right. At a rate equal to the original bit rate, it measures the baseband signal at the
optimum time to capture maximum energy of each bit. Simply stated, a 1 bit is
established by a positive voltage and a 0 bit by a negative voltage. The bit timing
recovery circuit, shown at the bottom, is a standard phased-locked loop configura-
tion where the I-channel signal is phased compared with respect to the bit timing
clock stream. Like the carrier recovery loop, it obtains ideal bit timing from the
incoming carrier on the bottom leg of the power divider.
The final step of obtaining ones and zeros is by the big decision circuit in the
basic form of an integrate and dump filter. This classical device captures the en-
ergy of the bit during the symbol period to deliver a peak voltage for the decision
process. Implementation in DSP provides for proper integration with minimal loss,
and achieves ideal performance in the presence of thermal noise. A favored tech-
nique is maximum-likelihood detection, which analyzes more data to minimize the
error probability.
24 The Radiocommunication System in an Interference-Limited Environment
��������������������������������������������������������������������
This description relates to the most basic form of digital modulation, BPSK. As
we move up the chain of methods that improve the spectrum use efficiency, most of
these features are retained. The two variables used in modulation, amplitude and
phase, allow designers to squeeze more bits into the same bandwidth. But, the basic
properties of modulation and demodulation do not change. Also, the carrier itself
retains the same shape, with its nearly constant spectral density and corresponding
tendency to produce and be affected by RFI. With more efficient modulation than
BPSK, the occupied bandwidth is established by the symbol rate rather than the bit
rate. It can be seen that the symbol contains some quantity of bits (1, 2, 3, 4, 5, 6,
etc.), so the symbol rate can be expressed using the simple equation
Rb
Rs =
log2 (q)
where Rb is the bit rate into the modulator (including FEC bits) and q is the number
of phase and amplitude states (2, 4, 8, 16, 32, 64, etc.).
has automatic frequency control to address the small variation in frequency on the
carrier itself.
The block diagrams in Figure 2.6 contain amplifier stages to provide the de-
sired power level either on input or output. If this is an upconverter, than the final
power amplifier will output sufficient RF power to close the link to the distant
receiver. Most RF systems employ solid state amplifiers using gallium arsenide field
effect transistor (GaAsFET) and gallium nitride (GaN) transistors. This provides
good efficiency and linearity. The transmit chain may also include channel equal-
ization in the form of gain and phase compensation, predistortion, and other forms
of linearization.
Up until this point, we have considered only a single channel of communica-
tions, from data input to RF output, and from RF input to data output. Often,
there are multiple channels of communications involving several modems and fre-
quency conversion stages. It is possible to use a single RF amplifier to provide suf-
ficient power to several carriers, but this must consider intermodulation distortion
generated in that amplifier stage. Amplifiers are characterized in terms of their
single-carrier 1-dB compression point and third-order intermodulation intercept
point. These parameters allow the designer to size the amplifier in terms of its
maximum power and consequential intermodulation properties.
The nature and configuration of station equipment depends to a certain degree
on application. We consider here three broad categories, relating more or less to
how the spectrum is allocated by the ITU and local jurisdictions.
26 The Radiocommunication System in an Interference-Limited Environment
��������������������������������������������������������������������
Figure 2.7 Simplified block diagram of a typical wireless RF front end. (Courtesy of Microwave
Journal.)
2.2 Transmitters and Receivers 27
discussed in Chapter 7. But its effect can be no less debilitating than in land-based
systems.
Figure 2.8 presents the block diagram of a satellite repeater such as that em-
ployed in the transmission of TV and broadband data services. This is the general
class of “bent pipe” repeater on board a geostationary satellite with coverage of an
area such as a country or continent; it operates in the Region 2 Ku band used for
fixed satellite services (FSS) (uplink in the range 14.0 to 14.5 GHz and downlink
in the range 11.7 to 12.2 GHz). As such, any ground transmitter within its uplink
beam and its operating frequency band (14.0 to 14.5 GHz) will enter the repeater
and be transmitted back to Earth on a frequency equal to the difference between
the uplink frequency and the translation frequency in wideband Section 1 (shown
as 2.3 GHz). The total bandwidth of 500 MHz is divided into 12 transponders of
36-MHz bandwidth. Each transponder is provided with a preamplifier/linearizer
and a traveling-wave tube (TWT) power amplifier. More complicated types of sat-
ellites with multiple beams and frequency bands are operating, but the basic prin-
ciple of the bent pipe is almost universal. An exception is the on-board processor
(OBP) type of repeater that provides high degrees of flexibility for signal routing
and regeneration. In the design of an OBP, the details of signal structure must often
be taken into account and the trend has been away from this approach in favor of
the bend pipe. In this way, the satellite does not modify the signal itself, yet the RFI
will be transferred along with it.
There are a variety of types of Earth stations that can employ an FSS satellite
of this type. Figure 2.9 provides a simplified block diagram of a ground terminal
used for Internet access services. VSATs of this type have proliferated in recent
Figure 2.8 Block diagram of a typical satellite repeater for use in TV broadcasting or data communications.
From the uplink range of 14.0 to 14.5 GHz, the band is amplified and translated to the downlink range of 11.7
to 12.2 GHz, and then separated into 12 RF channels that are approximately 36 MHz wide; each RF channel
is amplified by its own transponder TWT amplifier, and the resulting assembly is passively combined and re-
radiated to earth in the downlink beam.
28 The Radiocommunication System in an Interference-Limited Environment
��������������������������������������������������������������������
Figure 2.9 The configuration of a VSAT for broadband satellite communications. The antenna is configured
to use linear polarization and would have a diameter in the range of 0.6m to 1.8m, depending on the fre-
quency of operation and the satellite performance.
years and are unattended. They can use the type of repeater shown in Figure 2.8 to
reach a hub station with a large antenna; the hub transmits a wideband broadcast
signal to many VSATs, which return data typically using one of the multiple access
techniques, TDMA being most common.
The most critical VSAT element with respect to RFI is the antenna itself, usu-
ally a dish-type with feed. The example in Figure 2.9 employs an ortho-mode
transducer (OMT) to combine transmit and receive waveguides using orthogonal
linear polarization (nominally vertical and horizontal). RFI is generated by or en-
ters through the reflector, both in terms of cross-polarized interference (XPI) and
adjacent satellite interference (ASI). The latter is addressed in the subsequent sec-
tion on antennas.
We introduced the antenna as the primary means of RFI production and entry. Fig-
ure 1.1 presented the coupling of EMI between two different radio systems, where
RFI is illustrated as “radiated coupling” between the antennas; all other paths are
unintentional and avoidable. According to Stutzman and Thiele [4], “The radio
antenna is an essential component in any radio system. An antenna is a device that
provides a means for radiating or receiving radio waves. In other words, it provides
the transition from guided waves on a transmission line to a “free space” wave (and
vice versa in the receiving case).”
Any conductor can transmit and/or receive if exposed to free space. What dif-
ferentiates an antenna is that it is designed to deliver, or receive, a radio wave that
can reach the intended destination or destinations. Below, we consider the wide
range of alternatives that are employed across the broad range of applications. In
addition to different intended radiation properties, there are difference in their RFI
properties, something that is of utmost interest in the field of this book.
Figure 2.10 Nondirectional antennas formed from the dipole and the monopole [5].
These are the ideal dimensions and larger or smaller dimensions can be used with
appropriate impedance matching. The dipole in (A) is oriented vertically and pro-
duces what appears to be a bidirectional pattern with nulls pointing upward and
downward (indicated as “min.” or “no-signal” in the figure). Shown at (B) is a
three-dimensional representation where the circles in (A) are rotated about the axis
of the dipole itself. This has been described as the “donut” pattern because of its
familiar shape. The dipole is often used as an element of more complex antenna
designs, as will be discussed later. Now, if we substitute the ground for the lower
half of the dipole we obtain the monopole given in the lower figure. The lower half
is actually not there, and the ground provides a reflecting plane that mirrors the up-
per half of the pattern. With a good ground (such as a metal plate of sufficient size),
the pattern has the same shape as half of a donut sliced across the axis marked
“max.” in the upper figure. So the half-ellipse for the monopole is actually rotated
360 degrees about the vertical axis of the antenna.
The size of either the dipole or monopole is related to wavelength, and we
find numerous examples of both scaled to the respective frequency. So monopoles
at 10 MHz and 10 GHz are 7.5m and 0.75 cm, respectively. Both types of anten-
nas produce linear polarization in the direction of the conductor. The monopole is
popular as the microwave waveguide transition element to allow electrical signals
to be converted to electromagnetic waves. It is the main component of the common
coax-to-waveguide adaptor.
To obtain circular polarization, we need to add a second conductor at right
angles and with the same signal either delayed or advanced in phase by 90 degrees.
Such elemental antennas are illustrated in Figure 2.11 [6]. At the top are patch
antennas of the type used to receive from GPS satellites and in two-way mobile
satellite communications. The crossed dipole at (C) and drooping dipoles at (D), re-
spectively, work in the manner just described. Finally, the quadrafilar helix antenna
2.3 Antennas in Radiocommunication Services 31
Figure 2.11 Nondirectional antennas that provide circular polarization: (a) patch antennas, (b)
patch with power divider/phase shifter, (c) crossed dipoles, (d), drooping dipoles, and (e) quadrafilar
helix.
The general principle of a passive antenna of the type illustrated is that the
solid integral of the three-dimensional radiation pattern is unity, less losses. In
equation form,
P ( θ, ϕ)
G ( θ, ϕ) = 4 π
π 2π
P ( θ, ϕ) sin θd θd ϕ
∫∫0 0
The gain in the z direction (i.e., the peak gain) can be expressed as
P (0,0)
G (0,0) = 4
π
[P ( θ,0°) + P ( θ,90°) sin θd θ
∫ 0
The numerator is simply the relative power in the maximum direction, and the
denominator is the integrated area of a cross-section of the pattern, for example,
along the y-z axis. Once the pattern is known, the peak gain can be estimated using
this technique, called pattern integration [7].
It is common practice to prepare antenna radiation patterns in two dimensions
to simplify their application in link studies. Figure 2.15 presents four common pre-
sentations of antenna radiation pattern data, each derived from the common source
such as Figure 2.14. Antenna measurements are a difficult process and so only a
portion of the pattern is actually measured. Generally speaking, we are interested
in the maximum, or peak, gain value and the level of the sidelobes that contribute
to RFI. Cross-polarization performance is important as well, but this is usually
taken separately. The polar plot in (A) and the rectangular plot in (B) are popular
as they provide a clear representation of the major lobe and sidelobes. These are
shown as relative directivity where peak gain is 0 dB and all other values are nega-
tive in dB. More commonly, the peak gain is some positive value determined by the
aperture of the antenna, its frequency of operation, and various losses. The contour
plot in (C) is useful mainly in satellite antenna radiation patterns, indicating how
well the satellite “footprint” covers an area of interest. Last, the three-dimensional
plot in (D) provides quantitative data on the entire coverage of the antenna but is
challenging to obtain and apply in given situations.
These principles apply to all directional antennas because they perform by re-
directing energy as required and yet will allow some degree of undesired radiation
through sidelobes and the like. Side and back lobe levels are usually specified by
a mask that defines the maximum allowable gain in every direction. For example,
suppose that we were to specify -20 dB as the maximum allowable sidelobe level
in (B) of Figure 2.15. That can just about be met although the design would likely
need to be modified to satisfy the -20 dB. On the other hand, we see that the ac-
tual gain in the sidelobe region is typically below -20 dB and in fact has nulls at
particular angles.
Figure 2.15 A selection of different styles of antenna radiation patterns as intended for different
applications.
direction. Figure 2.16 shows the relationship of these elements and Figure 2.17 is
a typical radiation pattern in polar form. The additional elements are termed para-
sitic elements since they intercept energy from the radiating element and re-radiate
it in a constructive manner. Aside from the horn, it is one of the basic forms of direc-
tional antenna design. It is prominent at frequencies below about 100 MHz because
of its simplicity; however, it is still useful up to about 2 GHz where it is compact and
relatively inexpensive to construct.
The log-periodic array is a variant of the Yagi which contains many linear ele-
ments all fed with power. The elements have spacing and lengths that vary according
to a logarithmic progression, producing an antenna with a very wide bandwidth.
Figure 2.16 Typical arrangement of the Yagi-Uda array, providing a radiating element, a reflec-
tor and two directors. A supporting boom is not shown but needed to hold the elements at the
proper spacing, which is of the order of a quarter of a wavelength. The elements are nominally a
half-wavelength.
Figure 2.17 Polar radiation pattern of a four-element Yagi-Uda array; 0 dB corresponds to approxi-
mately 10 dBi.
a planar wavefront in the forward direction. A familiar form is the parabolic reflec-
tor and feed, generally known as the dish. Its first applications were in radio as-
tronomy, line-of- sight microwave systems and space communications. An example
of a classic parabolic reflector antenna is provided in Figure 2.18.
The geometry of the most basic parabolic reflector system is called the prime-
focus fed design, illustrated in Figure 2.19. The feed is located at the focus of the
paraboloidal surface, F, and all points of reflection, P, produce parallel rays, P – A.
The focal length, O – F, divided by the diameter, D, determines how concave the
surface will be and has secondary impact on beam characteristics. This places the
feed at the center of the path from the reflector where it produces blockage and
scattering of energy. The dish antenna is the most popular in use at microwave fre-
quencies, particularly because of its relatively low cost for the gain produced. The
dish is exploited in a number of configurations, with three common examples being
shown in Figure 2.20. In the prime-focus fed parabolic reflector system, discussed
above and shown at the left, we see that most of the energy is directed properly
along a parallel wavefront. However, there may be spillover at the edge of the
reflector, which happens when the feed illumination is not zero at the edge of the
reflector. In addition, there is blockage by the feed which also causes scattering
due to diffraction. In the offset feed design to its right, the feed is moved below the
direction of radiation so that there is no blockage or scattering. On the other hand,
2.3 Antennas in Radiocommunication Services 37
Figure 2.18 Parabolic reflector antenna with feed and subreflector. (Courtesy of L3 Datron.)
the beam is skewed from its most desirable orientation, producing a defocusing of
the beam. This has the effect of reducing gain as compared to the center fed design.
There is a tradeoff between defocusing and feed blockage, with the optimum choice
being a result of the interaction of these factors. The folded optics design on the
far right solves the mechanical problem of feed location, allowing all electronics as
well as the feed to be placed behind the reflector. The reflector of the Cassegrain
approach is hyperbolic (convex), as illustrated to the right of Figure 2.20 and in
Figure 2.18. Another equally popular folded-optics feed design is the Gregorian,
which uses a concave subreflector located further from the reflector and causing a
crossing of the beam along the axis of the antenna. Both folded optic designs can
be optimized for the best performance through fine tuning of feed and subreflector,
and adding absorptive material around the subreflector to reduce sidelobes. This is
discussed further below.
38 The Radiocommunication System in an Interference-Limited Environment
��������������������������������������������������������������������
Figure 2.20 Three common configurations of the parabolic reflector antenna: prime-focus fed,
offset fed, and folded optics.
beam and produces sidelobes comparable in level to the co-polarized case but in-
terleaving across the angular range.
One of the primary means of reducing sidelobe levels in reflector antennas
is through optimization of the pattern illumination of the reflector by the feed.
One ideal form is presented at the top of Figure 2.22: uniform illumination that
produces a constant level of radiation on the entire reflector surface. Through the
Fourier transform, the primary radiation pattern produces a secondary pattern in
the far field which, in this case, is determined by
sin u
2D
u
Figure 2.22 Pattern data for rectangular aperture antennas (spatial Fourier transform)
40 The Radiocommunication System in an Interference-Limited Environment
��������������������������������������������������������������������
The first sidelobe is approximately 13.2 dB down from the main beam and the
half-power beamwidth follows a relationship
λ0
25.4
D
While the sidelobe level is relatively high, the directivity gain factor is optimal
at 1. This means that from the standpoint of antenna efficiency, uniform illumina-
tion is best. On the other hand, sidelobe performance is the poorest. This is seen
by looking down the listing of alternative illumination patterns, to see a progress
to lower sidelobes while efficiency decreases as well. We can obtain the lowest sid-
elobe at -32 dB at the point of poorest efficiency, 0.67. The tradeoff is therefore
clear—if low sidelobes are needed, we may design the antenna with a correspond-
ingly larger dimension to compensate for lost gain. The mainbeam beamwidth
increases as we progress down the chart, but this too can be compensated with
increased dimension.
The beamwidth relationship also comes into play if the reflector is oval or rect-
angular in shape. What happens is that the beamdwidth is determined by diameter
of the reflector in the corresponding plane. Thus, if the reflector is rectangular and
oriented horizontally, then the resulting beam is nearly rectangular, but oriented
vertically.
Figure 2.23 Manufactured microstrip PCB with the patch array and integrated feeding network.
(Courtesy CST.)
2.3 Antennas in Radiocommunication Services 41
Figure 2.24 An example of a commercial planar array antenna with two bands of operation for
90-degree sectoral coverage in the horizontal plane; each circle represents a step of 5 dB. (Courtesy
Amphenol Antennas.)
42 The Radiocommunication System in an Interference-Limited Environment
��������������������������������������������������������������������
Figure 2.25 Array geometries for scanning in one plane: (A) individual radiators, (B) columns of
elements [8].
to any angle and the pattern is little changed except for translation [8]. Four such
arrays are combined in (B) to achieve a two-dimensional array.
The performance of a seven-element linear array with half-wavelength (λ/2)
and full wavelength (λ) spacing is given in Figure 2.26 (C) and (D), respectively [9].
All elements are driven in-phase in both examples and the main-beam is directed
toward zero degrees of azimuth. Note the reasonable sidelobe performance for λ/2
spacing and the extremely high grating lobe at ±80° for λ spacing.
The direction of pointing can be changed by introducing a phase progression
across the elements of 30°. In other words, the first element on the right is at 0°,
the next at 30°, the next at 60°, and so on. As a result, the main beam steers to the
left by approximately 10°. This angle can be computed from
1 λ0
θ = sin -1 δ
2π d
2.3 Antennas in Radiocommunication Services 43
Figure 2.26 Seven-element array with no phase shift, for λ/2 (plot above) and λ (plot below) spac-
ing of elements [9].
where d is the physical spacing and δ is a increment of phase used in the progres-
sion. The resulting pattern in Figure 2.27 has a noticeable asymmetry about it with
some sidelobes increasing and others decreasing.
44 The Radiocommunication System in an Interference-Limited Environment
��������������������������������������������������������������������
Figure 2.27 The mean beam of the λ/2 spacing pattern is repointed by 10 degrees relative to the
perpendicular using phase shifters; note the loss of symmetry and increase in left-most sidelobe [9].
Phased arrays are useful for interference rejection as well as beam pointing. In
fact, both can be accomplished at the same time provided that a number of con-
ditions are met. Generally speaking, cancellation works best outside of the main
beam where it is less likely to reduce the power of the desired signal. This is illus-
trated in Figure 2.28 for a small array (six in-line elements). The quiescent pattern
represents the array performance when aligned with the desired direction and no
cancellation in action. The cancellation pattern is that which the array provides
with appropriate amplitude and phase coefficients. Finally, the synthesized pattern
shows the null that is produced in the direction of the interfering signal. Key to the
operation of this array is knowing the direction to the interference. In the absence
of this, it is still possible to introduce a processor that can dynamically adjust the
weightings to produce the null. However, this only works under a number of stan-
dard conditions. More will be said about interference cancellation in Chapters 8
and 9.
Figure 2.28 Interference cancellation by a seven six-element phased array; “quiescent pattern”
is without cancellation, and “synthesis pattern” shows how an interfering carrier at 60 degrees is
canceled [10].
Figure 2.29 Boeing Satellite Systems technician Michael Mata builds part of the Spaceway phased-
array downlink antenna, which will be integrated into a Boeing 702 satellite. (Source: Boeing.)
principle of the phased array. The beams can be created dynamically and there is
the potential to introduce beam nulling to reduce RFI on receive or transmit. The
46 The Radiocommunication System in an Interference-Limited Environment
��������������������������������������������������������������������
theory is the same as previously described but implementation is perhaps more chal-
lenging as the active elements are exposed along with the antenna.
References
[1] Cheah, J. Y.C., Practical Wireless Modem Design, Norwood, MA: Artech House, 1999.
[2] Murphy, E., and C. Slattery, “All About Direct Digital Synthesis,” Ask the Application En-
gineer-33, Analog Dialogue 38-08, August 2004, http://www.analog.com/library/analog-
Dialogue/archives/38-08/dds.pdf.
[3] Wolf, R., et al, “RF SOI Solutions as a Platform for Wireless Front-End Applica-
tions,” Microwave Journal, June 2011, p. 44, http://www.mwjournal-digital.com/
mwjournal/201106?pg=68#pg66.
[4] Stutzman, W. L., and G. A. Thiele, Antenna Theory and Design, 2nd Edition, Hoboken,
NJ: John Wiley & Sons, Inc., 1998.
[5] Cheung, W. S., and F. H. Levien, Microwaves Made Simple: Principles and Applications,
Norwood, MA: Artech House, 1985
[6] Fujimoto, K., and J. R. James, Mobile Antenna Systems Handbook, Norwood, MA: Artech
House, 1994.
[7] Ludwig, A., “Gain Computations From Pattern Integration,” IEEE Transactions on Anten-
nas and Propagation, Vol. 15, No. 2, March, 1967, pp 309–311.
[8] Mailloux, R. J., Phased Array Antenna Handbook, Norwood, MA: Artech House, 1994.
[9] Monzingo, R.t A, and T. W. Miller, Introduction to Adaptive Arrays, Raleigh, NC: Scitech
Publishing, Inc, 2004.
[10] Wang, H., and Z. Zhang, “Improved Main-Beam Nulling Through Single Switchable Dis-
placed Element for Small Scale Adaptive Array,” IEEE Transactions on Antennas and
Propagation, Vol. 62, No. 5, pp 2522–2530, May 2014.
CHAPTER 3
47
48 �������������������������������
Key Concepts for Evaluating RFI
doing this. That aside, we can expect that the majority of RFI is produced in nor-
mal operation and consists of the radiated signals themselves, along with spurious
radiation of harmonics and intermodulation products. Spurious signals can look
like communication carriers that are operating on different frequencies, but can fall
inside our operating band due to the signals and non-linearities that produce them.
The majority of communication carriers in terrestrial and space service contain
digital modulation transmitted continuously (or at least continuously during an
“on” period). That applies to phase shift keying (PSK) and amplitude and phase
shift keying (APSK) modulation and to CDMA using spread spectrum as well.
Regarding CDMA, bandwidth is determined by the chip rate rather than the bit
rate (i.e., applying the spreading factor). Systems that employ TDMA have a duty
cycle ranging from under 10% during “on” periods to as much as 95%. Of course,
it will depend if we are looking at the transmission coming from a single station
(which is on the low end of the range) or what exists over a multiple access chan-
nel containing all users as is experienced by a satellite transponder or other type
of radio repeater. The basic spectrum of a digital carrier is presented in Figure
3.1 for constant data rate and modulation characteristics and assuming a symbol
rate of unity. The parameter in this figure is the spectrum roll-off factor, α, which
represents that amount of extra bandwidth due to the raised-cosine spectrum shap-
ing. Minimum occupied bandwidth is achieved at α equal to zero. This produces a
perfectly rectangular spectrum with a bandwidth equal to the symbol rate, Rs, and
a constant power spectral density in watts/Hertz that is simply the ratio
Pc
Rs
where Pc is the total carrier power in watts. Also shown are spectra for other values
of roll-off factor between 0 and 1; however, in every case the maximum value of
power spectral density is that computed above (which is the worst case). The total
power, Pc, in every case is the area under the spectrum curve. This tends to simplify
the analysis of the impact of RFI due to common carriers of this type.
Figure 3.1 Spectrum of a digital carrier with raised-cosine spectrum shaping, for various values of
spectrum roll-off factor, α, based on a normalized symbol rate of 1. The minimum bandwidth cor-
responds to α=0 and is equal to the symbol rate.
3.1 Manmade RFI (Unintentional and Intentional) 49
We must consider the total RFI carrier power, Pc, if it falls entirely within the
bandwidth of the victim carrier. In the limiting case, the bandwidth of the RFI is
so small relative to that of the victim that it can be treated as an unmodulated
carrier. If, on the other hand, the RFI has greater bandwidth and overlays the vic-
tim, then we employ the maximum power spectrum density of the RFI across the
bandwidth of the victim. This has the effect of increasing the noise floor by the
amount of power spectral density or by an equivalent noise temperature increase.
