Shun-Ping Chen · Heinz Schmiedel

RF Antenna
Beam Forming
Focusing and Steering in Near and Far Field
RF Antenna Beam Forming
Shun-Ping Chen · Heinz Schmiedel

RF Antenna Beam Forming

Focusing and Steering in Near and Far Field
Shun-Ping Chen Heinz Schmiedel
Hochschule Darmstadt Hochschule Darmstadt
University of Applied Science University of Applied Sciences
Darmstadt, Germany Darmstadt, Germany

ISBN 978-3-031-21764-7 ISBN 978-3-031-21765-4 (eBook)

https://doi.org/10.1007/978-3-031-21765-4

© The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Switzerland AG
2023
This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether
the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse
of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and
transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or
dissimilar methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does
not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective
laws and regulations and therefore free for general use.
The publisher, the authors, and the editors are safe to assume that the advice and information in this book
are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the
editors give a warranty, expressed or implied, with respect to the material contained herein or for any errors
or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in
published maps and institutional affiliations.

Responsible Editor: Reinhard Dapper

This Springer imprint is published by the registered company Springer Nature Switzerland AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
To Michael, Anna, Katja and Angela
Preface

This book has been written for students in Electrical and Electronics Engineering and
Information Technology. It is also written for radio frequency (RF) engineers to give a
technology oriented overview of phased array antennas, beam steering and beam form-
ing. This may allow engineers and also project managers, involved in RF design, to
develop visions and opportunities with their own applications.
Phased array antennas have been developed since the 1960s and have seen many
applications. Some of these use fixed arrays to increase the aperture of an antenna sys-
tem to obtain a desired gain and performance. Others use controlled phases and ampli-
tudes to allow beam steering and beam forming. The antenna system performance
depends on the antenna elements and on the proper phase and amplitude control. With
phase control for the individual antenna elements an antenna beam can be steered over
a wide angle range. By additionally controlling the amplitudes for all antenna ele-
ments individually the beam can be formed to given specification, e.g. side lobes sup-
pression. Well known, successful applications of steerable antenna arrays are e.g. part
of the PATRIOT air defense system. Also IRIDIUM relies on phased array panels with
steerable beams for worldwide telephone and data communication. Beam steering is also
essential for existing and future 5G mobile communication systems. A whole chapter in
this book has been dedicated to these important MIMO-antenna systems. Most readers
will be familiar with Maxwell’s equations and electromagnetic waves. Nevertheless a
first chapter is a brush up for Maxwell’s equations and electromagnetic wave theory, as
far as needed as foundation for the antennas described later in the book. It may help the
reader to get started into the antenna topic. In the following chapter basic antenna ele-
ments are described. All these may be used in phased array antenna systems. Well known
are dipole arrays for TV reception, helical antenna arrays on the GPS satellites, patch
antennas on the IRIDIUM satellites, and many more.
Next is a comprehensive chapter on linear antenna arrays. It explains a complete
set of arrangements for defined steering and beam forming. Comprehensive simulation
results, along with measurement results, are discussed in detail. Examples are shown
for defined beam steering angles. With an appropriate distribution of amplitude weights
the shape of the beam can be designed. Equal weight, Chebyshev and binomial, that

VII
VIII Preface

is similar to a Gaussian weight, are investigated, simulated and measured for different
beam steering angles. The performance of these linear antenna arrays is investigated for
far field and near field also.
The results for these linear arrays are put into a two-dimensional context in the next
chapter. Planar arrays, where the antenna elements are placed in a two-dimensional
plane, are investigated. Again beam steering is simulated and beamforming is applied
to obtain desired beam shapes. So far all descriptions were related to antenna arrays ori-
ented in one plane only.
In a logical next step so called conformal arrays will be considered. It begins with a
modification of the linear array which will have a defined curvature now. All antenna ele-
ments are placed on a contour which is a part of a circle’s circumference. There are two
options, one is the concave and the other the convex configuration. The concave arrange-
ment allows, as a special case, to irradiate a target in a focus point with equal distance
and equal phase to all antenna elements. The convex arrangement on the other side is
interesting to consider for an antenna tower and 360° coverage. Here the antenna ele-
ments can be controlled in phase, and amplitude, so that the antenna beam has full 360°
coverage. In our results a fraction of the whole circle is investigated in detail, inclusive
beam steering and beam forming.
After describing these specific antenna designs, applications are considered in the
next chapters. After an introduction the specific application of state-of-the art MIMO-
antenna systems is covered comprehensively. In MIMO applications antenna