To determine the total noise power in the victim bandwidth, take the product of the
RFI power spectral density and the victim bandwidth. Varying the roll-off doesn’t
change the total power of the carrier since it is represented by the area under the
spectrum. However, the RFI will tend to overlap more as the roll-off increases even
though its total power does not increase. The trend in radiocommunication systems
has been for low values of roll-off (≤0.2) as this improves spectrum usage efficiency.
As a result, spectrum shaping is more rectangular and the power spectral density
it produces is the constant value computed above. An interesting variant is the
North American Advanced Television Standards Committee (ATSC) digital carrier
using residual sideband modulation (Figure 3.2), composed of a digital transmis-
sion along with an unmodulated pilot carrier [3]. Both the digital signal and the
pilot must be considered individually when evaluating interference from an ATSC
signal into another radiocommunication service.
Systems that employ dynamic features to change modulation and coding will
produce and be affected by RFI in varying degrees according to the instantaneous
condition. For example, the digital video broadcasting satellite (DVB-S2) TV broad-
cast standard that has gained acceptance in many satellite applications offers a
wide range of modulation from quadrature phase shift keying (QPSK) to 64 APSK
and coding rates from ¼ to 9/10 [2]. This standard includes an optional feature
called Adaptive Coding and Modulation (ACM) that permits dynamic real-time
changes of the modulation and coding parameters in response to link conditions.
Our discussion of protection ratio (C/I), below, demonstrates that impact of circuit
quality (e.g., BER) will depend on these parameters and time-based statistics. A
Figure 3.2 The spectrum of the North American TV broadcast carrier ATSC, employing 8VSB, com-
posed of 8PSK with vestigial lower sideband.
50 �������������������������������
Key Concepts for Evaluating RFI
helpful aspect of ACM as provided through DVB-S2 satellite networks is that their
symbol rate is held constant (i.e,, constant bandwidth and power spectral density).
We depart from a discussion of RFI from real carriers to the consideration of sig-
nals that can be created or introduced through some type of nonlinear or abnormal
operation. Generally referred to as intermodulation (IM) products, they exist at
frequencies that can be predicted based on the frequencies that generate them. The
IM can come from either the system causing the RFI or the victim system that may
have susceptibility beyond what the basic communications capabilities would indi-
cate. A range of such interference sources is given in Table 3.1. The last category,
fundamental overload, is not the result of a nonlinearity in the system causing RFI
but more a consequence of proximity and/or extremely high radiated power of the
interfering carrier. It can reduce receiver sensitivity (“desensitization”) or produce
IM within the receiving system.
All forms of intermodulation are the result of the interaction of two or more
carriers on different frequencies within a non-linearity. This mechanism creates
new frequencies in the output, each being mathematically related to those present
at the input. The particular formulation depends on the nature of the nonlinear-
ity, as discussed next. Figure 3.3 presents this mechanism for two carriers that are
multiplied by each other, which is a basic nonlinear mechanism.
The nonlinearity represented by the X within a circle produces the product of
the two inputs, both of which are individual frequencies. The primary output of
the device consists of the sum and difference frequencies, the property of a double-
balanced mixer. The absence of the input frequency, which is suppressed, is clearly
undesirable in an amplifier and this type of device is more appropriate to the op-
eration of a modulator or frequency converter. A more general representation of
the nonlinearity is given in [4] in the form of a limiter that compresses the signals
as they approach a maximum or saturation level (Figure 3.4). This is typical of
Figure 3.3 The multiplication of two carriers produces a sum and difference frequency at the out-
put. This is the principle behind the operation of an upconverter or downconverter.
Figure 3.4 Definition of limiter: a hard limiter takes as input vi(t) and produces an output, vo(t) with
a constant voltage, i.e., a flattened waveform; a soft limiter takes the same input and has a smooth
limiting characteristic for the output, vo(t), similar to amplitude distortion of common RF amplifiers.
amplifiers but also occurs in metallic junctions as power approaches a critical level
to make that junction behave as a diode.
Assuming the same inputs as in Figure 3.3, the output will consist of the input
carriers along with a third carrier at a frequency equal to the difference of the two
input frequencies. This is illustrated in Figure 3.5 in terms of the frequency spec-
trum and the phasor diagram. Frequencies at a and b interact and produce an IM
product at the third frequency at c. (Another IM product will appear on the upper
side, but this nonlinearity produces less power than that of the lower sideband
IM.) The mechanism might be better understood as the response of the nonlinear-
ity to two vectors (referred to as phasors that present the amplitude and phase of
sinusoids)—the first being the vector produced by the large carrier and the second
produced by the small carrier attaching itself to the tip of the first and rotating in a
clockwise sense at the difference frequency. The nonlinearity produces the counter-
rotating dotted phasor at the difference frequency.
The characteristics for two carriers can be extended to three or more by recog-
nizing that the IM products, at a multiplicity of new frequencies, are determined
by the arithmetic permutations shown at the top of Figure 3.6. The frequency of
a given IM product is governed by the sum and difference of integer multiples of
52 �������������������������������
Key Concepts for Evaluating RFI
Figure 3.5 Representation of output signals from the non-linearity of Figure 3.4 for an input of a
large carrier a and a weaker carrier b, producing an IM product (dashed arrow); (A) presents the
spectrum of the output and (B) presents the phasor relationship
Figure 3.6 IM produced by two carriers in a soft limiter characteristic. The frequency of a given
intermodulation product is determined by the linear permutation of the carriers producing it. The
spacing between IM products is equal to the carrier spacing (i.e., f2 – f1 = Δf).
the generating frequencies, where m and n are integers the sum of which indicates
the order of the product. Thus, for m=1 and n=2, a third (3) order product exists
at 2f1 – f2 and 2f2 – f1. The fifth (5) order products are located at 3f1 – 2f2 and
3f2 – 2f1. As a general rule, the higher the order, the lower the power level of the
intermodulation product, illustrated in the output at the right. This is a good thing
since higher order products can affect more victim signals.
While the frequencies of all IM products are established by those of the signals
generating them, the levels of these products are determined by the unique transfer
characteristics (amplitude and phase) of the device. The data for relevant amplifiers
are presented in empirical graphs of the type shown in Figure 3.7. Measurements
are made with two carriers that are equal in power that are on separate frequencies
in the amplifier passband. The 3rd order intercept point (IP3), while not physical
and directly measurable, nevertheless can be projected from the data. More general
properties can be determined with a single carrier measurement of both amplitude
and relative phase shift [4].
3.2 RF Intermodulation and Harmonics 53
Figure 3.7 Typical transfer characteristics of a solid-state RF amplifier operating in its linear and
quasilinear regions. The third and fifth order intercept points indicate the virtual location of where
the single carrier linear characteristic intercepts the extension of graphs of the third order inter-
modulation product and fifth order intermodulation product, respectively. The absolute numerical
values are for illustration purposes and do not represent real characteristics; however, the slope of
the straight-line sections is correct for most solid state amplifiers.
Figure 3.8 An example of a transistor amplifier circuit intended for Class A operation but which
exhibits signal distortion resulting in the production of IM and harmonics [5].
Figure 3.9 Frequency multiplier consisting of back-to-back diodes, producing a comb of frequen-
cies at integer multiples of the input frequency. The same principle applies to any diode-like junction
in the transmit path.
Radar systems employ pulsed signals and behave in a very different way from typi-
cal communications transmitters and receivers. They also transmit this power at
elevated levels. In a study of potential radar interference to 4G long-term evolution
3.3 Interference from Radars and Other Pulsed Sources 55
(LTE) terrestrial wireless systems in the 3.5-GHz band, the authors characterized
the class of weather radar in the following manner [7]:
Traditional (legacy) long-range air-search radar waveforms often use pulse widths
(PWs) that are between 1 and 10 μs long, at pulse repetition rates (PRRs) of be-
tween 300 and 1,000 pulses/sec. The transmitted waveform duty cycle (or DC,
defined as time interval during which waveform power is transmitted divided by
total time of the waveform’s periodicity) of such radars is about 0.1 percent. Such
radars use tube-type output devices (e.g., magnetrons, klystrons, crossed-field am-
plifiers) that generate effective isotropic radiated peak power levels (peak EIRP)
of 1–10 GW (90 to 100 dBW). Newer radars often use solid-state output devices
radiating peak EIRPs of 10–100 MW (70 to 80 dBW) but at the cost of longer PWs
and higher DCs that can approach 10%. The longer PWs of current and future
solid state radars require, in turn, that transmitted pulses be frequency-modulated
or phase-coded to maintain adequate range resolution.
∞
AT nT
xp (t ) = ∑ sinc T e j 2 πfc t
T0 n = -∞ 0
The relationships of this equation are illustrated in Figure 3.10. In the first ex-
ample of radar, pulses are transmitted with a period of T seconds, which is equiva-
lent to a pulse rate of 1/T pulses per second (identified as f0 in the figure). The
pulse width is τ seconds, so that the duty cycle is simply τ/T (not as a percentage).
The spectrum plot indicates frequency components at odd multiples (1, 3, 5, etc.)
of f0; further, the envelope of the spectrum has nulls at multiples of 1/ τ, defined as
f. Summarizing this information, the location of the frequencies in the spectrum is
defined by the pulse repetition rate and the nulls in the spectrum are defined by the
inverse of the pulse duration. Reducing the pulse rate but not the pulse duration, in
the second illustration, has the effect of producing more frequencies closer together,
but the location of nulls remains the same. Finally, according to the Fourier Trans-
form of a single rectangular pulse of duration T,
56 �������������������������������
Key Concepts for Evaluating RFI
Figure 3.10 Spectra of pulsed waveforms; the top two illustrations present the Fourier Series repre-
sentation of the spectrum while the bottom figure is based on the Fourier Transform of a single pulse
[9]. The latter has no discrete frequencies as the spectrum is effectively filled in with noise.
t
rect = Tsinc ( fT )
T
the spectrum has all frequencies (the spectrum is filled in like by noise) and the nulls
remain at the inverse of the pulse duration.
This discussion explains the basic nature of radar and other pulsed signals as
they may produce RFI. As these are periodic in nature, the energy is distributed
across a wide spectrum that contains components at multiples of the pulse rate.
The effect on a victim receiver may be more complex even though the impact may
occur only during the “on” periods of the duty cycle, as discussed in Chapter 6.
the ratios by adding the noise and interference together on a power basis (in terms
of true values and not dB):
-1
C N th Iint I ext IM
= + + + +
N C C C C
The terms on the right are called interference “entries,” which recognizes that
they enter the victim radiocommunication system independently of each other. The
contributors will depend on the specific system design, but generally consist of
thermal noise, internal interference, external interference, intermodulation, and
possibly others. Each of the terms on the right side of the equation is the inverted
value of carrier to noise or interference and can be understood as a normalized
noise or interference contribution (normalized in terms of the carrier power). We
treat noise and interference as independent Gaussian random variables; for Gauss-
ian noise, the variance is the noise power. Hence, we can add powers to obtain the
total.
We discussed how RFI from radar is pulsed with a certain duty cycle and that
it is not like Gaussian noise. On the other hand, it can have a noise-like effect when
it passes through the receiver bandwidth. This effectively adds another entry in the
above equation to correspond to the “on” time of the entry.
We can see the impact of increasing external RFI by making the carrier-to-
interference (C/I) external entry into an independent variable and computing the
resulting C/N total (noise plus all sources of interference). This is given in Figure
3.11 for a constant received power level and varying the assumed external C/I from
a high initial value of 34 dB to the extremely low value of 0 dB (i.e., the total power
of noise plus interference equals the carrier power). Under the former condition,
external RFI is negligible (the right-hand side of the curve); but as we increase ex-
58 �������������������������������
Key Concepts for Evaluating RFI
ternal interference, the overall link first is unaffected until the external C/I reaches
approximately 20 dB (i.e., an increase in its level of 14 dB).
We next examine how external RFI can be mitigated through an increase in
receiver threshold. This is presented in Table 3.3 for the same link assumed in Table
3.2 and Figure 3.11, with a column added that reflects an increase in external RFI
of 17.5 dB (corresponding to a decrease in the C/I of the same amount). In order to
maintain an overall C/N at the demodulator of 14.4 dB (the bottom line in Table
3.3), the receiver threshold must increase by 9 dB. This is seen in two places in
Table 3.3: the transmitted power increases from 3 dBW to 12 dBW (i.e., from 2 to
16 watts) and the resulting received carrier power increases from approximately
–80 dBm to –71 dBm. We see that the combining of different ratios of carrier power
to noise or interference power has a softening effect in that the overall link is not
affected in equal proportion to the rise in the interference itself. However, there is
still a substantial impact on receiver sensitivity that nevertheless must be taken into
account.
This example includes an external RFI entry from an unspecified source. It
could be from a neighboring radiocommunication system that is considered during
the development of the victim system. Alternatively, it could appear in the future,
in which case the characteristics of the victim system (e.g., the receiver threshold)
3.4 Protection Ratio and Receiver Sensitivity 59
Figure 3.11 Example of the sensitivity of C/N total to the external RFI as measured by the C/I ex-
ternal; C/N thermal without RFI is constant at 16.1 dB, per link budget in Table 3.2.
may need modification. In many practical situations, the transmitter of the victim
system is boosted in power when the RFI appears. This may be impractical if the
power is already at a maximum for reasons related to the design of this particular
link or system. The course of action will generally follow a process similar to what
we discuss in Chapter 8.
The basic mechanism for receiver desensitization and overload can be under-
stood from Figure 3.7, which shows the input/output characteristic of a typical
solid state RF amplifier. The top curve applies to each of two equal carriers in the
amplifier, displaying saturation at the maximum power output. When the output
reaches saturation, there is no more power to be added and the amplifier acts as
a limiter. Intermodulation products rise to a maximum as well. For unequal input
carriers, the strong carrier dominates and will drive down the weaker carrier as the
former reaches saturation. Reception of the weaker carrier is thus degraded and
can force it below threshold in the demodulator. This suppression is in the range of
5 to 10 dB, depending on whether the strong signal is unmodulated or modulated
over some particular bandwidth [4].
An undesired strong carrier can be addressed once it is identified and its source
located. Techniques for this are discussed in Chapter 8. The primary approach is
to suppress this signal at the input to the receiving station, probably between the
antenna and the first active stage. It may also be possible to shield the antenna
from the interference if the direction is known and the shielding applied in an ef-
fective way. It may even be possible to use signal cancellation through the antenna
(discussed in Chapter 8).
3.4 Protection Ratio and Receiver Sensitivity 61
References
[1] “FCC Nixes Petition by Hotels, Disallows Wi-Fi Hotspot Blocking,” Electronista.com,
http://www.electronista.com/articles/15/01/28/preceding.rm.11737.resolved.with.public.
notice.against.blocking/#ixzz3RSberbeA.
[2] Morello, A., and V. Migone, “DVB-S2: The Second Generation Standard for Satellite
Broad-band Services,” Proceedings of the IEEE, Vol. 94, No. 1, January 2006.
[3] Sparano, D., “What Exactly Is 8-VSB Anyway?” http://www.arrl.org/files/file/Technology/
TV_Channels/8_Bit_VSB.pdf.
[4] Jo, K. Y., Satellite Communications Network Design and Analysis, Norwood, MA: Artech
House, 2011.
[5] “Amplifier Distortion,” Electronic Tutorials, http://www.electronics-tutorials.ws/amplifier/
amp_4.html.
[6] Presentation by R. F. Gillette, PE, to the IEEE Consultants Network, Austin, TX, October
2014.
[7] Sanders, G. A., J. E. Carroll, F. H. Sanders, and R. L. Sole, “Effects of Radar Interference
on LTE (FDD) eNodeB and UE Receiver Performance in the 3.5-GHz Band,” NTIA Report
14-506, July 2014.
[8] Sklar, B., Digital Communications—Fundamentals and Applications, 2nd Edition, Upper
Saddle River, NJ: Prentice Hall, 2001.
Part II
Key Aspects of Radio Wave Propagation
63
CHAPTER 4
Pt
Pr ∝
R2
where Pr is the received power as measured at the distant end, Pt is the transmitted
power in terms of the EIRP, and R is the range. The same relationship applies to the
RFI as well. This can be expressed as an equation rather than a proportionality by
taking into account the gain of the receiving antenna and the frequency or wave-
length. The range equation (also called Frii’s transmission equation wherein EIRP
is substituted by the equivalent transmit power times transmit gain) is most often
used to identify important factors in line-of-sight propagation:
2
λ
Pr = PG
t r
4 πR
65
66 �������������������������������������������
Link Properties Under Free Space Conditions
where Gr is the gain of the receive antenna and λ is the wavelength. This can be
expressed in terms of a loss factor in the denominator by inverting the bracketed
term and substituting f/c for 1/λ:
PG
Pr = t r
4 πfR
2
c
Pr = EIRP - A0
Where
This equation is strictly accurate for computing the free space loss in dB and
the received power (at the antenna) in dBW, based on a line of sight path where
the only loss is due to “spreading” of the signal as it propagates out from the
transmitting antenna. It is figured in the direction of the receiving antenna and
thus includes the gain of the transmitting antenna as part of the EIRP (see further
discussion in Section 4.2.1). Atmospheric losses, fading, blockage and other im-
pairments to the radio signal are not included but will be introduced as additional
losses in the calculation of Pr. This is the process that was employed in Chapter
3 in Tables 3.2 and 3.3. The range equation is plotted in Figure 4.1 with range as
an independent variable and frequency as a parameter for 100 MHz, 1 GHz, and
10 GHz. Owing to the square law relationship versus frequency, the curves are
20 dB apart. For example, each doubling of range increases the free space loss by
6 dB (owing to the squaring of frequency in the range equation) and likewise for
frequency. In the past, engineers have used specialized slide rules that take account
of different antenna design factors and unique aspects of the particular radiocom-
munication system. However, it is most convenient to use software for the personal
computer to compute losses for complex paths over variable Earth topography and
on Earth-to-space paths through the atmosphere. Lacking this, the analyst can still
use Microsoft Excel or a math package like MatLab with the equations and data in
this and the following chapters. Calculation of C/I, reviewed in Chapter 3, involves
two separate path calculations: one for the desired carrier, C, and the other for each
interference entry, I.
The previous discussion and analysis was for a line-of-sight path in free space and
in a vacuum. As such, there is nothing to alter the pure electromagnetic field of
the wave as it leaves the transmitting antenna, propagates through space over a
4.1 Path Geometries 67
Figure 4.1 Free-space loss calculated with the range equation (Friis equation) in dB for a range up
to 500 km and at frequencies of 100 MHz, 1 GHz and 10 GHz.
distance and is converted back into an electrical signal in the receiving antenna. The
next order of business is to establish realistic conditions for the ground and other
potential obstacles under which free-space propagation over the Earth is essentially
achieved. This brings in the concept of the Fresnel Zone [1], which is an ellipse ex-
tending between transmitting and receiving antennas that is the locus of a total path
length that is half a wavelength longer than the line-of-sight path. As such, a wave
that reflects off of a point on this ellipse arrives at the receiving antenna exactly in
the opposite phase (i.e., delayed by 180°). The dotted lines in Figure 2 represent
such a path. The radius of the first Fresnel Zone, r, is given by
d1d2
r= λ
D
where D is the direct line-of-sight path length, and d2 and d1 are the distances from
transmit and receiving antennas, respectively, to a particular point on this ellipse.
This is illustrated in Figure 4.2; the broken line segments, TP and PR, sum to be
the path length, D, plus a half wavelength, λ/2. In principle, a reflected wave will
cancel the direct wave if received at R with equal intensity. This would only happen
in practice when the transmitting and receiving antennas have the same gain along
these two paths and there is perfect reflection at the first Fresnel Zone. In practice,
the first Fresnel Zone is a generally accepted zone of required clearance to obtain
an effective line-of-sight path with no loss due to blockage. Reflection can still be
an issue and can introduce constant or time-varying fading.
In the following sections, we identify three unique types of paths encountered
in radiocommunication: ground-to-ground, air-to-air, and space-to-ground (space-
to-space is considered as well). Within each, we find a range of possibilities that
provide for an environment that facilitates propagation for the desired signal as
well as potentially interfering signals. Unguided electromagnetic waves can be
68 �������������������������������������������
Link Properties Under Free Space Conditions
Figure 4.2 The dimensions of the ellipse for the First Fresnel Zone [1].
treated independently of each other owing to linear properties of vacuum, air and
any passive medium the waves encounter. This is different from the types of nonlin-
earities in amplifiers and other devices that were discussed in Chapter 3.
Figure 4.3 Primary modes of ground-to-ground radio wave propagation (100 kHz to 100 GHz).
fc
MUF =
sin(α)
where fc = the critical frequency at the time of the communications (the lowest fre-
quency allowing vertical reflection), and α = the elevation angle of the signal radiat-
ing from the transmit antenna relative to ground. For a signal transmitted straight
up α = 90° and sin(α) = 1. In this case, MUF = fc, as it should.
As the angle α decreases, the value of sin(α) becomes smaller (less than 1),
causing the MUF to be greater than fc. This is an approximation that is useful to
obtain an idea of the relationship and actual data will differ, based on ionospheric
conditions at the time. Under these assumptions, and for the flat earth, the distance
that can be achieved at a given value of MUF becomes:
h
sin(α) =
d
2
2 + h
2
70 �������������������������������������������
Link Properties Under Free Space Conditions
This is illustrated in Figure 4.4 where the F-layer is the predominant layer of
the ionosphere at an altitude of approximately 200 km and f3 is the MUF yielding
the longest possible path.
The ability of the ionosphere to produce skywave is influenced greatly by solar
activity, measured by the number of sun spots. These exhibit an 11-year cycle that
have peaks that vary due to changes in overall solar activity, as illustrated in Figure
4.5. Sun spots and other solar activity increases the electron count in the iono-
sphere, which in turn helps skywave propagation. However, the immediate impact
of a solar event such as a coronal mass ejection, composed of gas and magnetic
flux, overpowers the ionosphere and can produce a short-term radio blackout.
Use of the HF portion of the spectrum is almost always associated with sky-
wave propagation as it provides beyond line-of-sight communications. The excep-
tion is a very short-range mode called Near Vertical Incidence Skywave (NVIS),
which provides a straight up-and-down path due to reflection off of the ionosphere
for frequencies below approximately 10 MHz. In Figure 4.4, this corresponds to
Station A and Station B being in close proximity. NVIS is consistent during day-
light hours when conditions establish the “critical frequency,” fc, for mirror-like
reflection. Skywave propagation is less likely as one goes above 100 MHz, giving
confidence that RFI will not reach the receiving point except by ground paths.
Originally a service based on manually controlled transmitters and receivers, HF
skywave is simplified by automated systems that identify the best usable frequency
and automatically establish and reestablish the link. This overcomes some of the
variability and RFI characteristics that are encountered on a day-to-day basis. Of
course, the better approach for services with greater bandwidths and the need for
high availability is to employ frequencies in the UHF range and above that offer
line-of-sight propagation using the direct wave in Figure 4.3.
Our focus in ground-to-ground radiocommunication is on paths through the
troposphere in the direct wave mode of propagation. This is actually not true line-
of-sight because the layered atmosphere acts like a lens to bend radiowaves toward
Earth. There is a variation in refractive index of air as density decreases as one
moves from lower to higher altitude, matching the decrease in temperature over the
Figure 4.4 The MUF, f3, is the highest HF that can be used on a particular point-to-point path on
Earth [3].
4.1 Path Geometries 71
same range. In still air, the refractive index will vary smoothly; or with turbulence,
the variation will be irregular so as to cause a change in the path of propagation.
See Section 4.2.5 for a discussion of tropospheric scatter, which relies on these ir-
regularities to cause a radio wave to extend beyond the horizon. The general basis
is for the smooth variation, resulting in a bending of the wave toward the earth.
Standard variation in refractive index of the troposphere is suggested by the fol-
lowing relationship [5]:
79 4,800e
(n - 1) × 106 = p-e+
T T
Engineering studies usually don’t need to go into depth on this matter and
refraction is taken into account with the assumption of 4/3 Earth curvature. This
is shown graphically in Figure 4.6 and can be approximated with the quadratic
relationship:
d2
h=
2kR
72 �������������������������������������������
Link Properties Under Free Space Conditions
Figure 4.6 Illustration of the how typical tropospheric refraction yields an effective curvature of
4/3 earth radius. What was supposed to be a path blocked by the earth (solid lines) turns out to
provide a direct wave without blockage (dotted lines). In this illustration, h corresponds to true earth
curvature (k = 1).
where h is the downward drop from a flat plane at a distance d (including refrac-
tion), k is the curvature correction factor (i.e., 4/3), and R is radius of Earth (i.e.,
6400 km).
A special condition called a temperature inversion occurs when a higher layer
is at a higher temperature than a lower layer. This can produce ducting, which is a
phenomenon whereby the signal follows Earth like a surface wave but in the VHF
to SHF range (30 MHz to 30 GHz). A path length of over 2,000 km has been dem-
onstrated over water at 144 MHz and 10 GHz [6], [7]. The paths at 10 GHz are
shown in Figure 4.7. This suggests that RFI can occur from very distant transmit-
ters, particularly for an over-the-water path. The same can be said for tropospheric
scatter, discussed in Section 4.2.5.
Terrestrial mobile wireless propagation can be complex owing to the wide vari-
ety of user conditions. We first examine how open, non-urban, conditions influence
signal transfer. Figure 4.8 presents a direct wave along with a wave that is reflected
Figure 4.7 The map shows the paths covered on 10 GHz in tests that took place in July 2010,
achieving a maximum range of 2,696 km [7].
4.1 Path Geometries 73
Figure 4.8 The direct line-of-sight path can experience cancellation from a reflected multipath
signal.
by the ground to introduce a potentially interfering signal at the receiver. The lon-
ger path length of the reflected wave will result in a time delay and, equivalently,
a difference of phase with the direct wave. Being the same carrier, the two will
combine on a vector basis at the receive antenna and produce either enhancement
or cancellation. There will be a loss of energy at the point of reflection as it will be
on the ground with less than ideal reflection properties. A representative loss is 10
dB, also called the K-factor. Another factor is the antenna radiation pattern in the
direct versus the reflected directions; both the transmitting and receiving antennas
must be considered when evaluating the relative signal strength of the reflected
wave.
A final point is that the rounding of Earth itself (and without the influence of
atmosphere) will produce a degree of diffraction or bending, allowing some energy
to extend beyond the horizon. Referred to as “smooth Earth” diffraction, it is in
addition to the effect of refraction in air with 4/3 earth radius. We will discuss the
mechanism of diffraction in Chapter 5.
Ground-to-ground radiocommunication is most popular for mobile wireless
networks, including cellular 3G and 4G. The base station antennas are arranged
in such a manner as to match height to the size of the respective cell being served.
A high tower as in Figure 4.8 creates a “macro-cell” with a radius of as much as
50 km, while microcells that are important in suburban and urban areas are much
smaller and allow greater frequency reuse across territory. The operator designs the
network to control the RFI from their distant towers that repeat the particular seg-
ment of spectrum. The spectrum of different operators, however, is segmented so
that there is no potential inter-operator RFI. This works as long as propagation
phenomena and intervening terrain maintain isolation between metropolitan areas
where the same spectrum might be reused. More will be said about this in Chapter 8.
Figure 4.10 Simplified block diagram of a conventional satellite repeater the contains several indi-
vidual transponder channels.
Figure 4.11 Frequency plan for a Ku-band geostationary satellite with 12 transponders in each of
two polarizations.
access to the Internet. Hub stations are restricted as to location as they must enter
the repeater through specific beams that are not contiguous with beams serving the
VSATs. This is shown in the coverage illustration in Figure 4.12. RFI impact here
is limited to specific uplink and/or downlink beams, an advantage over the area
coverage in Figure 4.9.
Either bent-pipe configuration, while being the most popular and simplest ap-
proach, has the property that the thermal noise on the uplink and downlink add on
a power basis. This is because uplink noise is transferred through the repeater to
the downlink along with the signal. On a spectrum analyzer connected to a down-
link ground antenna, you can see the uplink noise adding to the noise floor within
the bandwidth of the transponder. RFI on the uplink will likewise be transferred
4.1 Path Geometries 77
Figure 4.12 Multiple beam coverage of a high-throughput satellite (HTS) that provides Internet
access from locations in the eastern half of the United States and the West Coast. The gateway hub
stations that provide Internet access are located in the western Great Basin region to reuse spectrum.
to the downlink provided that it is within the transponder bandwidth. There is an-
other class of repeater not employing the bent-pipe within which the uplink signal
is demodulated and reduced back to the original data. Referred to as demod-remod
or regenerative repeater, the uplink and downlink are isolated from each other in a
noise and interference sense.
With the launch of the Iridium system in 1998, the satellite communications
industry saw a rebirth of non-GEO systems and a few others have since come or
are planned. Unlike the GEO approach, low Earth orbit (LEO) and medium Earth
orbit (MEO) systems employ moving satellites (relative to the ground), so their user
antennas either must have broad beams or the ability to maintain pointing at the
satellite in use. Iridium, illustrated in Figure 4.13, is a narrowband system that sup-
ports handsets for voice service and other simple terminals for low-speed data ap-
plications. By virtue of having 66 satellites in 6 near-polar orbits, the constellation
illuminates the entire globe (and is the only system to do so). The use of L-band
frequencies around 1.5 GHz is beneficial to propagation because it is insensitive to
rain and allows for some diffraction and penetration of nonmetallic (e.g., wooden)
structures. It is able to close the link because of the nature of the service and the low
altitude of 780 km. But the link geometry is highly dynamic, ranging from as much
as 3,000 km as the satellite rises from the horizon to as little as the altitude, which
causes the path to experience a variety of impairments due to multipath fading,
obstruction by buildings and vegetation, and the complexities of handoff among
the satellite’s beams and between satellites (which incorporate intersatellite links).
The satellites are also packet switches and perform the demod-remod approach
as a result. The line-of-sight properties of a LEO system like Iridium exist but are
confounded by the need to communicate through and around obstacles.
Another non-GEO system, O3b, employs 12 MEO satellites that orbit at an
altitude of 8,000 km in the plane of the equator. This system serves ground loca-
tions ranging approximately 8,000 km from the equator and provides Ka-band
78 �������������������������������������������
Link Properties Under Free Space Conditions
Figure 4.13 The Iridium constellation employs six near polar orbits to provide coverage of the
entire surface of Earth. The solid lines for the ground-to-space link are at L-band (1.5 GHz) and the
space-to-space intersatellite links are at Ka-band (20 GHz); gateway Earth stations also communicate
with the satellite in the Ka-band. Each satellite has 48 beams over a total footprint that covers ap-
proximately 2,200 km.
bent-pipe capacity to tracking antennas. With at least two antennas per station,
end users can maintain continuous service as satellites rise from and fall below the
horizon, provided that the two locations have a view of a common satellite. Paths
are intended to be line-of-sight as any blockage will severely impair communica-
tions. Adequate margin for rain attenuation is required due to the frequencies em-
ployed. The non-GEO concept for broadband service at Ka band was under con-
sideration at the time of this writing for direct-to-user service and in the backhaul
segment of cellular telephones.
Figure 4.14 The basic arrangement of spacecraft and ground stations in the TDRS system, owned and operated by NASA.
79
80 �������������������������������������������
Link Properties Under Free Space Conditions
also provides links to users within the atmosphere of Earth, including balloons,
unpiloted aeronautical vehicles (UAVs), and a ground station at the South Pole.
All paths are true line-of-sight, making the service very reliable in terms of
propagation; on the other hand, there are limits as to duration of a path due to the
motion of user spacecraft in LEO, which make a complete orbit in approximately
90 minutes. From an RFI perspective, TDRS is reasonably protected because the
frequencies have received international recognition through the ITU and are dedi-
cated to NASA by the NTIA of the U.S. Government. The ground stations are
located in fairly isolated areas like White Sands, NM, and Owens Valley, CA.
One of the more fascinating approaches used for planetary missions involves
a relay satellite revolving about the planet in question. The repeater provides the
path back to a ground station on Earth. Operation at the remote planet could be
viewed as RFI free provided that the internal operation from the planet’s floor to
the repeater is clean. From the repeater to a DSN antenna on Earth, the path is
much the same as any deep-space mission. The dishes at DSN stations tend to be
larger than those used in commercial communications and are treated more like ra-
dio telescopes. They tend to have lower sidelobes as well, owing to their design for
low noise and high efficiency. These factors can mitigate RFI from the surrounding
area.
As a final note on space-to-ground radiocommunication, the GPS system em-
ploys L-band spectrum and may be a factor in RFI studies. These satellites effec-
tively broadcast important ranging signals from MEO, and there are over a billion
receivers throughout the planet (with some in Earth orbit as well). The system has
dedicated spectrum but can be susceptible to out-of-band RFI at the receiving end
of the link. Russian and European GPS systems are to be considered as well.
We’ve looked at basic link properties for an unobstructed path and see that there
are variations based on the locations of the terminating points. The next step is to
consider the technical characteristics of the stations and atmospheric effects that
introduce attenuation through absorption and scattering of RF energy. As discussed
for refraction above, the latter property of the troposphere attenuates signals and
allows RFI to propagate along otherwise obstructed or extended paths. These come
into play for each of the systems discussed previously, depending on the particu-
lar link geometry and presence of absorptive or ionized media that surround our
planet.
only a few items of information. The block diagram of a typical transmitting sta-
tion, regardless of application and how it is employed, is illustrated in Figure 4.15.
Like most two-way communication stations, it employs a single antenna but takes
advantage of the reciprocity of the device to combine both transmit and receive
functions. The component that allows the transmitter (TX) and the receiver (RX)
to connect to the antenna is referred to as a duplexer (DUP). Two forms of DUP are
found in the variety of wireless services: frequency division duplex (FDD) and time
division duplex (TDD). These are defined as follows [8]:
•• FDD uses the idea that the transmission and reception of signals are achieved
simultaneously with two different frequencies. Using FDD, it is possible to
transmit and receive signals simultaneously as the receiver is not tuned to
the same frequency as the transmitter as shown. For the FDD scheme to
operate satisfactorily, it is necessary that the channel separation between
the transmission and reception frequencies must be sufficient to enable the
receiver not to be unduly affected by the transmitter signal. This is known
as the guard band.
•• TDD is used with data transmissions (data or digitized voice), transmitting
a short burst of data in each direction. With this constraint, the duplexer
is nothing more than a T/R switch. The transmission periods are relatively
short to minimize time delay on voice transmissions. While FDD transmis-
sions require a guard band between the transmitter and receiver frequencies,
TDD schemes require a guard time or guard interval between transmission
and reception. This must be sufficient to allow the signals traveling from the
remote transmitter to arrive before a transmission is started and the receiver
inhibited. For systems communicating over short distances (e.g., a few ki-
lometers), the guard interval is normally small and acceptable. For greater
distances it may become an issue. The guard interval required for TDD will
comprise two main elements:
• A time allowance for the propagation delay for any transmissions from
the remote transmitter to arrive at the receiver. This will depend upon
the distances involved, but it takes 3.3 ms to travel a kilometer, 5.4 ms
to travel 1 mile.
Either technique can be used over land, sea, and air paths. LEO Iridium actu-
ally used TDD since it provided the dual benefits of a low-cost DUP in the handset
and established the link in both directions with the same frequency. Multipath, in
particular, is frequency sensitive and would have caused greater difficulties in two-
way voice connections. FDD is the standard technique beyond LEO and requires
separate transmit and receive frequencies, the scheme contemplated in most ITU
frequency allocations for space communications. The DUP is actually a set of fil-
ters (and possibly an orthomode transducer that combines two polarizations). In
terrestrial radiocommunication (including OFDM), FDD is needed for broadband
communications owing to the difficulty of fast switching of transmit and receive
and the bandwidth required to meet service requirements.
Referring to the details of Figure 4.15 and considering only the transmit di-
rection (from left to right), we first encounter the transmitter itself. As discussed
above, its output power would be measured in one of the standard forms but can
be converted to dBW for this calculation. The carrier passes through a transmis-
sion line with a fixed attenuation (loss), Lt , the loss of the DUP, Ldup, and the loss
of the transmission line, Lg, to the antenna itself. These values correspond to the
transmission band that the system is intended for. The EIRP for this transmitting
station is simply
This equation assumes that there are no mismatch losses due to reflections at
junctions of cables and devices or at the antenna itself, which has a gain of Gt.
Any spurious outputs of the transmitter might be attenuated in the DUP, but this
is generally not assumed because the unit is designed to combine transmit and re-
ceive with minimal loss. A transmitter that emits out of band will probably require
a bandpass filter with the concomitant loss to be included in the computation of
EIRP. The transmission line that extends to the antenna itself will depend on the
application. If the antenna is located on a tower and the transmitter is at the base,
which is typical of wireless and microwave installations, then this line will be one
with low loss per foot or meter. For frequencies below approximately 6 GHz, rigid
coaxial cable with air dielectric and a helical spacer (heliax) is generally preferred.
At 6 GHz and above, the line would be waveguide sized for the particular transmit
and receive frequencies. These transmission lines provide very effective isolation
and protection from RFI that may couple in from around the locality.
The antenna on the right in Figure 4.15, discussed in detail in Chapter 2, pro-
vides the transition between electrical signals and electromagnetic signals, using
the principle that this conversion should be as efficient as possible. The maximum
gain of the antenna, Gt(0), for a circular aperture can be estimated with the for-
mula provided in Chapter 2 and typically demonstrates an aperture efficiency of
between 55% and 70%, or an equivalent loss ranging from a high of 2.6 dB to a
low of 1.5 dB. Antennas can do better or worse, based on the specific construction
and frequency of operation. Antenna gain is measured in the far field and typically
4.2 Line-of-Sight Propagation Characteristics in Atmosphere 83
quoted as the peak value at a specific frequency (because gain typically increases as
the square of the frequency, in nominal terms) in a specified direction. Gain at other
frequencies may not even be quoted and could be a concern with respect to out-of-
band radiation or coupling to space. The peak gain is important for point-to-point
links and occurs if and only if the antenna is aligned properly. Owing to installation
errors and the effects of propagation through the atmosphere, however, there can
be variation in gain due to the angle of arrival of the wave departing from ideal.
This can be addressed by allowing an error, such as the 3-dB point on the antenna
pattern, represented by the 3-dB beamwidth, θ3db. The loss for this angular offset
is, by definition, 3 dB. The actual angle involved can be estimated from
22
θ 3dB =
fD
where f is the frequency in GHz and D is the aperture diameter in meters. The actual
antenna pattern should be consulted when designing or evaluating a real system and
patterns of the type provided in Chapter 2 should be consulted. Of importance are
the sidelobes and backlobes shown dotted with the antenna in Figure 4.15. Their
contribution to RFI will be discussed in detail in Chapter 8.
Table 4.1 Comparison of Various Measures of Received RF Power and Signal Strength
Measure Form Application
Received power (dBW) 4 πR Link calculations and gain
Pr = EIRP - 20log10 + Gr budgets
λ
manipulation of the equations. Typical applications are indicated on the third col-
umn and will be used in the following chapters.
A0 + Aw
A=
sin(ϕ)
where φ is the elevation angle from the ground. A0 and Aw are attenuation values to
the zenith from the oxygen and water vapor curves, respectively, at the frequency of
operation; their sum is simply the total curve value. Both this relationship and the
graph are valid for elevation angles between 5° and 90°; below 5°, use Figure 4.16
and calculate the total attenuation based on path length.
of rain rates, with tropical regions having the greatest intensity. But temperate cli-
mates also experience thunderstorms that produce heavy rain and the attenuation
that comes with it. Rain attenuation is usually presented in terms of a probability
of a certain rain rate or greater occurring, expressed as the specified value in mm/
hour, or less. On this basis, you can plan on the link operating reliably if you design
86 �������������������������������������������
Link Properties Under Free Space Conditions
Figure 4.17 Total, dry-air and water-vapor zenith attenuation from sea level (ITU-R).
for this amount of rainfall for a specified probability, which is called the availability.
The rain rate can be turned into a corresponding rain attenuation value based on
the frequency and the elevation angle of the path through the rain. A horizontal
path in terrestrial communications would experience the most water and therefore
have the highest rain attenuation for a given percentage of time. As the elevation
4.2 Line-of-Sight Propagation Characteristics in Atmosphere 87
angle increases toward 90°, attenuation is reduced because there is less water to
overcome.
Rain attenuation is a strong function of frequency and rain rate, as illustrated
in Figure 4.18 [13]. The details for calculating the specific rain attenuation from
the rain rate can be found in ITU-R Recommendation P.838. Missing from the data
is the path length to be used to covert specific attenuation into an absolute value.
The characteristic of rain is that the path length is itself a function of rain rate and
other factors. Figure 4.18 indicates that generally rain attenuation increases mono-
tonically up to about 150 GHz, at which point it flattens. Rainfall at 150 mm/hr,
the highest on the graph, corresponds to tropical regions around the equator and
for perhaps only 0.01% of the time. Rates near 50 mm/hr are experienced during
thunderstorms in temperate climates.
As a simple example, consider a line-of-sight 14-GHz path that traverses a
raincell of 2-km dimension. Assume the location is in the Northeastern United
States with a corresponding rain rate of 50 mm/hr at 0.01% outage. Figure 4.18
gives a specific attenuation of approximately 2 dB/km, yielding a total rain attenu-
ation of 4 dB. This is based on an estimated path length through the cell of 2 km,
but this will depend on local conditions. Other examples are provided in [14].
The way rainfall distributes across Earth is illustrated in Figure 4.19, which is
taken from the commonly applied ITU Dissanyake, Allnutt, and Haidara (DAH)
model [15].
Figure 4.18 Specific attenuation due to rain versus frequency for various rainfall rates.
88 �������������������������������������������
Link Properties Under Free Space Conditions
What remains is a means to compute the total rain attenuation for a given
path, including its elevation angle if applicable. Much work was done for space-to-
ground communications, and this can be extended to terrestrial paths as well. The
ITU DAH model is the best place to start because it is the most extensively used on
a global basis. There is still the possibility of local conditions impacting the rain
on the path, and there is no substitute for experience with the particular location
and systems.
Rain also has the effect of scattering radiowaves, causing them to traverse be-
yond the local horizon. This, or course, only happens when it is raining but may
need to be considered when frequencies are reused with a range of approximately
100 km.
Figure 4.20 Illustration of the principle of over-the-horizon propagation with tropospheric scatter
and the common volume.
References
[12] “Specific Attenuation Model for Rain for Use in Prediction Methods,” ITU Recommenda-
tion ITU-R P.838.3, 2005.
[13] Kadish, J. E., and T. W.R. East, Satellite Communications Fundamentals, Norwood, MA
Artech House, 2000.
[14] Freeman, R.L., Telecommunications Transmission Handbook, 4th Ed., New York: John
Wiley & Sons, 1998.
[15] Allnut, J. E., Satellite to Ground Radiowave Propagation, 2nd Ed., Institution of Engineer-
ing and Technology, London: Peter Peregrinus, Ltd, 2011.
CHAPTER 5
We now consider the types of obstructed radio paths experienced on Earth, both
terrestrial and space-to-Earth. The discussion of line-of-sight propagation through
atmosphere still applies and the losses of the previous chapter must be included as
well (unless the obstructed path substantially alters the environment). Geographic
features like hills, mountain ranges, and valleys have a pronounced impact, par-
ticularly as frequency increases above about 30 MHz and especially in the micro-
wave range. Their characteristics have been well understood with the availability of
global 3D mapping and GPS. Buildings also introduce blockage and are important
in mobile systems in city streets and where interior penetration is desired. This is
difficult to predict and often relies on direct measurements. Vegetation brings in
another variable as there can be a seasonal variation, along with growth patterns
that take years. Finally, the precise location of an end user is another variable:
where they stand or place antennas, including use inside vehicles and other modes
of transportation. Consider also that RFI usually comes from a different direction
than the desired signal, resulting in wide variation in antenna gain and the value of
C/I that exists for a particular situation and timing. The static situation is one that
can more easily be addressed than one where either the user or the cause of RFI,
or both, are on the move. Another variable is due to the propagation path itself,
especially with regard to infrequent phenomena like ducting and rain scatter that
allow RFI to appear in an otherwise isolated system. These factors are considered
in models developed over the years and reviewed in this chapter.
The path profile is the primary tool for understanding the natural obstructions
that attenuate radio signals, introduce diffraction paths, and offer the potential for
extended range. It is plotted as the cross-section of Earth along the path between
transmitter and receiver. The transmitter could emit the desired signal or it could
be the source of RFI. Likewise, the receiver is the point of reception of desired and
undesired signals. As a result, we need to understand potential paths, for every
location where operations are intended. The tools discussed at the end of this chap-
ter offer an automated means of preparing and evaluating profiles; otherwise, we
91
92 �������������������������������
Propagation on Obstructed Paths
must use topographic maps of the type used in surveying and by the military. An
example of a hand-drawn path profile including 4/3 Earth curvature, as prepared
by the author in 1967, is given in Figure 5.1. This illustrates the virtual line-of-sight
along the bottom, which is blocked by the mountains, and the diffraction path over
the highest peak. The latter has the effect of avoiding the second, lower, peak. The
amount of diffraction loss can be computed using the procedure given in Section
5.3. Note that this path is beyond line of sight due to low height of the endpoints.
Diffraction loss is a straightforward calculation, based on the same theory as
affects the diffraction of light [1]. Figure 5.2 illustrates the arrangement of the geo-
metric elements and their relationship to the Fresnel Zone, discussed in Chapter 4.
The upper illustration shows how the obstacle blocks the line-of-sight path; yet,
some signal will propagate through diffraction with the amount depending on the
height, h, relative to the dimensions of the first Fresnel Zone. In the lower illustra-
tion, the path is not blocked, but the obstacle introduces some loss that is again
dependent on the degree of impingement of the first Fresnel Zone.
The dimensionless Fresnel-Kirchoff diffraction parameter, υ, is a necessary step
to calculate the diffraction loss:
2(d1 + d2 )
ν=h
λd1d2
The diffraction loss will depend on whether h is negative, as in the lower figure,
or positive, as in the upper figure. The following formulae can be used to calculate
the loss in dB as a function of the diffraction parameter, υ:
Figure 5.1 A hand drawn path profile that includes an obstacle and is corrected for 4/3 Earth
curvature.
5.1 Path Profiles and Obstructions 93
Figure 5.2 Illustration of geometries for diffraction paths: blocked path and line-of-sight path with
partial blockage [1]. The dotted ellipses represent the first Fresnel Zone.
0.225
20log10 , υ ≥ 2.4
υ
The combination of these equations are plotted in Figure 5.3 to cover the range
from the first impingement on the first Fresnel Zone, to the obstacle just touch-
ing the line-of-sight path (which is a loss of 6 dB a v = 0), to blockage that ranges
through total blockage of the signal.
Knife-edge diffraction requires good clearance at the respective antennas and
only the presence of the obstacle or obstacles over the path. In the latter case, each
knife edge is computed separately and their total loss simply added. This may pre-
dict more loss than actually experienced, as in the case of the path in Figure 5.1.
The large obstacle alters the path and any remaining obstacles need to be treated
with respect to a more complex geometry [2]. Obstacles are important sources of
isolation that protect one operating radiocommunication system from potential
RFI from others. Real geography can provide very stable isolation and is often the
best source of such protection.
Propagation across smooth Earth at frequencies above approximately 100
MHz is governed by a combination of free space and diffraction but is different
owning to the lack of a knife-edge terrain feature. This is illustrated in Figure 5.4
for a smooth Earth path with antennas elevated to a height of h1 and h2 in meters
and a total distance of D in km. Distances d1 and d2 correspond to the range from
94 �������������������������������
Propagation on Obstructed Paths
–5
–10
A(v), dB
–15
–20
–25
–2.0 –1.0 0.0 1.0 2.0 3.0
Fresnel-Kirchoff Diffraction Parameter, v
Figure 5.3 Knife-edge diffraction loss.
Figure 5.4 Smooth Earth path based on the model proposed by Bullington [3]. Segments d1 and d2
are line-of-sight paths from the respective antennas to the horizon, which d3 is the hidden or beyond
line-of-sight segment.
antenna to the horizon, and include K factor of 4/3 Earth curvature. Distance d3 is
that portion of D that is beyond line of sight for the total link. We provide sample
calculations using the procedure developed by Bullington [3] using Figure 6 (re-
printed below as Figure 5.5), a nomogram for determining each of three contribu-
tors to the total attenuation. This is in addition to the free space loss calculated
with the formula in Chapter 4. The original nomogram, created in 1957, uses feet
and miles and is fairly straightforward in its application. The same method was
made available in a slide ruled developed by J. Deygout [4]. We provide another
example in Table 5.1 using metric values: total path length, 100 km; height of an-
tenna, 50 m on both ends; and three different frequencies of operation, 100 MHz,
1 GHz, and 10 GHz. The results are presented in Table 5.1 and present diffraction
loss alone; the free space loss for the path is fixed and can be calculated from the
equation in Chapter 4.
According to the Bullington method, each segment of distance produces an
amount of loss, as indicated in the columns. The total loss for the path, exclu-
sive of the free space component, is simply the sum of the three segment losses.
We see from Table 5.1 that total smooth Earth attenuation increases rapidly with
frequency for column d3 (i.e., in which there is no line of sight). Operation in the
VHF spectrum still offers reasonable propagation across smooth Earth, but the
attenuation almost doubles in dB for an increase in frequency from 1 GHz to 10
GHz. Note also that free space loss increases as well, at the rate of 20 dB per factor
of 10 increase in frequency. This overall behavior is consistent with how spectrum
is reassigned across geography, but ignores the potential for knife-edge diffraction
from tall hills and buildings and other modes of propagation such as ducting or
scatter that can provide higher levels of RFI than are predicted in this manner. The
models and software discussed later in this chapter provide a means of evaluating
more terrain options.
5.1 Path Profiles and Obstructions 95
Figure 5.5 Diffraction loss relative to free space transmission at all locations beyond line-of-sight
over a smooth sphere [3]. Height in feet and distance in miles; frequency in Mc (MHz).
The topic of multipath was introduced in Chapter 4 and depicted for a line-of-sight
path in Figure 4.8. We have reproduced this in Figure 5.6 and will show two impor-
tant fading mechanisms in radiocommunication in the presence of reflections alone
and under the influence of heavy obstructions; these are Ricean fading and Rayleigh
fading, respectively. Either form of fading occurs through the vector addition of
multiple samples of the same radio signal which result from the aforementioned
reflection and obstruction. It is vector addition because the signal at its particular
frequency can be represented by its amplitude and phase, presented as a phasor.
The tower in Figure 5.6 could be a terrestrial repeater or base station, or it could
represent a communications satellite or aircraft.
Ar e j ϕr = A1e j ϕ1 + A2 e j ϕ2
Figure 5.6 Simplified diagram depicting a mobile receiver that is subject to multipath.
5.2 Fading on Direct and Blocked Paths 97
Figure 5.7 Vector combination of a direct signal and its multipath that combine at the receiving
antenna.
apply to an interfering signal as well. This argues for additional protection margin
if multipath is anticipated.
Figure 5.8 S-band (2.2 GHz) TDRS to handheld signal variations: transition, inside to outside [5].
98 �������������������������������
Propagation on Obstructed Paths
level. Note also that there are peaks of signal enhancement of up to about 3 dB,
behavior that is expected in light of the phasor diagram in Figure 5.7.
The periodic nature of the trace is the result of the slow motion of walking
as the reflected signal vector rotates around the direct path signal vector and is a
consequence of the particular wavelength and velocity of the walker. There is also
a fixed point of reflection such as another building causing the geometry to change
with motion of the receiver. Alternatively, if the point of reflection was from flat
ground in the direction of the satellite (i.e., a point that moves with motion) the
amount of attenuation (or enhancement) would have been constant or unvarying.
This is because the satellite is so distant (36,000 km away) that the angle of arrival
at the receiver does not change with motion.
The Ricean fading mechanism is well understand and its behavior is fairly be-
nign. It can be addressed with a reasonable amount of margin, typically 6 dB, so
that both terrestrial and satellite links can be maintained in its presence. Interfer-
ence conditions at the receiver may involve more dynamic properties, particularly
if the desired signal is distant but the RFI is close.
Figure 5.9 Radiocommunication in the wireless world often involves using a non-line-of-sight path
to the emitter, relying instead on signals that have been absorbed, reflected and diffracted before
reception. There is also the presence of additional such signals that combine on a vectoral basis,
producing fading that is dependent on position and is widely variable.
5.2 Fading on Direct and Blocked Paths 99
internal repeaters is even more complex and difficult as the ability to receive ad-
equate signal (direct or indirect) depends on the precise location inside and the
construction of the building itself.
A phasor diagram for Rayleigh fading is a jumble of vectors that represent frag-
ments of the original signal that have experienced a variety of propagation mecha-
nisms. As illustrated in Figure 5.10, the resultant, Ar , can actually be stronger than
what exists of the direct signal, A1. If all elements are static, then the resultant will
also be static; however, if any contributor is variable (changing power, direction,
distance, velocity, etc.), that the resultant can experience wide variations and at
times disappear altogether. This is illustrated on the left side of Figure 5.7 where we
see more than 10 dB of attenuation relative to free space combined with even deep-
er fades of the signal as one moves around the interior of the building in question.
Further to the subject of building penetration, Figure 5.11 shows locations and
orientations of structures in the same testing performed at Jet Propulsion Laborato-
ry in La Cañada, California, using the TDRS satellite. Its low elevation angle of 22°
makes this relevant for terrestrial radiocommunication as well as space to ground
communication. Engineering measurements were made with a handheld receiver as
the holder worked through and outside of Building 161. In addition to blockage by
the building walls, the path was at times blocked by a high rise building located to
the west-southwest. These features produce a fading environment that covers liter-
ally all facets, especially Rayleigh fading within much of the building. The actual
Figure 5.10 How the various signal fragments combine in a blocked environment and produce
Rayleigh fading.
Figure 5.11 Satellite-to-indoor measurements: JPL building 161 picture, floor plan and handheld
test route [5].
100 �������������������������������
Propagation on Obstructed Paths
measured data covering the path is presented in Figure 5.12. The X-axis presents
the time of day as the holder of the receiver proceeds along the described path. This
produces the variations in signal strength over a 50-dB dynamic range, allowing a
precise study of the Rayleigh-induced signal fading.
The text boxes in Figure 5.12 indicate the location of the receiver, beginning
from outside the building and entering some interior offices. Without blockage
from the adjacent structure, the signal levels appear to average about 15 dB of at-
tenuation with some excursions below –30 dB. The section on the second floor out-
side stairs presents minor shadowing loss of about 3 dB along with Ricean fading
of up to 6 or 8 dB, and one point of near cancellation at –20 dB. Upon reentering
the building and being blocked by the adjacent eight-story building, the signal level
drops below –20 dB with many nulls of over –40 dB to as much as –50 dB. This
is a region of extremely poor coverage. Upon exiting the building to the northeast
and proceeding to the open portion, the signal gains line of sight to TDRS and the
fading is again Ricean. These measurements provide the reader with a good picture
of how the fading mechanisms perform on a real world path.
Figure 5.12 Satellite-to-indoor measurements: JPL building 161 signal power time series [5].
5.3 Geographic Coverage Analysis 101
mathematical sense. Hence the need for these models for estimating propagation
over a wide area. The good news is that the RFI is only one component of the total
C/N, which provides some mitigation in real world operations.
There have been attempts to do this by establishing categories of the kinds of
environments that would be encountered in a given radiocommunication system.
These could include such distinctions as: open and unobstructed, open with foli-
age, suburban with trees but without structures, suburban with structures up to
two stories, urban with buildings up to eight stories, and urban concrete canyons.
There is also the question of the locations of the end points: outside structures in
the open, inside buildings, exterior to moving vehicles or aircraft, or inside a means
of transportation. Buildings offer a wide range of variables that affect penetration:
wood frame structures, exterior rooms in concrete structures with windows, inte-
rior rooms in concrete structures, high floors versus low floors, and a variety of
variables having to do with the nature of the windows and ceilings. Much of this
was investigated in connection with the mobile satellite systems under development
in the late 1990s [6]. However, just applying a few words as a descriptor leaves out
much of what actually differentiates environments given similar names. The inter-
action with the design of the radiocommunication system itself is critical as well.
We review below the techniques used to establish geographic coverage in various
environments, recognizing that the most effective methods use computation and/
or measurement based on the actual conditions of the radiocommunication system
and the RFI that can affect it.
Figure 5.13 Antennas at the Fillmore, CA, earth station. The RFI victim antenna is pointed at an
Intelsat satellite at 180° longitude and is close to the horizon.
source identified was from S-band radar systems on board U.S. Navy ships when in
harbor at Port Hueneme, CA. The problem was resolved by installing a bandpass
filter between the antenna feed and the low noise amplifier, a common solution
to problems of this nature. This was suspected because the antenna in question,
outlined in Figure 5.13, has an alignment that suggests a high level of reception
through the antenna close-in sidelobes.
We used this case as a basis to re-create the problem using Terrain Analysis
Package (TAP™), PC-based terrestrial RF propagation software from Softwright
LLC. The line-of-sight path across the terrain from Port Hueneme to Fillmore is
given in Figure 5.14, indicating that the intervening mountains block the signal.
This means that propagation will include a substantial diffraction loss across the
mountain peaks. The actual profile is plotted using TAP in Figure 5.15.
The computation of losses by TAP for two frequencies, 300 MHz and 3 GHz,
is summarized as follows:
These losses at 300 MHz and 3 GHz are extremely high and more than enough
to protect the receiver at Fillmore. So, the question is, how can the RFI in fact
happen? If we look at the map, it is possible to recognize that there is a clear path
5.3 Geographic Coverage Analysis 103
Figure 5.14 The path between Port Hueneme and the Fillmore Earth Station, indicating obstruction
by an intervening mountain range. However, there is a clear path to the ocean along the valley that
includes Ventura, CA.
Figure 5.15 Screen shot of the path profile computed by TAP, a PC-based terrestrial RF propagation
software from Softwright LLC. Note that the profile has been corrected for 4/3 Earth curvature. The
intervening mountain range provides substantial blockage, particularly at frequencies above 1 GHz.
across the valley encompassing Ventura. The most likely source of the radar signal
is from the ship when it is in coastal waters away from port but in the line of sight
afforded by the valley path. This theory has not been verified, a step that might be
needed if the filter had not cured the problem.
follow the same process but are done in rapid succession to give the appearance of
a blanket of radiation to see where there might be gaps and holes. Once this is done,
an operator would use scouting vehicles to make measurements around the cover-
age area to validate the prediction and help understand if additional base stations
are needed. The same process can be used to evaluate how undesired signals might
extend beyond the intended area of coverage and present RFI to neighboring users.
Spectrum managers and regulators have the difficult task of assigning bands
of frequencies to different users who in many cases must use the same frequencies.
This can be accommodated if there is sufficient terrain isolation between areas on
the same bandwidths. We have discussed how propagation can extend beyond line
of sight through mechanisms such as refraction, diffraction, ducting and scatter.
The practice is to establish boundaries based on generalized models of coverage,
which are discussed in the next section. This can be validated or modified through
detailed investigations along point-to-point paths across the geography using soft-
ware such as TAP. Figures 5.16 and 5.17 compare two approaches for service in
a common area [7]. The FCC model in Figure 5.16 portrays clean coverage of the
area around the northeast corner of Jefferson County, CO, with some variation due
to a general trend of geography. It indicates areas of high signal strength as well as
areas where the signals are perhaps measurable but not of high intensity. Beyond
the outer area, no signal is predicted into the surrounding counties.
When we look at a more accurate portrayal of coverage, however, taking ac-
count of actual paths emanating from the central point, signal strength variation
is much greater and range can extend well beyond the expected limits posed in the
FCC model. Note that the area of most intense coverage around the center is very
similar between the two approaches. The extensive departure occurs beyond line of
sight where propagation over and around terrain obstacles dominates. A compre-
hensive discussion and presentation of the various propagation models at a detailed
level can be found in [8].
Figure 5.16 FCC site general coverage based on complex propagation model [7].
5.4 Complex Propagation Models 105
Figure 5.17 Site-specific model prepared with terrain analysis software from EDX Wireless [7].
The evaluation of propagation and RFI over wide areas has been a challenge to the
rapid expansion of radiocommunication. An approach is to employ some type of
comprehensive model that can be agreed-upon and employed at least as a starting
point in the system development and spectrum management process. We review
a few of the more prominent models below; several references are suggested for
further study and application; [8] in particular, presents an excellent discussion
of models and their relative merits. Many of the details of the propagation char-
acteristics were given previously, including relevant formulas and data. However,
the application of the models is still problematic owing to the irregular behavior
of physical topography and characteristics of the radio systems themselves. The
atmosphere also produces considerable variability. The following caveats are given
in the fine paper cited here:
•• None of the models studied provided data on sector transmission loss dis-
tributions due to multipath reception, except to assume a loss deviation of
8.2 dB (which is well below expectation for conditions of Rayleigh fading).
The models we have chosen, particularly that of the Institute for Telecom-
munications Sciences (ITS) of the US National Institute of Science and Technology
(NIST), formerly the National Bureau of Standards (NBS), have fairly wide accep-
tance for general studies of radiowave propagation in the terrestrial environment.
In addition, they are best used to perform general and preliminary studies of the
RFI that would exist in or could be caused by operation of a radiocommunication
system. The ITU-R offers their methodology for predicting interference between
radio stations on the surface of the Earth, employing various propagation models
as previously discussed [9].
•• Urban: “Built-up city or large town crowded with large buildings and two-
or-more-storied houses, or in a larger village closely interspersed with houses
and thickly-grown tall trees.” These are subclassified for the purposes of cor-
recting for mobile antenna heights as follows:
• Large city
• Medium city
•• Suburban: “Village or highway scattered with trees and houses—the area
having some obstacles near the mobile radio car, but still not very congested.”
•• Open: “No obstacles like tall trees or buildings in the propagation path and a
plot of land which is cleared of anything 300 to 400m ahead, as, for instance,
farmland, rice field, open fields, etc.” These are subclassified as follows:
• Open
• Quasiopen (defined as midway between Open and Suburban)
5.4 Complex Propagation Models 107
The basic Okumura model employs a simple equation in dB, similar in con-
cept to the link budget of Chapter 4. The formula addresses the free space loss
as well as a terrain loss term for the median loss from an urban environment.
The assumed transmit and receive antennas have elevations of 200 m and 3 m,
respectively. There are adjustment terms, discussed below, to account for different
antenna heights (more loss if the antenna is lower, less loss if the antenna is higher).
The total path loss for a terrestrial radio link from a base station tower to a mobile
terminal can be expressed as
where
L50 is the median path loss between transmitter and receiver;
LFS is the free space computed from the range equation, reduced by the sum of
the gain of the transmit and receive antennas;
Amu is the basic median attenuation due to propagation in an urban environment;
Htu is the height gain correction factor (variation from the assumed base sta-
tion height of 200m); and
Hru is the height gain correction factor (variation from the assumed mobile
terminal antenna height of 3m)
The basic median attenuation is taken from a graph in Okumura, while Hata
has developed curve-fitted equations that eliminate the graphs. Figure 5.18 is from
[11] and presents the basic mean attenuation as a function of frequency with path
Figure 5.18 Basic median attenuation as a function of frequency and path distance. After Okumura
[11].
108 �������������������������������
Propagation on Obstructed Paths
length from 1 to 100 km as a parameter. As an example from the graph, the basic
median attenuation at 1,000 MHz (1 GHz) and a range of 60 km is approximately
50 dB. This accounts for the existence of terrain in the urban environment and is in
addition to the free-space loss, which can be computed as 128 dB. Summing these
values gives 178 dB as the median path loss between transmitter and receiver for
the assumed heights and environment.
The height gain correction factors can be estimated from data presented in Fig-
ures 5.19 and 5.20 for the base station and mobile terminal antennas, respectively.
The effective antenna height is calculated as the height of the antenna’s radiation
above the average terrain, calculated along the direction of the radio path for the
distance between 3 and 15 km of the respective antenna. A base station antenna
above 200m yields a positive gain, while one below 200m yields a negative gain.
Applying Figure 5.19, at 100m height and a 1-km distance, the base station an-
tenna gain, Htu, is approximately -4 dB. In the case of the vehicular antenna (Figure
5.20) at 800 MHz and with a height of 5 m yields a gain of approximately 2 dB
relative to the 3m reference antenna in a large city (e.g., Tokyo). According to [13],
corrections in Figures 5.19 and 5.20 are seldom used.
These graphs are for the urban/large city environment. The basic median
field strength will be increased (reduced attenuation) for suburban (6 to 14 dB),
quasiopen (18.5 to 27.5 dB), and open areas (23.5 to 32.5 dB). According to [14]:
Care should be exercised in selecting the area type to use, keeping in mind the fact
that these definitions are based on Japanese environs. Application to U.S. and other
cities should make appropriate adjustments.
As discussed previously, the Hata models provide empirical formulas that replace
the graphs found in Okumura. These employ linear and parabolic approximations
Figure 5.20 Mobile station height correction gain (after Okumura [11]).
that are within 0.5 dB to 1 dB of the original data. Owing to the greater variability
and uncertainty of our knowledge of the precise topography and atmospherics,
these errors can be ignored.
•• Frequency;
•• Polarization;
•• Path length;
•• Antenna heights above ground;
•• Surface refractivity;
110 �������������������������������
Propagation on Obstructed Paths
There are two prediction modes: point-to-point and area coverage. To perform
the analysis using point-to-point mode, the actual path profile must be known so
that individual paths can be evaluated. While unavailable at the time, the graphical
computing tools given in Section 5.3.1 are invaluable for point-to-point evalua-
tions. Lacking these specifics, the area coverage mode allows prediction of cover-
age based on assumed topographical environment characteristics as in the Oku-
mura model. Such a comparison was given in Figures 5.16 and 5.17.
The documents are all available on line and software is available both as free
download and from suppliers like Softwright LLC. and EDX Wireless Inc., so we
will not reproduce the details here. They are also subject to many revisions and
extensions, and it can be concluded that the expression “the devil is in the details”
clearly applies here. The U.S. DoD has also developed its own model called Terrain-
Integrated Rough Earth Model (TIREM), described in [17]. These approaches have
the benefit of giving answers when data is lacking or incomplete but bring with
them uncertainty of the type discussed in Section 5.3. This uncertainty translates
into RFI experiences that might not have been detected based on analysis using
area-coverage type analysis. Point-to-point analysis, therefore, will always be pre-
ferred and is how we addressed the existence of RFI in the example in Section 5.3.1.
The age of computer analysis and large databases has greatly eased the chal-
lenge of coverage and link studies. With the right tools, almost any radiocommuni-
cation system can be evaluated as to the range of transmission, both for the desired
and undesired signals. The art keeps advancing every year, based mainly on the
sophistication of data processing, manipulation and presentation. The underlying
theory, on the other hand, never changes.
References
[1] Mararic, K., EMI Protection for Communications Systems, Norwood, MA: Artech House,
2010.
[2] Deygout, J., “Multiple knife-edge diffraction of microwaves,” IEEE Transactions on An-
tennas and Propagation, Vol. AP-14, No. 4, July 1966.
[3] Bullington, K., “Radio Wave Propagation Fundamentals,” Bell System Technical Journal,
Vol 36, No. 3, May 1957.
5.4 Complex Propagation Models 111
[4] Deygout, J., Summary of Speech Given by Lt. Colonel J. Deygout, French Liaison Officer,
at USASCS on 5 February 1964, http://www.applicationstrategy.com/deygout.
[5] Bell, D., “Satellite Communications Networks and Applications—Satellite Mobile Com-
munications,” UCLA Extension Engineering, 2006, http://www.applicationstrategy.com/
bell.
[6] Goldhirsh, J., and W. J. Vogel, Handbook of Propagation Effects for Vehicular and Per-
sonal Mobile Satellite Systems—Overview of Experimental and Modeling Results, The
Johns Hopkins University Applied Physics Laboratory, and the University of Texas at Aus-
tin, A2A-98-U-0-021 (APL) EERL-98-12A (EERL), December 1998.
[7] Greg Leon, Director, EDX Wireless Inc., “Spectrum Management - Propagation and Day
to Day Spectrum Management,” TLEN: 5230.
[8] IEEE Vehicular Technology Society Committee on Radio Propagation: Coverage Predic-
tion for Mobile Radio Systems Operating in the 800/900-MHz Frequency Range,” IEEE
Transactions on Vehicular Technology, Vol. 31, No. 1, February 1988.
[9] Prediction procedure for the evaluation of interference between stations on the surface of
the Earth at frequencies above about 0.1 GHz, International Telecommunication Union,
Recommendation ITU-R P.452-16 (07/2015).
[10] NTT Technical Review, Vol. 11, No. 4, April 2013.
[11] Okumura, Yoshihisa, et al, “Field Strength and its Variability in VHF and UHF Land-
Mobile Radio Service,” Review of the Electrical Communication Laboratory, Vol. 16,
Numbers 9–10, Sept–Oct, 1968.
[12] Hata, Masaharu, “Empirical Formula for Propagation Loss in Land Mobile Services,”
IEEE Transactions on Vehicular Technology, Vol VT-29, No. 3, August 1980.
[13] Kostanic, Ivica, “Okumura and Hata Macroscopic Propagatioin Models,” ECE 5221, Per-
sonal Communication Systems, Spring 2011, Florida Institute of Technology, http://my.fit.
edu/~kostanic/RF%20Propagation/Old%20Notes/RF%20Propagation%20-%2007-Oku-
mura%20and%20Hata%20Macroscopic%20Propagation%20Models.pdf
[14] “Notes on Okumura Propagation,” Softwright LLC, 1997, http://www.softwright.com/
faq/engineering/prop_okumura.html.
[15] Rice, P. L., A. G. Longley, K. A. Norton, and A. P. Barsis, Technical Note 101, “Transmis-
sion Loss Predictions for Tropospheric Communication Circuits,” Volume 2, National Bu-
reau of Standards, issued May 7, 1965.
[16] Hufford, G.A., et al, A Guide to the Use of the ITS Irregular Terrain Model in the Area
Prediction Mode, NTIA Report 82-100, April 1982.
[17] Eppink, David and Wolf Kuebler, TIREM/SEM Handbook, Department of Defense, Elec-
tromagnetic Compatibility Analysis Center, IIT Research Institute under contract to the
Department of Defense, March 1994.
Part III
RFI Assessment and Resolution
Methodologies
113
CHAPTER 6
Eb C B
= ×
N0 I Rb
Es C B
= ×
N0 I Rs
115
116 Interference
�������������������������������������������������������
Protection Ratio (C/I) and Its Application
Figure 6.1 General relationship of carrier to interference, both in band and in the adjacent channel. Inter-
ference that occupies a bandwidth greater than the desired signal can usually be treated as noise, which is
shown as a constant power spectral density. The C/I is the ratio of desired power (under the matched filter
curve) divided by the noise in the same equivalent bandwidth.
where Rb is the information bit rate prior to forward error correction coding, and
Rs is the total symbol rate once the carrier is modulated. The relationship between
these values is
Rb
R
Rs =
log 2 (q)
where R is the code rate and q is the number of phase and amplitude states of the
modulation (q = 2, 4, 8, 16, 32, 64, etc.).
Figure 6.1 shows upper and lower adjacent channels that are contiguous with
the symbol rate bandwidth of the desired carrier. While this is feasible for certain
types of modulations, it is a common practice to allow for roll-off of the filter
bandwidth and place the adjacent carriers accordingly. The roll-off factor, α, ac-
commodates the filtering along with providing control of intersymbol interference
(ISI) in the demodulated signal.
When the RFI occupies less bandwidth than the desired carrier, the value of I is
simply the total power of the interfering carrier. We treat this just like the equiva-
lent amount of noise from a broadband interferer within the same bandwidth (i.e.,
N0B). This is a reasonable and perhaps conservative assumption for conventional
digital demodulators of the type discussed in Chapter 2. Certain radio systems, par-
ticularly radar, behave differently due to their pulsed nature, as discussed in Section
3.3 . We present an example later in this chapter.
Another special case arose regarding interference from a wideband FM car-
rier sweeping across a narrow-band digital transmission using a technique known
as single channel per carrier (SCPC). The case is described in Section 7.5.2 as an
6.1 C/I Criteria 117
example of how hardware testing can be used to verify a prediction based on the
previous assumptions about a narrow or unmodulated carrier have the same effect
on BER as an equivalent amount of thermal noise in the receiver bandwidth. Basi-
cally, the BER increased only during the time that the interfering carrier was within
the victim demodulator bandwidth; conversely, when the carrier swept outside of
this bandwidth, the BER returned to its uninterfered-with state.
Many radiocommunication systems are subject to multiple sources of interfer-
ence in addition to the thermal noise present within the system itself. A methodol-
ogy for adding noise and interference powers is discussed in Chapter 4, wherein
each interference “entry” is treated as an independent source that can be added on
a power basis with thermal noise to produce a total value of C/(N+I).
One of the more complex and potentially contentious areas of RFI investigation
and mitigation is the issue of the proper value of C/I to be used. As stated above,
there are often multiple RFI entries and associated sources, and each may need to be
evaluated individually. Then, the total must be considered against what is produced
at the input to every demodulator on the receiving end. We can look at this in terms
of the worst case which produces the lowest possible value of C/I. If a given system
is not affected by the worst case RFI, then all is well. On the other hand, it is often
the situation that most receivers will operate effectively but perhaps a limiting or
even nonexistent case may end up driving the study and ultimately be the bone of
contention between the parties involved. In the latter case, more study and measure-
ment is probably called for as there are likely to be answers that allow the victim
and interfering systems to coexist. This is, after all, the ultimate objective of RFI
investigation and resolution.
A good example of how a worst case analysis is used as an initial filter for de-
termining if there could be unacceptable RFI is the approach within ITU regulation
of geostationary satellites. ITU member administrations should assign frequencies
to operators within their borders (or under their jurisdiction) in a way that doesn’t
cause unacceptable interference to the operation of an existing system in another
country. The Advance Publication of Information (API) initial stage of obtaining a
satellite orbital location is described in Article 9, Section 1, of the Radio Regula-
tions [1]. An administration is required to compile technical information about the
planned satellite and Earth station network and provide this to the ITU for review
and publication to all administrations. From there, the data are used by other ad-
ministrations to determine the potential for unacceptable RFI using the technique
described in Appendix 8. Boiling this all down, the methodology provides a worst
case assessment of how much the total received noise in the victim receiver will
be increased by the maximum possible power spectral density in that system that
could be produced by operation of the new entrant. The threshold is set at 6%,
which is equivalent to an offending C/I which is 12 dB higher than the total C/N of
the victim system. The latter is derived for a link operating with the planned C/N
(with margin) while the former corresponds to the highest possible power spectral
density, assumed to exist at the same frequency. As a result, the threshold C/I would
probably be a relatively high value (making the coordination threshold low). The
118 Interference
�������������������������������������������������������
Protection Ratio (C/I) and Its Application
Figure 6.2 The desired carrier with thermal noise (constant spectral density, N0), a narrowband interfering
carrier, and a wideband interfering carrier.
6.1 C/I Criteria 119
-1
C Nth I
= +
N C C
-1
C N I
∴ = -
Nth C C
where C/Nth is the thermal C/N (excluding any interference entries) and C/I is the
ratio of carrier to total interference power in the bandwidth of the carrier. Note that
these are true values and not in dB. This expression gives us the necessary value of
C/Nth before interference is included and for a specific desired value of link C/N.
If that value of C/N is assume to be 10 dB, Figure 6.3 presents the relationship be-
tween C/Nth and C/I per entry, for cases of a single RFI entry (N = 1) and 10 RFI
entries (N = 10). We use the simple mathematical model which assumes that noise
and interference can be added on a power basis, which is justified in the common
case of independent signals. The graphs are presented in terms of dB, allowing us to
look over a relatively wide range of C/Nth (10 to 15 dB) and C/I per RFI entry (10
to 35 dB). A review of the two curves in Figure 6.3 can be very instructive about the
interplay of basic link characteristics in the presence of noise and RFI. The lower
curve assumes one entry while the upper (dotted) curve assumes 10 RFI entries. On
a simple basis with RFI entries that are of equal power but uncorrelated (i.e., from
independent sources), each of the ten entries can only be one tenth the power (e.g.,
C/I 10 dB greater) than the case of the single RFI entry. As a general rule, the more
entries, the higher the C/I per entry must be (because they add on a power basis to
produce an effective total C/I reduced by 10 log10(N), where N is the number of
entries). Stated another way, the two curves are separated by the same 10 log10(N),
which is 10 dB in this example.
The case of N = 1 will come about where there is a single interference source
degrading service; it is addressed by increasing link C/N in the absence of RFI to the
120 Interference
�������������������������������������������������������
Protection Ratio (C/I) and Its Application
Figure 6.3 The relationship between C/Nth and C/I per entry, for cases of a single RFI entry (N =
1) and 10 RFI entries (N=10). Total link C/N is fixed at 10 dB to represent an overall service quality
requirement.
amount shown on the X-axis corresponding to the value of C/I on the Y-axis. For
example, an RFI condition of C/I = 20 dB with one entry (N = 1) requires that the
link operate at a C/Nth greater than or equal to 10.5 dB (assuming the combined
link C/N = 10 dB). This would require a nominal increase of 0.5 dB in link power,
or about 10%. On the other hand, the condition of C/I = 13 dB requires a 3 dB
increase in C/Nth to the same value of 13 dB. This is expected since both the power
of the RFI and the noise in the bandwidth of the carrier are equal, resulting in a
combined C/N that is 3 dB less than either. The asymptotic nature of the graph to
the right indicates that one cannot allow a C/I approaching 10 dB, for which the
thermal noise would have to be zero in power terms (or –∞ in dB).
Taking a look at the dotted curve for 10 RFI entries, the C/I of each must be
10 dB greater than for the N = 1 example below it. The horizontal line presents
how a C/N = 10 dB can be achieved with C/I = 21 dB for both conditions. However,
the C/Nth would have to be about 5 dB greater to accommodate 10 such carriers.
This seems like a good trade because 10 dB in RFI can be accommodated with only
about 5 dB of carrier power, which is a factor of about 3 (rather than 10). The rea-
son for this is the presence of the thermal noise in the link which helps compensate
for RFI. All of this discussion is meant to provide some insight into the practical
relationships between these critical variables. These values interact in the manner
presented and the analyst must examine all cases that can arise for a particular
system.
There is a special case for multiple RFI carriers if they are unmodulated (CW).
Pure sinusoids when added together can produce peaks in voltage and hence power
6.1 C/I Criteria 121
that can overdrive receiving devices that lack sufficient dynamic range. For exam-
ple, two CW interferers if equal in power cause peak values according to the square
of the voltage (i.e., 6 dB). On the other hand, if the RFI signals are modulated as
they would be if they contained digital data, then the summation used in Figure 6.2
will be appropriate. Section 6.4 provides a real example of RFI that is Gaussian
noise, CW, and pulsed or swept,
Figure 6.4 FER as a function of Eb/N0 for the case of perfect power control versus power control
with 1 dB error standard deviation [2].
help accommodate this fade by increasing transmit power, thus raising the Eb/N0
at the base station. However, as more stations employ this strategy, the total power
radiated increases as does the interference. The system tends quickly toward over-
load and capacity drops rapidly. Alternatives to this involve complex strategies that
group users according to common properties, and grant priorities based on service
levels and the like. Dynamic adjustment of MODCOD and throughput are also
prevalent and effective countermeasures.
computed using a link budget for the desired system (discussed in Chapter 4), while
I is computed for the path taken by the interference. The following equation is used
to represent the interference, where the primes (‘) indicate interference (I) source
and direction):
4 πR′
I = Pt ′ + G′ ( θ ′ ) - 20log10 + G ( θ ) dB
λ
The illustration and equation are deceptively simple since only a few parame-
ters are involved and the relationship is obvious to one familiar with radio systems.
The challenge, however, is to lay the problem out precisely and accurately, and to
obtain the requisite values of antenna gain and power. Calculating free space loss,
4 πR ′
20log10 , is a simple step and yields a very accurate value (based only on
λ
knowledge of the frequency or wavelength, and the distance). What we have not
considered so far are the propagation characteristics involved in the paths for both
C and I. The gains, G(θ) and G′(θ′) for the victim and interfering antennas, respec-
tively, are in directions that correspond to their paths of interference coupling. The
respective sidelobes of these antennas are effectively opposite each other, yielding
a transmission path not unlike one for actual communications. Before we can ad-
dress suppression of this path of distance R′, we must understand its nature and
constituents.
Lb ( p) = Pt + Gt + Gr – Pr ( p) dB (1)
is the gain toward the physical horizon on a given azimuth; for a receiving terres-
trial station, the maximum main beam axis antenna gain is to be used. In the case
of a receiving Earth station, the permissible interference power Pr(p) is specified
with respect to the actual percentage of time the receiver is in operation, rather
than the total elapsed time.
For propagation mode (2), a volume scattering process is involved and a modi-
fication of the above approach is necessary. Where the coordinating earth station
antenna beam intersects a rain cell, a common volume may be formed with a ter-
restrial station beam or an earth station beam (operating in the opposite direction
of transmission in bidirectionally allocated frequency bands). In the case of a ter-
restrial station, the assumptions are made that the terrestrial station beamwidth is
relatively large in comparison with that of the coordinating earth station and that
the terrestrial station is some distance from the common volume. The terrestrial
station beam is therefore assumed to illuminate the whole rain cell, which is repre-
sented by a vertical cylinder filled with hydrometeors that give rise to isotropically
scattered signals. This scattering process may give rise to unwanted coupling be-
tween the coordinating Earth station and terrestrial stations or other Earth stations
operating in bidirectionally allocated frequency bands, via the common volume.
•• Fundamentally, BER has an inverse relationship with the total value of C/N
in the bandwidth of the carrier. Service quality in digital services is directly
related to the BER, as well as packet error rate (PER). Established values and
curves of this performance are available for commercial modem products
and wireless systems that employ them. The C/N must be maintained high
enough so that BER is satisfactory; a drop in C/N from this threshold condi-
tion will result in reduced data throughput and loss of link synchronization.
•• C/I is combined with link C/N without interference using the approach given
in 6.1.1. The resulting value of C/(N+I) will determine the BER for the link.
•• There are aspects of link quality and reliability that go beyond simple BER
prediction. Pulsed or sweeping RFI need to be considered based on the tim-
ing of the pulse or sweep (i.e., the temporal characteristics) along with its
impact on actual receivers and demodulators. The case of radar interference,
covered in Section 6.4.1, is a good example. Another example is provided in
Section 7.5.2 for SCPC satellite communications with swept FM interference.
126 Interference
�������������������������������������������������������
Protection Ratio (C/I) and Its Application
We can discuss RFI and its impact in abstract, but it is probably more instructive
to look at some real systems and the interference situations they have faced. RFI is
a part of how typical radiocommunication systems are designed, implemented and
operated. These examples provide a jumping-off place for looking at systems that
differ in important ways; however, the principles as discussed previously do not
change.
Figure 6.6 Schematic diagram showing coupling scenario between littoral (coastal) radar transmit-
ters and possible future 3.5 GHz LTE systems. By geometry, coupling from radar transmitters should
occur more into LTE enhanced Node B (eNB) base stations that into UE located on low terrain.
aircraft could easily experience RFI similar to that received by the eNB illustrated
in Figure 6.6.
NTIA provides the following insight into the RFI-producing characteristics of
radar (discussed in Section 3.3).
Traditional (legacy) long range air-search radar waveforms often use pulse
widths (PWs) that are between 1 and 10 μs long, at PRRs of between 300 and
1,000 pulses/sec. The transmitted waveform duty cycle (DC, defined as time inter-
val during which waveform power is transmitted divided by total time of the wave-
form’s periodicity) of such radars is about 0.1 percent. Such radars use tube-type
output devices (e.g., magnetrons, klystrons, crossed-field amplifiers) that generate
effective isotropic radiated peak power levels (peak EIRP) of 1–10 GW (90 to 100
dBW). Newer radars often use solid-state output devices radiating peak EIRPs of
10–100 MW (70 to 80 dBW), but at the cost of longer PWs and higher DCs that
can approach 10%. The longer PWs of current and future solid state radars re-
quire, in turn, that transmitted pulses be frequency-modulated or phase coded to
maintain adequate range resolution. Solid-state radars can cause more interference
than higher-power tube-transmitter radars because the peak EIRPs of the newer ra-
dars are still high enough to exceed interference thresholds for many systems, while
their higher DCs of around 10 percent have more impact in the time domain than
the 0.1 percent DCs of older radars. Radar pulses were modeled as either fixed-
frequency carrier waves, designated by NTIA as [an unmodulated pulse sequence
containing no information] P0N or [an angle-modulated carrier with a single chan-
nel but containing no information, e.g., chirping] Q3N [5].
The investigation centered on actual testing of radar interference into the
aforementioned LTE base station. The radar interference waveform parameter
space was designed to span the entire range of all existing and possible fu-
ture radar waveforms that do/may occur in 3.5-GHz band. Measurements in-
cluded LTE data throughput, block error rate (BLER), and modulation and coding
scheme (MODCOD) usage for both the downlink and uplink, and additionally UE
transmit power and resource block (RB) usage for the uplink. Each RB represents a
block of bandwidth equal to 180 kHz which is dynamically assignable by the eNB.
These diagnostics were recorded for a wide range of interference power levels, I,
by varying the interference duty cycle (DC) values in an approximately logarithmic
progression: 1%, 3%, and 10%. Pulse width (PW) and pulse repetition rate (PRR)
values were adjusted to achieve these DCs. For the P0N pulses, setting DC was
128 Interference
�������������������������������������������������������
Protection Ratio (C/I) and Its Application
straightforward: DC = PW ∙ PRR. The P0N pulses were centered at 3,600 MHz for
the downlink testing and 3,520 MHz for the uplink testing. Two additional P0N
tests were performed on the downlink path at +2 MHz and +4 MHz offsets from
3,600 MHz (3,602 MHz and 3,604 MHz, respectively). The target value of BLER
was 10%.
Measured spectra for the carrier transmitted by the base station (eNB down-
link) is shown in Figure 6.7 under peak and average conditions. The spectrum is
similar to Worldwide Interoperability for Microwave Access (WiMAX) emission
spectra presented in a recent NTIA Report [5], which employ orthogonal frequency
division multiplex (OFDM). It is also noted that the spectrum is from a prototype
3.5-GHz LTE eNB.
The emission spectrum structure of the LTE UE transmitter, as measured with
both peak and average detection, is shown in Figure 6.8. While the UE transmis-
sion is centered at f0 = 3,520 MHz, the lowest (in frequency) 10 resource blocks
(RBs) were unused, leaving only the upper 40 RBs active. As each RB is 180 kHz
wide, the total occupied bandwidth of the emission spectrum of the UE is 7.2 MHz,
causing the center of the intentional emissions to appear to be at 3,520.9 MHz. It
should be noted that in the presence of interference, this spectrum can change dra-
matically as the UE changes the number of RBs in use. A change in RBs will change
the total occupied bandwidth of the UE emission spectrum
BW = RB ⋅ BWRB
Figure 6.8 LTE user terminal transmitted spectrum (uplink from UE).
Figure 6.9 Data throughput and BLER for 10-MHz-wide Gaussian noise interference to the LTE UE
receiver.
Figure 6.10 Data throughput and BLER for P0N-4 (PW = 10 ms, PRR = 1,000/sec, DC = 1.0%) in-
terference to the LTE UE receiver.
6.4 Interference Among Various Systems 131
Figure 6.11 Data throughput and BLER for Q3N-1 (PW = 10 ms, PRR = 1,000/sec, DC = 1.0%)
interference to the LTE UE receiver.
modulation (called chirped radar and designated Q3N) into the LTE UE receiver
where the BLER blows up to greater than 50% in a 20 dB range of interference
power. This unexpected behavior could not be explained by the authors from the
data and must relate to the specifics of link dynamics and modem design. This
would require detailed investigation of receiver properties, which was not part of
the investigation.
Measurements were also conducted in the uplink direction, where the UT is the
transmitter and the eNB base station is the receiver. The system is very dynamic and
adaptive in nature, so the best measure is how throughput is maintained during RFI
exposure. The received signal level at the eNB was –85 dBm. Figure 6.12 presents
the case with wideband Gaussian noise interference while Figure 6.13 presents the
case with swept radar-type interference (Q3N), again displaying a substantial ad-
vantage for chirped interference as compared to continuous noise. With Gaussian
noise, throughput measured in RB transfer falls off when interference exceeds -80
dBm; while it does not vary with chirped radar even as the interference exceeds -30
dBm.
The NTIA concluded that theoretical analysis is recommended to better un-
derstand why various radar interference waveforms have the particular effects that
have been published in this report. LTE signal detection needs to be understood to
perform such analyses. We learn a lot about how the radiocommunication system
behaves in the presence of the expected RFI environment. Clearly, LTE and radar
in the same band have significant sharing problems. The challenge remains to com-
pletely understand what is going on and how the system will respond when oper-
ating in a live environment. Sharing concepts for future systems are presented in
132 Interference
�������������������������������������������������������
Protection Ratio (C/I) and Its Application
Figure 6.12 Data throughput, UE transmit (Tx) power, BLER, and RB usage for 10-MHz-wide Gauss-
ian noise interference.
Figure 6.13 Data throughput, UE transmit (Tx) power, BLER, and RB usage for Q3N-1 (PW = 10 ms,
PRR = 1,000/sec, DC = 1%) interference.
6.4 Interference Among Various Systems 133
Figure 6.14 Hybrid terrestrial/satellite wireless network based on the ATC [6].
134 Interference
�������������������������������������������������������
Protection Ratio (C/I) and Its Application
6.4.2.2 Assessing RFI from MSS into GPS: Technical and Business Issues
An independent technical working group (TWG) examined the performance of GPS
receivers in the presence of LightSquared’s terrestrial operations [9]. The TWG
testing, discussed in detail in the TWG Report, shows a wide range of individual
GPS receiver performance, illustrating conclusively that the problem is one of GPS
device design and characteristics, rather than an issue of fundamental incompatibil-
ity between LightSquared’s planned transmissions and all GPS receivers. The tested
GPS receivers had varying abilities to reject LightSquared’s transmissions depend-
ing, as expected, on the nature of the filtering employed in the receivers as well as
other aspects of receiver design.
Although the results vary among devices, the overall conclusion of the test-
ing is that transmissions in the upper 10-MHz channel—the channel nearest to
the 1,559-1,610-MHz GPS band—would adversely affect the performance of a
Figure 6.15 LightSquared Signal Spectral Occupancy; from “Preliminary Results on Overload Char-
acteristics of GPS Receivers in Proximity to LightSquared’s L-Band Terrestrial Base Stations (BTS) and
User Equipment (UE),” 3GPP TSG-RAN4 #54AH R4-110470, Austin, TX, USA Jan. 17–21, 2011.
6.5 Satellite Communications vs. Terrestrial Microwave Systems 135
significant number of legacy GPS receivers. The tests confirm that this problem is
not caused by emissions from LightSquared’s base stations into the GPS band, but
from the failure of these legacy GPS receivers to reject transmissions from Light-
Squared’s licensed frequencies, which are adjacent to the spectrum allocated for use
by GPS. Ultimately, the FCC pulled the ATC authorization, and LightSquared was
unable to introduce terrestrial services using MSS spectrum.
The FSS was established before 1965 using C-band (6-GHz uplink and 4-GHz
downlink) to provide spectrum for introduction of long-distance satellite commu-
nications between fixed locations on Earth. It allowed the establishment of the
global system originally known as the International Telecommunications Satellite
Consortium (INTELSAT), a joint-venture organization owned by major public and
private telecommunications operators in the respective countries that INTELSAT
interconnected. The particular spectrum provided by the ITU for this service previ-
ously had been allocated to the terrestrial Fixed Service (FS), which basically was
for point-to-point microwave links of the same operators. The FSS was co-primary
with FS, which is to say that the spectrum was to be shared.
The forced partnering of FSS and FS of necessity introduced requirements for
coordination, the complicated ITU process to identify and resolve international
conflicts between radiocommunication services that result from RFI. Many rules
were put in place to manage the interference:
•• Space to Earth,
•• Earth to Space, and
•• Earth to Earth.
The approach depends on the use of fixed locations that are coordinated and li-
censed by the respective nations. To control interference, C-band Earth stations had
to be located in remote places that had isolation from terrestrial microwave links.
This system worked fairly well but tended to hinder application of space commu-
nications, until an unshared Ku-band allocation arrived on the scene around 1972.
An important and often-challenging aspect deals with the coordination of the satel-
lites themselves. From 1965 through the time of this writing, geostationary satel-
lites have been the foundation, and these have the benefit of being fixed in relation
to Earth and to one another. Consequently, coordination amounts to adequately
defining how a given satellite is to be used and what RFI it will cause and receive
with its neighbors when put into use (moved into the assigned orbit longitude and
turned on so it can relay signals from the ground).
These procedures are laid out in Article 9 of the ITU Radio Regulations and
conducted by the Radio Regulations Board of the ITU in Geneva, Switzerland. We
will not go into the details of these regulations nor about how nations and their
respective satellite operators interact on formal and informal bases. Suffice it to say
that the process favors established operators who have a presumption of the ability
to replace satellites as they reach end of life, but for which careful following of the
136 Interference
�������������������������������������������������������
Protection Ratio (C/I) and Its Application
rules is mandatory. A new operator will face challenges in finding a usable orbit
longitude (slot) and completing this process in time to launch the first satellite into
orbit. This is probably why potential new entrants are focusing their attention on
nongeostationary satellite orbit (NGSO) satellite systems or constellations operat-
ing in the low Earth orbit (LEO) and medium Earth orbit (MEO) regimes.
The classical example of an NGSO would be the Iridium system, implemented
by Motorola in the 1990s using 66 LEO satellites in a near-polar orbit at an alti-
tude of 780 km. To proceed with Iridium, Motorola required a dedicated piece of
spectrum allocated by the ITU and subsequently assigned to them on a sole-user
basis by the FCC. To do the latter, the FCC segmented the band to give them ap-
proximately 8 MHz and like amount to their main competitor, a joint venture of
Loral Space and Communications and Qualcomm. The MSS band in question is in
proximity to the same segment of L-band used by LightSquared and GPS; their re-
spective protection is again provided by having adequate frequency separation and
controls on power. Coordination with any terrestrial uses of this spectrum were
nearly unnecessary as few if an existed at the time of introduction. One major ex-
ception was for Radio Astronomy (RA) also using a proximate band but sensitive
to out-or-band emissions from space. A protocol was established wherein Iridium
would not exceed a certain specified flux density in areas populated by radio tele-
scopes like Greenbank, WV. Iridium satellite operation was automated in such a
way that the appropriate beam was throttled when a satellite was in view of this
area of the ground.
The various RFI paths that can exist between Earth stations, satellites and ter-
restrial microwave stations is illustrated in Figure 6.16. The paths labeled (2) and
(3) are addressed through the process of terrestrial coordination, discussed previ-
ously for Article 7 of the ITU Radio Regulations. Path (1), which is directly from
the terrestrial microwave station to the satellite in orbit, is unlikely and in any case
is addressed by the general rule that terrestrial microwave stations must not point
their antennas at the geostationary arc.
Section 5.3.1 provides a clear example of how RFI is introduced into an Earth
station. In this example, the RFI is from a ship-board radar operating in an adjacent
Figure 6.16 Interference between terrestrial microwave stations, and Earth station and a satellite.
6.5 Satellite Communications vs. Terrestrial Microwave Systems 137
frequency band. It produces front-end overload in the LNB and was cured using
an RF filter. The situation with respect to a shared frequency band between FS
and FSS, particularly C-band, demands more than filtering as the interference is
generally within the passband of either the earth station LNB or the terrestrial
microwave receiver.
RFI along Path (4), which is directly from the satellite into the terrestrial micro-
wave station, is controlled through a flux density limitation on satellite transmis-
sion to the earth. An example of such a limitation is presented in the set of formulas
in Table 6.1 [10].
This somewhat obscure specification limits the power flux density on Earth’s
surface, with the more severe limit toward the limb of the Earth (i.e., where a ter-
restrial microwave antenna might be pointed at a satellite near the local horizon).
The two reference bandwidths consider the carrier bandwidth, further restricting
the power spectral density. The variable Y in the second set of formulas is specified
in a footnote to the table of frequency allocations:
The value of Y is defined as Y = 0 for max(NN, NS) ≤ 2; Y = 5 log(max(NN,
NS)) for max(NN, NS) > 2, where NN is the maximum number of space stations in
a system simultaneously transmitting on a cofrequency basis in the fixed-satellite
service in the Northern Hemisphere, and NS is the maximum number of space sta-
tions in the same system simultaneously transmitting on a co-frequency basis in
the fixed-satellite service in the Southern Hemisphere. In determining NN and NS,
two space stations simultaneously transmitting during periods of short-duration
handover shall be considered as one satellite.
We now address the local problem of RFI between an Earth station and one
or more terrestrial microwave stations (i.e., paths (2) and (3), respectively). Figure
6.17 illustrates the problem in terms of a terrain profile where the earth station is
shielded to some extent by a hill or other obstacle. This is the general approach
since radiation is diminished by blockage and diffraction loss. Section 5.1 addresses
the theory and analysis of this aspect of RFI under an obstructed path.
While most of the INTELSAT C-band Earth stations were placed outside of
urban areas, new entrants saw benefits for having teleport facilities in proximity
to customers like cable TV networks and television producers. The teleport in Cul-
ver City, CA, was perhaps the first to construct and operate antennas right in the
middle of an urban center. They exploited the isolation provided by tall buildings
that blocked the path of offending microwave links, and placed their frequencies
to be noninterfering.
Figure 6.17 Propagation of RFI between an Earth station and terrestrial microwave stations that
share the same frequency band.
References
[1] International Telecommunication Union, Radio Regulations, Issue of 2014, ITU, Geneva,
Switzerland.
[2] Alsuwaidi, J., and S. Mahmoud, “Enhancing the Performance of W-CDMA Downlink in
Mobile Satellite Systems for Web Traffic,” Institute of Electrical and Electronics Engineers,
0-7803-7632-3/02/, 2002.
[3] Lee, J. S., and L. E. Miller, CDMA Systems Engineering Handbook, Norwood, MA: Artech
House, 1998.
[4] Sanders, G. A., J. E. Carroll, F. H. Sanders, and R.t L. Sole, “Effects of Radar Interference
on LTE (FDD) eNodeB and UE Receiver Performance in the 3.5- GHz Band,” NTIA Report
14-506, July 2014.
[5] National Telecommunication and Information Agency, “Preparation of Applications for
Frequency Assignment Action,” Chapter 9, 1/2008, http://www.ntia.doc.gov/legacy/osm-
home/redbook/9.pdf.
[6] Sanders, F. H., R. L. Sole, J. E. Carroll, G. S. Secrest, and T. Lynn Allmon, “Analysis and
Resolution of RF Interference to Radars Operating in the Band 2,700-2,900 MHz from
Broadband Communication Transmitters,” NTIA Technical Report TR-13-490, U.S. Dept.
of Commerce, Oct. 2012.
[7] Parsons, G. M., and R. Singh, An ATC Primer: The Future of Communications, Mobile
Satellite Ventures, 2006.
[8] National Space-Based Positioning, Navigation, and Timing Systems Engineering Forum
(NPEF), “Assessment of LightSquared Terrestrial Broadband System Effects on GPS Re-
ceivers and GPS-Dependent Applications,” June 2011.
6.5 Satellite Communications vs. Terrestrial Microwave Systems 139
Up until this point, we have worked to identify the various types of RFI that affect
radiocommunication systems and services. An understanding of the causes of this
issue is essential for proceeding with the key step of preventing and/or removing
RFI once it has occurred and had its impact. For this, we need to consider two
regimens:
1. RFI resolution: the process of quantifying the type and source of the RFI
and its effect on the planned or operating radiocommunication system.
2. RFI mitigation: the important step of identifying ways to remove or correct
for the presence of the RFI when and if it occurs. This includes remediation
of the unacceptable state of operation.
This is not unlike how a medical doctor first diagnoses a disease and deter-
mines the best treatment for its cure or reduction in the symptoms. This chapter is
dedicated to the first process and provides some proven steps and guidelines, but
the actual methodology for a given case will need to be worked out based on the
specifics. The remaining chapters address the step of RFI mitigation.
Spectrum analysis is a basic and vital aspect of RFI investigation and its resolu-
tion and so is worth some explanation. The measurement tool that performs this
function so admirably is called, simply, the spectrum analyzer. It is to spectrum
analysis what the oscilloscope is to time waveform analysis. According to Keysight
Technologies, the company originally known as HP and a strong proponent of this
device [1]:
141
142 RFI
�������������������������
Resolution Techniques
and the ability to measure complex signals using advanced DSP techniques. Addi-
tional benefits of using the FFT approach include
•• Fast capture of waveform: In view of the fact that the waveform is analyzed
digitally, the waveform can be captured in a relatively short time.
•• Ability to capture nonrepetitive events: The short capture time means that
the FFT analyzer can capture nonrepetitive waveforms, giving them a capa-
bility not possible with analog spectrum analyzers.
•• Ability to analyze signal phase: As part of the signal-capture process, data is
gained that can be processed to reveal the phase of signals (i.e., “i” and “q”
components).
•• Storage of waveforms: Using FFT technology, it is possible to capture the
waveform and analyze it later should this be required. This allows compari-
son over time and trending.
There are two domains where the spectrum analyzer has great benefit: (1) the
identification of unknown signals that are present within a bandwidth that con-
tains noise and existing signals (RFI investigation), and (2) detailed measurement
of signal characteristics to verify prediction based on mathematical analysis or lab-
oratory measurement. The former can be qualitative, while the latter is both quali-
tative and quantitative. Whenever quantitative measurements are needed, there is
the basic question of accuracy and precision of measurement. With improvements
in direct ADC at RF, spectrum analysis is has moved quickly in the FFT direction.
The applications of spectrum analysis are varied, testifying to the versatility of
this type of instrument. It provides a clear window into what is going on in a par-
ticular piece of radio spectrum, ranging from kilohertz up into the gigahertz region.
Basic measurement capabilities include carrier power, noise spectral density, band-
width, and spectral shape. Furthermore, a spectrum analyzer is capable of contrib-
uting to our understanding of underlying characteristics of a given communica-
tions signal. In this regard, the other common spectrum analyzer measurements
include modulation, intermodulation, and noise [3]. Measuring the quality of the
modulation is important for making sure a system is working properly and that
the information is being transmitted correctly. Understanding the spectral content
is important, especially in communications with limited bandwidth. The amount
of power being transmitted (e.g., to overcome the channel impairments in wire-
less systems) is another key measurement in communications. Tests such as modu-
lation degree, sideband amplitude, modulation quality, and occupied bandwidth
are examples of common modulation measurements. In communications, measur-
ing distortion is critical for both the receiver and transmitter. Common distortion
measurements include intermodulation, harmonics, and spurious emissions; any of
these can produce RFI internally and externally as well. Tests such as noise figure
and signal-to-noise ratio (SNR) are important for characterizing the performance
of a device and/or its contribution to overall system noise. For all of these spectrum
analyzer measurements, it is important to understand the operation of the spec-
trum analyzer (discussed below) and the spectrum analyzer performance required
for a specific measurement to help choose the right analyzer for the application as
well as get the most out of it.
144 RFI
�������������������������
Resolution Techniques
Figure 7.3 (a) General arrangement of controls and display on a commercial spectrum analyzer.
(Courtesy of Keysight Technologies, Inc.) (b) A black box spectrum analyzer (on the right) along
with the laptop computer used to display and control all analyzer functions. (Courtesy of Tektronix.)
7.1 Spectrum Analysis and Monitoring 145
•• Frequency range: This is the operating range of the analyzer, ending at some
maximum frequency in the GHz range. Keep in mind that the frequency of
observation may need to be a multiple of the expected carrier so that inter-
modulation products and harmonics can be observed as well.
•• Accuracy, frequency and amplitude: These are dictated by the respective sen-
sors within the analyzer, typically ±1 dB.
•• Resolution: Frequency resolution is the ability of a spectrum analyzer to
separate two input sinusoids into distinct displays. While simple in concept,
realization of this capability requires narrow filtering in the analyzer IF. On
the other hand, a narrow filter has a slow response time, which means that
the sweep rate must be low enough to permit resolution of the sine waves.
Fast sweep is preferred unless the display has memory or persistence.
•• Sensitivity: Range of input signal, minimum to maximum, typically in dBm.
Consider also the internal noise floor. The low end of the range is needed
when measuring into deep signal fades or nulls.
•• Distortion: Produced when signals exceed the linear range of spectrum ana-
lyzer input. The RF input attenuator is a step attenuator located between
the input connector and the first mixer. This is used to adjust the level of the
signal incident upon the first mixer to prevent mixer gain compression and
distortion due to high-level and/or broadband signals. The resulting range
may need to be as much as 100 dB.
•• Dynamic range: Dynamic range is defined as the maximum ratio of two signal
levels simultaneously present at the input that can be measured to a speci-
fied accuracy. This interacts with distortion and sensitivity. This is important
when observing weak signals, such as RFI, present along with strong carriers.
A dynamic range of 60 to 100 dB may be needed for particular measurements.
shown in Figure 7.4, the IF follows the swept oscillator, producing what is called
the resolution bandwidth (RBW). Specifically, the IF filter defines the RBW of the
analyzer and can be changed via the front panel of the analyzer (or menu setting of
the FFT processor). A broad range of variable resolution bandwidth settings can
be used to optimize sweep for various signal conditions, yielding a tradeoff among
frequency selectivity (the ability to resolve signals), SNR, and measurement speed
(sweep rate).
We can see in Figure 7.4 that as RBW is narrowed, selectivity is improved (i.e.,
we are able to resolve the two input signals). This can also often improve SNR. The
sweep speed and trace update rate, however, will degrade with narrower RBWs.
The optimal RBW setting depends heavily on the characteristics of the signals of
interest and is best determined by experimentation using the various controls. For a
signal with defined sidebands, an excessively wide value of RB will cause the power
of these sidebands to be lumped in with the overall carrier and thereby become
invisible in the broadened display. Narrowing the RB allows the sidebands and
other spectral details to become clearly visible. The sweep speed will need to be de-
creased, however, to permit spectral energy to pass through the filter or, in the case
of an FFT analyzer, to be computed quickly enough to display the spectral details.
The video filter is a low-pass filter that is located after the envelope detector
and before display processing. This filter determines the bandwidth of the video
amplifier and is used to average or smooth the trace seen on the screen. These ele-
ments are illustrated in Figure 7.5.
Analyzers with DSP have capabilities beyond what the RBW and video fil-
ters are capable of doing. For example, the data after capture can be processed to
produce various types of averages. A moving average, for example, smooths the
Figure 7.4 The IF section of the analyzer provides the RBW, which permits finer definition of spec-
tral properties. (Courtesy of Keysight Technologies, Inc.)
7.2 Spectrum Monitoring and Carrier Measurement 147
Figure 7.5 The circuit shows the sweep and downconversion within the spectrum analyzer, pro-
ducing the baseband for envelope detection and video filtering. The scan on the left shows signal
in noise and without video filtering; the benefit of video filtering is clear from the scan on the right.
(Courtesy of Keysight Technologies, Inc.)
spectral shape by removing short-term fluctuations that are due to various analyzer
artifacts and internal and external noise.
The spectrum analyzer displays signal plus noise so that the closer a signal is to
the noise level, the more the noise makes the signal difficult to see. When signal and
noise are equal in power in the same RB, the signal appears to be 3 dB above the
spectrum analyzer noise floor. By changing the video bandwidth (VBW) setting, we
can decrease the peak-to-peak variations of noise. This type of display smoothing
can be used to help find signals that otherwise might be obscured in the noise. To
do this, we need to first narrow down the scan range and sweep rate; this means
that the particular signal must be isolated so that it can be analyzed in this manner.
Another postdetection feature displays a history of past traces, called peak hold. It
is provided by a frame store of past traces to capture intermittent or highly variable
signals such as RFI. Modern DSP analyzers offer many other features such as the
waterfall display, helpful when observing intermitted signals like radar RFI.
the latter depending on software and databases to collect, process and display the
spectral data.
According to infoDev [4], the purpose of spectrum monitoring should be to
achieve
•• Monitors and captures all types of signals (as expected in the application):
• Continuous, transients, high level, low level, adjacent channel;
•• Uses commercially available hardware and computer operating system;
•• Modular design to permit update/add of subsystem components
independently;
•• Intuitive GUI operation;
•• Complementary interworking among applications, using a consistent data-
base of previous measurements;
Figure 7.6 General arrangement of elements of a spectrum monitoring system. The types of anten-
nas and frequency range will depend on the application.
7.2 Spectrum Monitoring and Carrier Measurement 149
Figure 7.7 The SMR system employs a WAN, which is essentially the Internet, to connect remote sites with
antennas and spectrum analyzers [5]. (Courtesy of Crystal Solutions, LLC.)
•• Interactive spectrum analyzer control: one or more users may view, monitor
and control local or remote analyzers. Users working with a common ana-
lyzer can simultaneously view different measurements on different segments
of a spectrum or different switch ports.
•• Background monitoring: measures transponders/carriers behind the scenes.
Monitoring continues even when operators may not actively view the spec-
trum thus continuing to detect any anomalous carrier condition or RFI, and
log relevant carrier measurements for later analysis.
•• Historical trending and spectral playback: users can graph and view histori-
cal measurements for a transponder/carrier. Users can also play back the
recorded spectrums to view the spectrum as it changes. Playback includes
leap-to-anomaly feature to improve RFI investigation.
•• Spectrum recording: records the spectral trace for transponders/carriers full
time or “On Anomaly” (i.e., recording automatically launches when an
alarm condition occurs on a monitored spectrum segment).
Figure 7.8 SMR software uses one or more spectrum analyzers to monitor hundreds of RF sources.
When integrated with Crystal’s NMS+ solution, SMR can react to detected anomalies and automati-
cally route affected signals to digital signal processors for further investigation or reconfigure the
signal path to an alternative path (either on the same site or to an alternative site) to restore service
integrity. (Courtesy of Crystal Solutions, LLC.)
Figure 7.9 RDF using a directional antenna and a spectrum analyzer. (Courtesy of Rhode and
Schwartz.)
Figure 7.10 Use of RDF and triangulation to locate a source on a topographical map. The dotted
triangles (angle uncertainty) and shaded polygon (position uncertainty) provide an initial indication
of location.
that there is one tester who must relocate from measurement position 1 to 2, a time
period during which the RFI must remain on. Having the simultaneous measure-
ments will, of course, eliminate this delay. Location of RFI based on RDF amounts
to RF snooping in the general vicinity of the source. Sometimes, propagation af-
fords us the opportunity to extend the range to hundreds or even thousands of
kilometers. Yet the accuracy could suffer as a result and ultimately a close in RDF
will be needed. Use of aircraft or satellites as points of reception is also a possibility,
as discussed later in this section.
7.3 Interference Location and Radio Direction Finding 153
Figure 7.11 Basic arrangement for the use of TDOA using two satellites and two receiving anten-
nas [6].
slightly different frequencies since they are not tied to the same source. Also, better
location accuracy is provided by having accurate satellite orbit data, called ephem-
eris, to perform the calculations. Several system suppliers compete; some offer RFI
location services using extensive database of satellites and Earth stations.
SAT Corporation, a wholly owned subsidiary of Kratos Corporation, provides
a system called Sat ID, which is their TDOA/FDOA system that can provide an
uncertainty ellipse of only 1 km in diameter as opposed to 300 km under the con-
dition of poor ephemeris data. These systems are owned by satellite operators and
governments, and SAT Corporation also operates its own system to provide RFI
location as a service.
7.4 Transmitter ID
We rely on the spectrum analyzer and spectrum analysis to identify and categorize
RFI signals as we receive them through the victim system. Also, ancillary antennas
can be used for this purpose as long as they are able to detect enough of the of-
fending signal. In terrestrial systems such as cellular and microwave line-of-sight
radiocommunication, the distances involved are relatively short. A path including a
satellite can extend the range many thousands of kilometers, however. As a result,
the potential population for RFI sources can be extremely large and far removed.
For this reason, there have been various efforts to create standards for inclusion of
an identifying signal in the transmitter (transmitter ID), allowing a receiver the po-
tential to determine the source and follow up as appropriate. With a closed system
using proprietary equipment, it is usually a simple matter to provide identification
in the basic transmission or protocol structure. Open systems in a multivendor
environment, however, can pose a challenge unless there are rigorous controls or
standards that can be applied. This is the case in push-to-talk services used by pub-
lic service and the military, because all transmissions must be identified by the user.
A technique for introducing the ID data onto a standard TV transmission was
first introduced into the satellite broadcasting industry around 1988, again as a
7.4 Transmitter ID 155
result of the Captain Midnight incident. As proposed by Home Box Office (HBO),
a baseband subcarrier with Morse code was added to the normal video and audio
modulation. This carried the industry well until broadcast went digital with digital
satellite TV and digital cable TV. There are currently two forms of carrier ID in
use in the commercial satellite industry: those that are embedded in MPEG Stream,
or in the Network Information Table (NIT) field, and a subcarrier underneath the
main carrier (spread spectrum). The NIT conveys information relating to the physi-
cal organization of the multiplexes carried via a given network, and the character-
istics of the network itself [7]. This approach works well on broadcast (DVB) video
and data carriers that contain the MPEG stream, whereas the subcarrier technique
can be applied more generally, but not on burst-mode TDMA signals from remote
earth stations (VSATs). The latter is still unresolved yet viewed as a significant
source of RFI into adjacent satellites.
Carrier ID was implemented for the 2014 Olympics using the NIT. ISI has
a system for reading the NIT and operators will have their capability. Comtech
EF Data introduced an alternative and more generally applicable system using a
spread-spectrum subcarrier hidden under the main carrier, indicated in Figure 7.13.
Here is another caveat about applying transmitter ID to satellite signals. RFI
on the satellite path can be introduced from a separate terrestrial source by leakage
Figure 7.13 Carrier ID provided using a spread-spectrum signal placed underneath the carrier to be identi-
fied. (Courtesy Comtech EF Data.)
156 RFI
�������������������������
Resolution Techniques
into or from IF cables at an uplink Earth station. In the past, it was mostly from
local FM stations around 88 MHz getting inside the 70-MHz IF found in older
equipment. Now, with Earth stations using L-band IF systems (950 to 2,150 MHz),
signals from cell towers can leak into the uplink side of VSATs. GSM 2G signals
are being retransmitted around 950 MHz, which is at the bottom of the range.
3G signals come higher in the United States. Also, 4G can potentially leak in at
1,900 MHz. The Block Upconverter (BUC) that transmits at the VSAT can also be
a source of leakage. Société Européene des Satéllites (SES) uses older Nokia GSM
cell phones in the engineering mode, which allows a unit to identify the number of
the cell tower that it is receiving. The cell tower in question is probably in another
country leaking into the IF transmit (L-band) of a VSAT uplink. SES uses that
phone in their teleport and in the field to read out the identification of the offend-
ing cell tower.
But this must be balanced with the need to keep what is working in order. In the
past, the dynamics made for interesting reading and even spectacle. For example,
the prospective broadband satellite operator of the late 1990s, Teledesic, was able
to corral literally all delegations at a key ITU World Radiocommunication Confer-
ence and thereby obtained their own sole-user block of spectrum at Ka-band. The
spectrum in question was carved out of the existing FSS spectrum that had rela-
tively light usage. This trade would allow Teledesic to proceed without much risk
of causing or receiving RFI. Alas, their efforts to build a system fell apart because
of unacceptable cost of development, and the spectrum remained unused for more
than a decade. Now, in 2015, a number of new entrants intend to follow Teledesic
with their respective technical strategies, while the GEO community has discovered
the benefits of Ka-band for their own broadband and broadcast services. Hav-
ing multiple LEO operators in the Teledesic spectrum will present an interesting
challenge, and the planners will work out the details based on the past record and
whatever technical data can be produced ahead of initial system operation. Strate-
gies for spectrum allocation in such instances are covered in Section 8.1.3.
Similarly, we saw that NTIA has studied the possibility of sharing C-band spec-
trum between 4G LTE operators and the currently existing maritime radar com-
munity. The study presented in Section 6.4.1 gave only a small part of the picture
of what could happen when these systems collide in space and time. Sharing among
very different systems is not so easily done, and the rules of the road can impose
harsh limitations. An incumbent like the U.S. Navy might not appreciate having a
new source of RFI appear on their screens, so to speak. And the wireless operators
might not find the value they are looking for in spectrum that, to them, is polluted
with powerful and intermittent radar signals. The discussion of cognitive radio in
Section 9.2, however, along with good spectrum monitoring and carrier ID, offers
a path forward for intelligent spectrum sharing.
Figure 7.14 General geometry and signal structure involved in the RFI investigation of FM/TV into
digital SCPC QPSK satellite transmission.
kbps corresponding to the DS0 channel, while the FT/TV transmission contains an
energy dispersal (ED) frequency sweep of 2-MHz bandwidth to reduce the radiated
power spectral density (PSD). The rate of sweep is only 15 kHz corresponding to
the line sweep rate of analog TV. As a result, the FM/TV signal appears as a quasi-
static unmodulated carrier when it is within the bandwidth of the SCPC receiver/
demodulator. With adjacent satellite isolation of approximately 30 dB and a dif-
ference in carrier power in the range of 10 to 20 dB (the TV carrier is at maximum
power and the SCPC carrier is backed off), the carrier C/I is in the range of 20 dB
to 10 dB. But this is only present when the FM carrier passes through the victim
SCPC receiver bandwidth.
Let us assume that this increases the BER from 10–8 to approximately 10–5.
The interfering carrier sweeps back and forth across the bandwidth shown with
a constant power. Thus, RFI exists during the fraction of time (e.g., duty cycle):
RS
Bi
where Rs is the symbol rate. For example, if we assume QPSK then Rs = 32 kHz and
the ratio above is 32/2,000 or 0.016. Simply stated, the BER is unaffected during
98.4% of the period and experiences RFI at the assumed protection ratio, C/I, only
during 1.6% of the time. Taking the worst case C/I of 10 dB and the corresponding
BER of 10–5, we can estimate the average BER by the following relationship:
( ) ( )
BER = 10-8 × 0.984 + 10-5 × 0.016 = 1.7 × 10-7
7.5 Additional Resources and Strategies 159
The first term in the equation counts bit errors when the interfering carrier is
out of band and the second term counts bit errors during the actual interference
when the sweeping carrier is within the bandwidth of the SCPC signal. We see
that the average BER is more in line with the out-of-band condition. Remediation
would consist of increasing carrier power and hence C/N in the absence of RFI by
some amount, perhaps 1 dB.
The analysis and its result, above, were taken by the other party as supposition
of a theory that needs verification. Verification was performed at the time using an
off-the-shelf digital SCPC modem and FM/TV modulator, confirming the analysis.
An expanded test program, subsequently conducted and reported by Eutelsat
[8], involved an updated set of link parameters, reflecting the improvements in
satellite operation over the intervening period of ten years. Specifically, the early
studies summarized above had been operated under relatively high carrier-to-ther-
mal noise ratio (15 to 16 dB) in the SCPC channel bandwidth (i.e., 32-kHz noise
equivalent bandwidth) to achieve a BER of less than 10-6 in the absence of inter-
ference. Data rates available in the then-existing Eutelsat Satellite Multiservices
System (SMS) in the range 64 kbps to 1,920 kbps needed to be evaluated under
this RFI environment. Furthermore, the system uses rate ½ convolutional FEC en-
coding with eight-level soft-decision Viterbi decoding as a means of improving the
tradeoff between space segment bandwidth and power. As a result of this encoding,
the SCPC carriers in the SMS operated with much lower carrier-to-thermal noise
ratio (by approximately 6 dB) than those considered in the previous studies. The
changed conditions could invalidate the conclusions from ten years prior.
The interference from a co-channel FM television signal harming an encoded
SCPC channel raises several questions:
How will the decoder react to a potential burst of errors created by a high-
energy component falling within in the channel bandwidth?
What C/I values can an SCPC channel tolerate when encoded?
What is the dependence on C/I for SCPC channels having different bandwidths?
The laboratory tests were performed with the configuration shown in Figure
7.15 employing commercially available equipment to evaluate the interference
caused by the same FM/TV carrier to the encoded SCPC/QPSK channel.
Figure 7.15 FM/TV interference into QPSK/SPCP test configuration employed by Eutelsat [8].
160 RFI
�������������������������
Resolution Techniques
Table 7.1 C/I Required to Limit Eb/N0 Degradation on Victim QPSK Link
Caused by FM/TV to Either 0.5 dB or 1.0 dB
2 Mbps Modem 64 kbps Modem
Peak-to-Peak
ED Deviation Sync C/I required C/I required C/I required C/I required
(MHz) Pulses for 0.5 dB for 1.0 dB for 0.5 dB for 1.0 dB
CW No 14.4 11.1 13.8 10.9
1 No 13.8 10.9 8.8 6.5
1 Yes 13.8 10.4 7.8 6.0
2 No 13.5 10.3 8.3 6.3
2 Yes 13.3 10.2 7.1 5.5
4 No 11.8 9.1 7.0 4.9
4 Yes 11.7 9.1 6.4 4.3
7.5 Additional Resources and Strategies 161
7.5.3 Remediation
Remediation is defined as simply the correction of something that is bad or defec-
tive. We have to live with some level of RFI, especially because it is usually due to
the normal operation of internal and external transmitting sources of RF energy.
But remediation has to do with our response to RFI that causes unacceptable deg-
radation to our service. As we have discussed previously, the first task after the RFI
is identified is to establish if it is internal to our own system or network, the case
perhaps 50% or more of the time. We can then work to improve our internal opera-
tion and either avoid or remove the cause (or make the link less susceptible). If it is
an external RFI source, then we need some kind of cooperative approach; barring
that, we need to find ways to operate in the presence of this external RFI.
All of these approaches can work. The choice will depend on the time, cost,
and resulting performance. The next sections address the nature of our corrective
action: is it a short-term “stopgap” solution, or is it a long-term type of perma-
nent resolution of the problem (until something changes that introduces a new RFI
source and effect).
with the standard formula for sidelobes and a 9m antenna. There were temporary
operators, however, who used transportable antennas of the 6m class (reducing
the C/I by 4 dB or more) to provide uplinks at venues like stadiums and news
scenes. One night, we received a complaint from another operator about RFI to
one of their broadcast channels, and it was determined that it was in fact a tempo-
rary transmission on our satellite from one of these transportable stations. As the
technical manager, I placed a call to my counterpart at the other satellite operator,
which was our primary competitor, to discuss the situation going forward. In a
friendly way, he informed me that the FCC had classified these smaller antennas
as nonstandard and that if they were to cause interference with others using stan-
dard antennas then they must shut down. After thanking him, I contacted our
legal counsel in Washington, D.C., who followed up and confirmed the FCC order.
From that point, we issued instructions to our network operation center that these
smaller antennas were only allowed to transmit on frequencies that would not
cause adjacent satellite interference. The total time from incident to final resolution
was less than a week.
Here we see that an incident of RFI was really the precursor for how our sys-
tem must be operated in the new environment. Other situations arise where the RFI
is indeed a single case requiring only a limited corrective action. A case in point was
out-of-band radar interference that overloaded the front end of an Earth station
operating in an adjacent band, discussed in Section 5.3.1. The simple and effec-
tive solution was to insert a bandpass filter ahead of the low-noise amplifier in the
Earth station antenna. The same kind of solution removed RFI in a C-band receiver
due to RFI from radar on train engines and to automobile radar detectors. Filters of
that type are readily available and one wonders why they are not standard design
practice. But, we often have to relearn well-proven solutions.
An example of unexpected RFI to a GSM cellular operation was reported by
COMSEACH, a unit of CommScope Company [9]. A U.S. cellular operator was
experiencing a high degree of dropped calls in one sector of a base station. When
COMSEARCH investigated, however, the problem did not immediately present
itself. Eventually, it was found to be due to private and improper use of vehicular
cellphone amplifiers in the 800-MHz band capable of causing in-band RFI to the
GSM network. This device couples to a cellphone using a small antenna that is
connected to the amplifier and an external vehicular antenna to reach the base sta-
tion. If the cellphone is too close to this antenna, the amplifier overloads and emits
spurious signals that are as much as 30 dB higher that what is required. It was no
wonder that when a vehicle functions in this manner that the operation of the cell
site is disrupted. Any solution to this kind of issue will require controls placed on
customers who might employ such nonstandard and RFI-prone equipment. One
idea might be to detect if a subscriber is using one of these amplifiers and then to
move them to an open segment of the band. More strategies are suggested in Sec-
tion 8.1.1, which particularly involve changes in frequency assignment.
appearance of RFI in the future. New systems like 4G wireless and advanced broad-
band satellite networks include many dynamic features with inherent ability to miti-
gate RFI as it appears. With regard to satellites, the ITU regulatory process has been
particularly effective in promoting innovation [10]. But, even these will need human
intervention and design work as performance degrades beyond what was initially
recognized. The long term involves a thorough understanding of all of the principles
discussed previously in the volume, as well as the details of the characteristics of the
potentially victim system. The examples covered in this and the previous chapter
suggest the range of what can be involved. The extent of investigation and testing
can grow well beyond what one initially thinks is adequate. Most RFI involves
other parties and players so it is partly engineering and partly conflict resolution.
The latter is best addressed in friendly collaboration or negotiation. If it becomes
an unfriendly type of conflict, there may be no currently available resolution with-
out one party vacating the spectrum or drastically changing how they operate. The
good news is that RFI is a two-sided issue: the party causing RFI is probably trying
to conduct its own radiocommunication and wishes to enjoy its rights to spectrum
with minimal impairment. Methodologies covered in the remaining chapter offer
state-of-the-art schemes for adaptive systems that address these challenges.
References
[1] Keysight Technologies, Spectrum Analysis Basics, Application Note 150, February 25,
2014, http://cp.literature.agilent.com/litweb/pdf/5952-0292.pdf.
[2] Poole, I., “FFT Analyzer Tutorial,” Radio–Electronics.com, http://www.radio-electronics.
com/info/t_and_m/spectrum_analyser/fft-analyzer.php.
[3] Anritsu, “The Basis of Spectrum Analyzers,” http://www.naic.edu/~phil/hardware/Misc/
anritsu/SpectrumAnalyzer_basis_of.pdf, 2008.
[4] Aerosystems International Inc. © 2015, 3538 Ashby, Montreal, Quebec, Canada H4R
2C1.
[5] Crystal Solutions, LLC., “Case Study—Spectrum Monitoring and Recording in the Real
World,” http://crystalcc.com/docs/case-studies/Spectrum%20Monitoring%20and%20Re-
cording%20in%20the%20Real%20World%20Case%20Study.pdf.
[6] Steffes, P., “Research and Development of Satellite Interference Location System (SILS) at
Georgia Tech,” January 20, 1987, through June 30, 1990, https://smartech.gatech.edu/
bitstream/handle/1853/35196/e-21-610_338506_fr.pdf.
[7] European Telecommunications Standards Institute, “Digital Video Broadcasting (DVB);
Specification for Service Information (SI) in DVB systems”, ETSI EN 300 468 V1.11.1
(2010-04).
[8] Dutronc, J, et al., “FM/TV Interference into Rate ½ FEC Encoded SCPC/SPSK,” Interna-
tional Journal of Satellite Communications, Vol. 4, pp. 111–126, 1986, p 111.
[9] Branin, J., “Interference from Cellular Amplifiers—A Case Study,” COMSEARCH, Ash-
burn, VA, 2011, http://www.comsearch.com/articles/CellAmpInterf.pdf.
[10] Ebadi, A. R., Communications Satellite Fundamentals—From Design to Launch and Op-
eration, Oxford, UK: Oxford University Press, 2013, p.49.
CHAPTER 8
This chapter is about the tactics of addressing RFI before it rears its ugly head or
after it causes trouble and/or harm to an operating radiocommunication system.
As a critical type of troubleshooting in telecommunications, RFI is a menace that
appears when least expected and often with dire results. There is a host of activity
we can do in response to an RFI incident, whether short term (nice, because it even-
tually disappears), long term (troublesome, but addressable through a change in
technical design or operation), or intermittent (often nasty and usually the hardest
to deal with due to the “cat and mouse” nature of getting to the problem). First we
consider the mind-set of the RFI planner/avoider as well as the troubleshooter who
must take the most appropriate corrective steps. In anticipation of RFI in a system,
we address planning steps and interference-reduction strategies for the long term.
The impact will be lessened if RFI has a localized impact and the frequency change
does not affect the overall system.
Frequency planning consists of making assignments that avoid RFI in the first
place and allow it to be mitigated when it does occur. We first consider making a
frequency change as a simple but effective means of sidestepping RFI. A pool of
available frequencies is perhaps the best means of being ready for an RFI incident.
Employing this reserve to dodge RFI, however, means that there must be a means of
issuing and installing the new frequencies in a rapid and reliable manner.
165
166 How
�����������������������������������������������������
to Identify, Prevent, and Fix Common RFI problems
transmit. This principle is often applied to automated systems such as trunk radio
and some satellite networks.
This short case study represents the real world, and we can find additional
examples in a more modern era. Spectrum that is supposed to be clear can end up
having RFI within it, either initially or in the future. Excess channel capacity is
a means to assure ourselves that the service can continue as we conduct our RFI
investigation and remediation activities. The difficult part is having the means to
distribute the new frequencies to the radio terminals that need them and then verify
that all is working correctly. An example in satellite communications involved an
SCPC satellite network using a central hub station and remote VSATs operated by
independent subscribers who required Internet access. A portion of the satellite
transponder experienced RFI that was being investigated by the satellite operator;
however, service had to be maintained until the RFI could be removed. In the mean-
time, we used the network management system to direct the affected terminals to
a different segment of the band. This is a simple solution that provided “breathing
room” until the satellite operator had located and removed the source of RFI.
The NTIA test of 4G LTE operation in the presence of radar, described in
Section 6.4.1, highlighted how modern networks automatically correct for the
presence of high traffic loading and RFI. Through the process of assigning RBs,
8.1 Frequency Planning Techniques 167
the base station commands the remote user equipment to employ a particular fre-
quency or block of frequencies. These are taken from a pool of resource and that
pool can be adjusted to reflect an understanding of the local RFI environment. Ad
hoc networks such as Wi-Fi perform this function automatically as well, being able
to change modulation and bandwidth dynamically in response to increases in the
combination of noise and interference. Furthermore, Fluke Networks, provider of
electronic measurement systems, offers a Wi-Fi monitor that can identify and locate
RFI affecting Wi-Fi network performance. We can watch for the presence of RFI
through monitoring the behavior of such systems.
not acknowledge their responsibility. A good system of event logging (e.g., trouble
ticketing) and spectrum recording provides evidence for RFI forensic investigation.
Once identified, the removal would involve one of the following procedures:
Additional steps and techniques are covered in the following paragraphs. The
body of knowledge and experience is great, but there is perhaps no limit to the
steps that can be taken. As this must be tailored to the specific situation, the inci-
dent management approach covered in Section 8.3 should be considered in cases of
RFI that seriously hamper or endanger an operating radiocommunication system.
A source of RFI that is a licensed transmitter will usually self-identify through
the modulation that it contains. Standardized systems like 3G and 4G wireless base
stations and user equipment, digital TV (MPEG and DVB) transmissions over the
air and satellites, common-carrier microwave links, and many HF and VHF radio
transmitters all incorporate identification along with other internal network con-
trol data. Network operators in these domains all have provision to receive and
interpret such data, which is a good reason why they are the first line of defense
when RFI appears. Barring transmitter ID, we are left with the various RDF tech-
niques discussed in Section 7.3.
RFI response is often difficult and time consuming. Again according to Hol-
mes, “Genius is an infinite capacity for taking pains.” You must “dig in” and work
it hard to identify the potential causes, narrow down to the one that needs to be
addressed, and finding a solution – specifically short term and/or long term – that
allows the services to continue or be restored swiftly.
Figure 8.1 Three different allocation strategies to manage RFI in the operation of different user
communities.
Figure 8.2 Different user groups operating in same region with three cells and employing spectrum
interleaving (strategy B) to control RFI.
8.2 Avoiding Intermodulation 171
Intermodulation in this context refers to signals that result from the nonlinear in-
teraction of multiple signals on different frequencies within RF equipment or a
172 How
�����������������������������������������������������
to Identify, Prevent, and Fix Common RFI problems
Where M and N are integers (1, 2, 3, …), ftx is a transmit frequency, and flo is
another frequency from a local oscillator or other proximate source of RF. The new
frequencies appear across a wide spectrum and their respective levels are highly
nonlinear. Generally speaking, the designer or operator of a facility seeks to avoid
those frequencies that fall in band or close to in band of receivers. This is illustrated
in Figure 8.3 for an equipment chain that could represent a more complex facility
with unintended nonlinear elements within transmission lines and antennas (i.e.,
passive intermodulation). The first oscillator is multiplied to produce a desired RF
frequency, filtered and applied to another nonlinearity. Within the latter, the input
frequencies interact and produce a broad spectrum of intermodulation products.
Filtering can only reduce the quantity of spurious (spur) carriers; it cannot remove
those that fall close to or within the bandwidth of the desired signal.
Equations of the above form are programmed into software tools and graphi-
cal charts that guide engineers so that intermodulation products remain outside
of the band of interest. An example of such a chart is provided in Figure 8.4,
produced using the Keysight Genesis analysis software tool that evaluates “what
if” scenarios by calculating all frequency combinations. Operation in the solid bar
channels must be avoided as they contain the strongest spurious frequencies. Other
regions are deemed “spur free,” meaning that there is no risk of interference. This
type of presentation has also been referred to as a “spotty dog chart” in recognition
Figure 8.3 Illustration of intermodulation creation within an equipment chain, where the input
spectrum consists of a strong carrier with weak sideband frequencies and the output after the xN
multiplier contains many potentially problematic frequencies.
8.3 Interference Cancellation 173
Figure 8.4 An example of an intermodulation avoidance chart using Genesis analysis software
(“spotty dog chart”) indicating unusable and preferred channels. Courtesy of Keysight Technologies.
Figure 8.5 RFI is introduced within the receive antenna through the main beam or sidelobes. RFI
cancellation techniques are possible at any level in the reception process, from within the antenna
and feed (RF), the intermediate frequency where the signals are still modulated; after demodulation
or detection, cancellation involves detailed knowledge of data characteristics and/or some form of
access control [1].
signals received from the primary antenna. This is basically the same principle of
noise cancelling headphones. Cancellation can also be performed after the antenna
with this interference sample in the IF, baseband, and even after the data itself are
recovered. The latter generally only applies where the detailed characteristics of
interference modulation are known to some degree [1]. Figure 8.5 illustrates a de-
sired signal, C, and two potential interferers, I1 and I2. The various stages are areas
where cancellation could potentially be introduced, to be discussed in the following
paragraphs. Cancellation of more than one interferer will require greater complex-
ity such as more antenna elements (e.g., a phased array) or digital processing in the
subsequent elements of the receive chain. Furthermore, cancellation after demodu-
lation and detection requires knowledge of the interfering signal. This presumes a
somewhat closed system with controlled access.
Figure 8.6 Example of sidelobe RFI cancellation using an auxiliary antenna to receive the same RFI
signal along with processing that provides a close replica of the RFI for cancellation.
carrier power, C. The technique in Figure 8.6 derives from a radar receiver [2 ] but
has more general application in radiocommunication.
The complexity of the RF processor and modulation circuit (shown as a circle
with an X) will depend on the nature of the signals and environment. If the signals
are stationary in time, all that is required is that the amplitude and phase of the
interfering carrier be adjusted to match the interference as it comes through the
main antenna. This produces subtraction in the circle shown prior to the receiver.
There is a commercial satellite communications technique that provides for can-
cellation of a carrier using an adaptive approach. Termed “carrier-in-carrier,” the
concept is to allow two Earth stations to transmit to each other to provide a duplex
point-to-point link. The carriers are transferred over a bent-piper repeater within
the same area-coverage satellite beam. Normally, the carriers must be on different
frequencies so that reception of the opposite end’s carrier is not jammed by one’s
own carrier. In carrier-in-carrier, both stations transmit on the same frequency but
use cancellation (at IF) to remove the RFI. This is possible because each end knows
what its own signal is at any instant and can thus cancel it out on receive. The mo-
dem on each end is presented with the received carrier with a sufficient C/N, with
only residual RFI, for the process of demodulation.
A phased-array antenna system offers the opportunity to include RFI cancel-
lation within the requirements and capability of the original design [6]. Even the
two-element array in Figure 8.7 has the ability to cancel one RFI source. The same
signal is received by both antennas and adds constructively for the desired carrier.
Cancellation occurs at the off-angle toward the interference (see Figure 2.28) [3].
For the two-element example in Figure 8.7, the amplitude and phase adjust-
ment “weighting factors” for each of the two elements can be expressed in complex
form by
Element 1: w1 + jw2
Element 2: w3 + jw4
176 How
�����������������������������������������������������
to Identify, Prevent, and Fix Common RFI problems
Figure 8.7 Interference cancellation using a two-element phased array. Pattern shape is illustrative
and not analytically produced.
The weightings, contained within the boxes labeled “Amplitude and phase ad-
justment,” must satisfy the following so that the array accepts the desired signal
while simultaneously rejecting the interferer:
1 1 1 1
w1 = , w 2 = - , w3 = , w 4 =
2 2 2 2
P0
( ω) = 1 +
{ }
1 + cos τ ( ω - ω0 ) PJ
Pn PN
8.3 Interference Cancellation 177
Where:
PJ
= the ratio of interfering “jammer” power spectral density to the internal
PN
noise power spectral density of the receiving system,
d
τ = sin θ
v
RF cancellation can be applied to almost any signal and RFI type because it ad-
dresses the modulated carrier in free space. Also, it can be introduced into the sys-
tem after it is already in operation provided that suitable antennas can be installed.
The approach of beam-nulling has been applied for almost as long as radiocom-
munication has existed, and improvements have employed adaptive technology to
address dynamic RFI conditions.
8.3.2 Cancellation at IF
Cancellation of IF signals proceeds in much the same way as RF since the carriers
contain their original modulation (i.e., predetection cancellation). The dominant
approach today is perform A/D conversion on the entire IF band and rely on com-
plex digital processing to select the desired carrier and accomplish all processing
including detection and decoding. This approach also allows the greatest complex-
ity in terms of matching the relative gain and phase of signal and RFI. It would
be possible to attempt cancellation based on expected characteristics of the RFI,
something that can be updated through external means of measurement. There
could be some difficulty achieving the accuracy needed for good cancellation be-
cause of gain and phase variation in the analog devices in the frequency translation
and amplification stages between the RF section and the processor. A phased array
that accomplishes phase and amplitude combining at IF and after A/D conversion
178 How
�����������������������������������������������������
to Identify, Prevent, and Fix Common RFI problems
offers the means to enhance signal and cancel noise as well. A system such as this
can include adaptive channel characterization to improve suppression of channel
noise and impairments.
•• Base station to user terminal, called the forward or outbound link (the uplink
in cellular networks) and
•• User terminal to base station, called the return or inbound link (the downlink
in cellular networks).
There is a third type of called mesh or point-to-point link that avoids connec-
tion to a hub station. The use of the terms uplink and downlink to describe the
forward and return links, respectively, is an item of confusion for those in satel-
lite communications since the hub Earth station and the user VSAT terminal each
have respective uplinks and downlinks. So, we will use the more general terms of
forward (hub transmit) and return (remote transmit) links in the current discus-
sion. Forward links use the principle of digital code orthogonality of base-to-user
transmissions because all are synchronized. Return links come from unsynchro-
nized transmitters, however, and hence cannot employ orthogonality to reduce RFI
impacts.
Terrestrial wireless networks generally use some form of spread-spectrum
transmission to simplify assignment of channels to cells and to reduce RFI. Gener-
ally speaking, the forward direction is the easier to deal with since, in a given cell,
there is one dominant forward carrier and RFI is caused by a few distant base
stations transmitting their respective forward carriers on the same frequency. The
user terminal receiver then only has to cancel RFI from a limited number of sources
(e.g., one or two). On the other hand, the fact that there are many user terminals
transmitting return carriers, which are transmitted independently of each, results in
a more difficult RFI control environment. Mitigating this factor is that the receiver
8.3 Interference Cancellation 179
is now at the base station and can be more complex and sophisticated in terms of
its ability to process and potentially remove RFI. The base station can employ mul-
tiple receivers of greater complexity to decode all desired and potentially interfer-
ing signals and perform the type of cancelling that is to be discussed.
Modern wireless systems are employing antennas that are based on the concept
of multiple-input, multiple-output (MIMO). A MIMO antenna might have four
beams that cover some angular range. Each antenna connects to its own receive
chain, which is typical of how sectors were laid out in the past. The connections
are made through some type of phase and amplitude combining network similar
to what is shown in Figure 8.7 for two elements or Figure 2.26 for seven elements,
however. We have discussed how RF cancellation can be used to null the pattern
in a specific direction. The context here, however, is for systems where interference
cancellation is done at baseband after demodulation. This improves data quality in
the same way as forward error correction but does it by processing samples of the
RFI to enhance the recovered data.
There appear to be two viable approaches to interference cancellation at base-
band: parallel interference cancellation (PIC) and successive interference cancella-
tion (SIC), illustrated in Figures 8.8 and 8.9, respectively [1]. As shown in Figure
8.8, PIC is characterized in that all users are detected individually and processed si-
multaneously. The technique can be repeated to improve cancellation (in much the
same manner as turbo codes do). The initial very course estimate may be used to
cancel some of the interference, allowing parallel detection to be repeated. Because
of this repetition, PIC has been referred to as multistage interference cancellation.
Soft interference cancellation is necessary since the first state usually results in very
noisy data estimates.
SIC, shown in Figure 8.9, detects one user per stage, where the strongest signal
is detected first and used to cancel itself out of later stages that detect the weaker
signals. The received signal for each signal is reconstructed by recreating the trans-
mit signal, allowing subsequent users to experience a cleaner signal. All users have
improved performance because the earlier user starts with a stronger signal while
later users experience better cancellation of this same stronger signal.
Figure 8.8 Parallel interference cancellation, which detects all users simultaneously. The initial es-
timate may be used to cancel some interference, followed by repeated parallel cancellation. (Repro-
duced from [1] with permission.)
180 How
�����������������������������������������������������
to Identify, Prevent, and Fix Common RFI problems
Figure 8.9 Successive interference cancellation, which detects just one user per stage. The stron-
gest received signal is detected first, then the next strongest, and so on. (Reproduced from [1] with
permission.)
PIC has decreased latency because all users must be processed in parallel; but
increased complexity because there must be a quantity of cancellation stages for
each user as well. In contrast, SIC only requires one stage per user, reducing com-
plexity; latency increases at the same time because of the sequential nature of pro-
cessing (i.e., one stage after another). The analysis shows that if all users are equal
in power (a condition rarely met in practice), PIC has superior performance; how-
ever, SIC wins out in the real-world where power levels vary. This imbalance of
power can be exploited in the system by assigning users at cell boundaries a lower
power to reduce interference caused to their neighboring cells.
In summary, base station return-channel receivers need to suppress RFI from
many independent user transmitters, in contrast to forward-channel receivers in
user terminals that must be able to suppress interference from just a few neighbor-
ing cells. The latter employ a single antenna or at most two antennas and so RF
cancellation is not available as a part of the solution. The benefit will be the reduc-
tion in the required distance between cells operating on the same frequency or an
increase in throughput within the same cell.
Work continues in the wireless arena to address RFI as the primary limita-
tion of coverage and service capability. The 3rd Generation Partnership Project
(3GPP™) is studying Network-Assisted Interference Cancellation and Suppression
(NAIC) as a primary means to boost capacity in denser cell deployment [4]. This
is because cochannel interference, either from independent intercell or coordinated
intracell users is expected to become the limiting factor for achieving higher net-
work capacity. The idea is to increase the role of the receiver to achieve interference
rejection combining (IRC) in user equipment, applying network assistance where
appropriate. The tradeoff in evaluating various IRC receiver structures include:
performance, complexity, network coordination, and network signaling. The block
diagram of the link abstraction model for one form of IRC receiving system is given
in Figure 8.10. This covers the Link-to-System (L2S) as applied to system-level
interference cancellation (SLIC), described by Andrews [1] as NAIC.
According to 3GPP, synchronous network deployment is assumed for NAIC
receivers but receiver performance degradation from timing, frequency synchro-
nization error and asynchronous deployment is for future study. Another area for
8.3 Interference Cancellation
Figure 8.10 Block diagram of the L2S model (only modulation part of the modulation and coding system is needed for SLIC) [4].
181
182 How
�����������������������������������������������������
to Identify, Prevent, and Fix Common RFI problems
future study is misalignment of control regions and whether better network signal-
ing or UE detection of interference are needed. The trend then involves intelligence
in network management and improved signal waveforms that are more resilient to
RFI.
The fault tree identifies the cause and FMEA allows us to choose the best means
of resolution. The academic approach, part of the field of industrial engineering,
8.4 RFI as an Incident Needing Effective Tools and Management 183
This tool is based on the Pareto principle, which states that in a given situation, a
small number of causes (usually 20% of the total) account for most of the prob-
lems (usually 80% of the total). Often called the “80/20 Rule,” the Pareto concept
suggests that you can get the greatest results for the least amount of effort by
identifying and focusing on a few key issues. It requires collecting and organizing
data into root causes or categories of causes. The data are then displayed in a bar
chart [histogram] to give a visual representation of the relative importance of the
root causes. A cumulative percentage line shows the contribution of each category
to the total problem.
•• Incident command: Sets the incident objectives, strategies, and priorities and
has overall responsibility for the incident;
•• Operations: Conducts operations to reach the incident objectives. Establishes
the tactics and directs all operational resources;
•• Planning: Supports the incident action planning process by tracking re-
sources, collecting/analyzing information, and maintaining documentation;
•• Logistics: Provides resources and needed services to support the achievement
of the incident objectives;
•• Finance and aministration: Monitors costs related to the incident. Provides
accounting, procurement, time recording, and cost analyses.
FEMA adopted the Incident Management System (IMS) to improve the rapid
response of emergency resources that come from different agencies and groups.
The key is coordination of efforts, elimination of duplication and gaps, and im-
provement of information availability during troubled times. For the typical RFI
incident, we can probably focus on the first three or four functions, leaving part
of logistics and finance and accounting to an existing administrative structure. (In-
cidentally, “logistics” includes emergency communications, using radio and other
means.) Command and control, operations, and planning are key to a successful
response and correction of RFI issues that affect a working radiocommunication
system.
Frequent physical and virtual meetings are an effective means of brainstorm-
ing, generating action items, and creating data and tactics for evaluation by the
184 How
�����������������������������������������������������
to Identify, Prevent, and Fix Common RFI problems
team. From this background, we move into the specifics of how an RFI condition
can be addressed once we understand its causes. We include the basic means of tak-
ing corrective action: frequency planning, physical separation filtering and shield-
ing, and very important, RFI cancellation.
References
[1] Andrews, J. G., “Interference Cancellation for Cellular Systems: A Contemporary Over-
view,” IEEE Wireless Communications, April 2005, p. 19.
[2] Steinberg, B. D., Principles of Aperture and Array System Design, New York: John Wiley
& Sons, 1976, p. 282.
[3] Monzingo, R.A., and T. W. Miller, Introduction to Adaptive Arrays, Raleigh, NC: Scitech
Publishing, Inc., 2004, p. 48.
[4] 3GPP™, 3rd Generation Partnership Project; “Technical Specification Group Radio Access
Network, Study on Network-Assisted Interference Cancellation and Suppression (NAIC)
for LTE (REleast 12),” 3GPP TR 36,866, V12.0.1, (2014-03), Technical Report.
[5] FEMA, ICS-100: ICS for Higher Education—Student Manual, Unit 4: Incident Com-
mander and Command Staff Functions, November, 2008, https://www.training.fema.gov/
emiweb/is/is100he/student%20manual/l4_ics100highered_sm.pdf.
[6] Marshall, J., An Introduction to Fault Tree Analysis (FTA), The University of Warwick, PE-
USS 2011/2012, http://www2.warwick.ac.uk/fac/sci/wmg/ftmsc/modules/modulelist/peuss/
slides/section_11b_fta_lecture_slides_compatibility_mode.pdf.
[7] Thayer-Hart, N. (ed.), Facilitator Tool Kit—A Guide for Helping Groups Get Results,
Office of Quality Improvement, University of Wisconsin, Madison, Version 2.0, Revised
September 2007. http://oqi.wisc.edu/resourcelibrary/uploads/resources/Facilitator%20
Tool%20Kit.pdf.
CHAPTER 9
185
186 Prospective
�������������������������������������������������������������������
for RFI Resolution in Future Radiocommunication Systems
The term cognitive radio (CR) and its concept have been around for approximately
15 years at the time of this writing. The ITU defines CR as follows [2]:
A radio system employing technology that allows the system to obtain knowledge
of its operational and geographical environment, established policies and its in-
ternal state; to dynamically and autonomously adjust its operational parameters
and protocols according to its obtained knowledge in order to achieve predefined
objectives; and to learn from the results obtained.
including the ITU, European telecom regulatory agencies, and the FCC, are work-
ing to see it implemented at some point. The intent is to have non-CR users op-
erating in the same spectrum and physical space as CR users. Historically, these
would be designated in regulatory terms as primary and secondary spectrum users,
respectively. The primary user should enjoy full use of the spectrum without harm
or unacceptable interference from the secondary users; while the secondary users
employ available spectrum resources in a way that does not conflict with the en-
joyment of the primary users. Further, secondary users must accept all RFI from
primary users who operate in accordance with their licenses. We know that in the
absence of one of the strategies given in Figure 8.1, introduction of secondary users
can produced unacceptable RFI absent something like CR.
Simon Haykin provided clear definition and exposition on CR in his seminal
paper published by the IEEE [3]. In it, he defines CR as:
Figure 9.1 The cycle of a cognitive radio, encompassing receiver tasks of radio-scene analysis and
channel identification, and a transmitter task of transmitter control and dynamic spectrum manage-
ment [3].
however, the devices can remain in the licensed cellular spectrum but operate in a
way that does not interfere with the licensed carrier’s network. By communicating
directly on a single-hop basis, the devices use half the spectrum they would have
needed if a base station had been employed, and with less power provided they
are in close proximity (typically the case with such applications). This is illustrated
in Figure 9.2, where in (a) we have D2D in operation that is independent and
unaware of the enhanced Node B (eNB) base station (BS) and its backhaul to the
mobile switching center. In case (b), direct D2D communication is allowed, but it
is managed through the standard wireless link with the base station that is aware
of D2D services.
According to the authors in [7], “We have proposed a semi-distributed CSA
solution in which cognition at the D2D terminals allows interference-aware deci-
sion making and limited control at the BSs helps the D2D users in selecting the
spectrum band with the least interference.” They describe allowing simultaneous
D2D and cellular communication in the same spectrum (i.e., “in band” communi-
cation). There are challenges involving interference management among the coex-
isting cellular and D2D users along with the need for a low-complexity centralized
or semi-distributed spectrum access and interference mitigation techniques with
minimal signaling overheads, and the notion of priority for the resource manage-
ment of D2D and cellular user equipment. They suggest a tiered approach rather
than all-in with CR. This is because the complexity of CR may only be needed un-
der conditions of excessive RFI or traffic loading. The rest of the time, D2D users
can transmit in spectrum holes that appear static while cellular users continue to
operate with their associated base station sites. The overhead associated with CR,
requiring communication between D2D users and the base stations, will probably
result in less resource for their use; however, they will have access and the ability to
transfer some quantity of data.
Spectrum sharing between cellular users and D2D pairs allows higher spectrum
reuse. However, it may lead to severe cross-tier interference at D2D links when they
coexist with cellular users and small cell tiers in a multi-tier cellular network. The
concept of multitier celluar is part of the forthcoming body of standards. Accord-
ing to Ekram Hossain, 5G systems will adopt a multitier architecture consisting
of macrocells, different types of licensed small cells, relays, and D2D networks to
serve users with different quality-of-service (QoS) requirements in a spectrum and
Figure 9.2 D2D wireless communications (a) unaware channel and (b) aware channel either with-
out network control or with network control. (From [5], with permission.)
9.2 Cognitive Radio 191
•• D2D-unaware spectrum access (DUSA): A base station can utilize any chan-
nel to serve any cellular user based only on its scheduling policy, whereas
each D2D pair selects a random channel.
•• D2D-aware spectrum access (DASA): A base station first assigns one of the
channels (say cd) for D2D transmissions and exchanges the ID of this channel
with the D2D users in its coverage area over the standard LTE user to user
interface. Once the D2D users are informed of the D2D channel, they are
responsible for initiating the communication session by exchanging control
and data signals with no further supervision from the BS. This channel is not
exclusive for D2D users and can be used for cellular transmissions.
In Figure 9.4 [6], cognition improves the CAP for a D2D transmit terminal for
both the DUSA and DASA strategies. Cognition is most effective in dense networks
where RFI is greatest, however. For example, the improvement in CAP with DUSA
due to cognition is only 3% (i.e., from 0.85 to 0.87) when the density of CUE is
10 CUEs/km2 compared to 85 percent (i.e., from 0.29 to 0.53) when the density of
CUEs is 50 users/km2. This result suggests that CSA is not crucial when the number
of CUEs is low and the D2D TXs have a good chance of finding free channels. On
the other hand, for dense networks, CSA is crucial to avoid the nearby interferers
and increase the efficiency of using the available resources.
192
Prospective
Figure 9.3 A multitier network composed of macrocells, picocells, femtocells, relays, and D2D links. Arrows indicate wireless links, whereas the dashed lines denote
the backhaul connections [7].
for RFI Resolution in Future Radiocommunication Systems
�������������������������������������������������������������������
9.2 Cognitive Radio 193
Figure 9.4 Channel access probability (CAP) for DUSA and DASA policies, with and without cogni-
tive spectrum access (CSA) [6].
One of the promises of CR as conceived early on was some means of the new
CR system actually improving the ability of non-CR users to transfer their informa-
tion. A way to accomplish this is to reduce the total noise within the band of the
non-CR user. This is done with a new regime called transform-domain communi-
cation system (TDCS) that removes its energy from areas where there is or could
be other carrier energy. In a basic TDCS implementation, spectral interference and
friendly signal presence are estimated using Fourier-based or general spectral esti-
mation techniques. According to the Air Force Institute of Technology [14], TDCS
would accomplish the following:
detection estimation are employed at the receiver. It is interesting to note that the
idea of notching CDMA signals was proposed by Dr. Larry Milstein of University
of California, San Diego, in a presentation he made to the IEEE prior to 2000 [16].
We have briefly explored the concept and structure of CR based on the theory
and one possible application in cellular wireless systems. The logical ability of CR
to employ spectrum holes cannot be disputed. However, the challenge still remains
to find ready applications including the needed intelligence and communication to
provide immediate information about the RFI environment and the sensitivities of
non-CR occupants. Ideally, CR is a way to employ the common band approach
given in Figure 8.1 and still manage RFI as part of providing an acceptable service
to all who require it.
Our review in Section 7.1 of current spectrum analysis and monitoring devices and
systems establishes the basis for knowing the current status of desired signals as
well as RFI within the spectrum of interest. Discussions of interference cancellation,
spectrum sharing and CR provide some hope for new methodologies to mitigate
interference effects and increase utility of and access to valuable spectrum. We now
move in this section to the regulatory side of the equation to see how international
and national agencies may apply the tools, some of which are still experimental and
some which are purely theoretical at this point in time. The conclusion addresses
advanced concepts in spectrum monitoring and RFI remediation.
Figure 9.5 National Model City framework, involving the FCC, NTIA and supported by the Na-
tional Institute of Standards and Technology: Center for Advanced Communications.
9.3 Spectrum Management Tools and Resources 197
Section 6.4.1 and mentioned above. Figure 9.6 presents the basic elements un-
der evaluation, including the incumbent 3.5-GHz government (Navy) radar and
proposed access for licensed LTE and unlicensed Wi-Fi. In it, we see a 3.5-GHz
prototype “small cell” connected to a network on one side and to separate base sta-
tions for LTE and Wi-Fi. Also shown is a Naval radar operating in the same band.
This is more comprehensive than the testing by NTIA because it includes back end
functions and a scheme for managing service in a manner to protect the Naval ra-
dar from operation of commercial users. Recall from Section 6.4.1 that there are
conditions where the LTE user downlink is detrimentally affected by operation of
the radar while interference to the radar was not evaluated in the particular tests.
Critical to Alcatel Lucent’s operation is the Spectrum Access Server (SAS),
which provides the following functions:
This all sounds very good but the proof will come as testing goes forward and
eventually as real projects unfold. We saw that ATC and spectrum sharing between
radar and LTE can be problematic owing to the level of RFI and its effect on victim
receivers (some of which may not be part of either radiocommunication system).
Adding a layer of spectrum management, with CR capabilities, can mitigate much
of the RFI. But, there must be restrictions on power levels, spectrum occupancy in
certain circumstances, and the quantity of users on both sides of the sharing space.
We still have the situation that RFI to one end is communication to the other. Us-
ers that already enjoy their services will find it difficult to accept new entrants who
operate in different and potentially interfering ways. Generally, it requires a more
powerful authority that can impose new rules and require that incumbents accept
some level of RFI or operating restrictions even if they had clear bandwidth in the
Figure 9.6 Alcatel-Lucent 3.5-GHz Network Platform for Research and Testbed [12].
198 Prospective
�������������������������������������������������������������������
for RFI Resolution in Future Radiocommunication Systems
past. The latter can be expected to do this provided that the pain is manageable.
The new comers, on the other hand, need to understand what the restrictions on
their use means to the applications they provide or employ. This appears to be an
objective of the Model City program, where the various sides come together in a
test case that foretells a successful offering of services. While the results may be
encouraging for the particular case under investigation, we have to be careful not
to generalize about something as complex as spectrum sharing using CR or other
unproven scheme that looks good on paper.
The Defense Research Projects Agency (DARPA), made famous by its support
of ARPAnet that evolved into the Internet, has considered the use of CR in a de-
fense context. Here, the object is to achieve a fault tolerant radiocommunication
system that provides diverse and automated routing much like that of the Internet
Protocol family [11]. This is the DARPA Wireless Network After Next (WNaN)
cognitive radio program, among whose objectives are:
Someone who has been around DoD advanced research and development for a
long time will recognize much of this as the standard wish list. However, we may be
at the point where critical aspects are within reach. One of these facilities is DTN,
which can be defined as follows:
The delay and disruption tolerant network demonstrated the benefits of hop by
transport of contents, and the use of adaptive caching and persistence within wire-
less networks.
This recognizes that radiocommunication must provide real-time access for the
broad range of uses, but there are cases where some delay is allowable but with
guaranteed delivery.
systems will require better RFI mitigation resources, enumerated next (but this is
not a complete list):
Some of the capabilities needed now in RFI identification and resolution could
include:
•• Improved spectrum analysis that can display signals in both frequency and
time, adjusting automatically and dynamically, without human intervention;
•• Measuring signal power by integrating spectra and waveforms, suggesting
the nature of the modulation or material modulating the signal;
•• Ability to compile a database of signals, automatically indexing signals
and providing a clear user interface to better understand both desired and
undesired signals (being sure the characteristics are clearly displayed and
recorded);
•• Using a knowledge-based system to suggest most likely sources of RFI and
approaches for resolution;
•• Improved interference location with high accuracy and fast response;
•• Means to interconnect various radiocommunication operators and users in
a network that provides situational awareness and direction for identifying
and removing RFI.
The last item recognizes that even with better RFI analysis and resolution tools,
we still need excellent communication to “get the word out” when service is dis-
rupted from an unknown RFI source. We are sure that these capabilities and more
will arise in coming months and years. The fact that communications have largely
gone “wireless” provides motivation to technology providers to be as innovative
in this area as we have seen them to be in social networks and smart phones apps.
Underlying these innovations is the basic precept that the party being interfered
with will experience a loss of some kind. This can be measured in a variety of terms:
References
[1] Goldsmith, A., “Breaking Spectrum Gridlock With Cognitive Radios: An Information The-
oretic Perspective,” Proceedings of the IEEE. Vol. 97, No. 5, May 2009, pp 894.
[2] ITU-R, “Definitions of Software Defined Radio (SDR) and Cognitive Radio System (CRS),”
Report ITU-R SM.2152 (09/2009).
[3] Haykin, S., “Cognitive Radio: Brain-Empowered Wireless Communications,” IEEE Jour-
nal on Selected Areas in Communications, Vol. 23, No. 2, February 2005, pp. 201.
[4] Mitola, J. “Cognitive Radio: An Integrated Agent Architecture for Software Defined Ra-
dio,” Doctor of Technology, Royal Inst. Technol. (KTH), Stockholm, Sweden, 2000.
[5] Faussurier, E., “Introduction of New Spectrum Sharing Concepts: LSA and WSD,” ITU-
R SG 1/WP 1B Workshop: Spectrum Management Issues on the Use of White Spaces by
Cognitive Radion Systems, Geneva, January 20, 2014.
[6] Sakr, A. H., et al, “Cognitive Spectrum Access in Device-to-Device-Enabled Cellular Net-
works,” IEEE Communications Magazine, July 2015, pp. 126.
[7] Hossain, E., et al, “Evolution Towards 5G Multi-tier Cellular Wireless Networks: An In-
terference Management Perspective,” Wireless Communications, IEEE, Vol. 2, No. 3, June
2014.
[8] Zhang, R., et al, “Exploiting Multi-Antennas for Opportunistic Spectrum Sharing in Cog-
nitive Radio Networks,” IEEE International Symposium on Personal, Indoor and Mobile
Radio Communications (PIMRC), Athens, Greece, September 2007.
[9] Department of Commerce, National Telecommunications and Information Administra-
tion [Docket Number: 140708559–4559–01], RIN 0660–XC011, FCC, [Docket No. ET
Docket No. 14–99], “Model City for Demonstrating and Evaluating Advanced Spectrum
Sharing Technologies,” Federal Register/Vol. 79, No. 135/Tuesday, July 15, 2014/Notices.
[10] FCC, Workshop on Model City Program, April 14, 2015.
[11] Moorefield, F. D., Jr., DoD CIO Brief to FCC-NTIA Model City Workshop, April 15, 2015.
[12] Buddhikot, Milind M., Technologies for 3.5 GHz Shared Spectrum, FCC Model City
Workshop, MMB, April 15, 2015.
[13] Marshall, Preston, Quantitative Analysis of Cognitive Radio and Network Performance,
Norwood, MA: Artech House, 2010.
[14] Chakravarthy, V. et al., “TDCS, OFDM, and MC-CDMA: A Brief Tutorial,” IEEE Radio
Communications, September 2005, p. S11.
[15] Zhong, W., et al, “Joint Resource Allocation for Device-to-Device Communications Under-
laying Uplink MIMO Cellular Networks,” IEEE Journal on Selected Areas in Communica-
tions, Vol. 33, No. 1, January 2015, p. 41.
[16] Milstein, L., “Wideband Code Division Multiple Access,” invited paper IEEE Journal on
Selected Areas in Communications, Vol. 18, No.8, August 2000, p. 1344.
[17] ArrayComm, “Cooper’s Law,” http://www.arraycomm.com/technology/coopers-law/.
List of Acronyms
201
202 List
����������������
of Acronyms
RX receiver
S-band frequencies between 2 and 3 GHz (specifically, 2.2 to 2.5 GHz space
application)
SA single access system on TDRS
SAS spectrum access server
SCPC single channel per carrier
SDR software defined radio
SFD saturation flux density
SHF super high frequency (3 to 30 GHz)
SIC successive interference cancellation
SLIC system-level interference cancellation
SNR signal to noise ratio (usually post-detection)
TAP Terrain Analysis Package, software by Softwright, Inc.
TDD time division duplex
TDCS transfer-domain communications system
TDMA time division multiple access
TDRS tracking and data relay satellite system
TIREM Terrain Integrated Rough Earth Model
TWG Technical Working Group
TWT traveling wave tube
TX transmitter
UAV unmanned aerial vehicle
UE User equipment
UHF ultra-high frequency (300 to 3,000 MHz)
VBW video bandwidth (spectrum analyzer setting)
VHF very-high frequency (30 to 300 MHz)
VSAT very small aperture terminal (two-way satellite communications
terminal)
WiMAX Worldwide Interoperability for Microwave Access standard
XPI cross-polarized interference
WNaN wireless network after next
WRC World Radiocommunication Conference
About the Author
205
Index
A due to atmospheric gases, 85
Acronyms, this book, 201–4 free space, 7
Air-to-ground radiocommunication, 73–74 median, 107
Alcatel-Lucent Network Platform, 196–97 path, 7
Amplifiers rain, 84–88
low-noise (LNA), 44 smooth Earth, 95
transfer characteristics, 53 Automatic gain control (AGC), 22
traveling-wave tube (TWT), 27
Amplitude and phase shift keying (APSK), 48 B
Amplitude modulation (AM), 74 Band segmentation, 171
Ancillary Terrestrial Component (ATC) Baseband cancellation, 178–82
defined, 133 Baseband processing, 178
hybrid terrestrial/satellite wireless network Bent-pipe configuration, 76
based on, 133 Bit error rate (BER), 22, 121, 125, 158
introduction of, 156 Block error rate (BLER)
Antennas, 29–46 defined, 127
array of elements, 40–41 Gaussian noise interference, 128, 129, 130
characteristics and isolation, 122–23 interference to LTE UE receiver, 130, 131
coupling to, 80–83 Block Upconverter (BUC), 156
dipole, 29–30 Broadcast Satellite Service (BSS), 9
directional, 33–34
discone, 32 C
Fillmore, CA earth station, 102 Cancellation
flat-panel array, 40–41 bandwidth, 176–77
interference cancellation at, 174–77 at baseband, 178–82
log-periodic array, 35 at IF, 177–78
MIMO, 178, 191 parallel, 179, 180
monopole, 29–30 at RF, 174–77
multiple element, 34–35 sidelobe, 175
nondirectional, 29–31 successive, 179–80
phased array, 41–44 system-level, 180–82
reflector aperture, 35–40 with two-element phased array, 176
RF electronics integration with, 44–46 Carrier ID, 155
waveguide horn, 31–33 Carrier-to-noise (C/N) ratio
Yagi-Uda array, 34–35 analysis, 56–57
Array of elements, 40–41 BER and, 125
Atmospheric loss, 84 C/I relationship, 120
Attenuation link, 125
dry-air and water-vapor, 86
207
208 Index
Low Earth orbit (LEO) satellites, 73–74, 77, Network-Assisted Interference Cancellation
136 and Suppression (NAIC), 180
Low-noise amplifiers (LNAs), 44 Nondirectional antenna radiation patterns,
29–31
M Nongeostationary satellite orbit (NGSO), 136
Maximum usable frequency (MUF), 68, 69, 70 Nonlinearity, 52
Medium Earth orbit (MEO) satellites, 77, 136 Non-steady state propagation, 123–25
Medium frequency (MF), 68
Mobile Satellite Service (MSS) O
ATC authorization, 134 Obstructed paths
domain, 9 complex propagation models, 105–10
reuse, 134 diffracted, 92–93
RFI assessment from, 134–35 fading on, 96–100
spectrum availability, 134 geographic coverage analysis, 100–105
Model City program, 195–98 profiles, 91–95
Alcatel-Lucent Network Platform, 196–97 propagation on, 91–110
defined, 195 Ofcom, 9
framework, 196 Okumura propagation models, 106–8
objective, 198 Orbit locations, process structure for obtaining,
Modems 10
defined, 22 Orthogonal frequency division multiplex
demodulation, 23 (OFDM), 128
elements of, 22 Ortho-mode transducer (OMT), 28
power divider, 23 Oscillators, 172
properties of, 22–24 Over-the-horizon propagation, 89
Modulation and coding scheme (MODCOD),
127, 156 P
Monopole antennas, 29–30 Packet error rate (PER), 125
Multipath Parallel interference cancellation (PIC), 179,
link quality and, 97 180
mobile receiver subject to, 96 Path attenuation, 7
Ricean fading, 97–98 Path computation using software, 101–3
signal occurrence, 98 Path geometries, 66–80
vector combination of direct signal and, 97 air to ground, 73–74
Multiple element antennas, 34–35 ground to ground, 68
Multiple-input, multiple-output (MIMO) space to Earth, 74–78
antennas, 178, 191 space to space, 78–80
Multitier network, 192 Path profiles
defined, 91
N illustrated, 92
National Advanced Spectrum and obstructions and, 91–95
Communications Test Network Phased arrays
(NASCTN), 196 defined, 41
National Telecommunication and Information geometries, 42
Agency (NTIA), 126–27, 157 performance of, 42
Near Vertical Incidence Skywave (NVIS), 70 seven-element, 43
212 Index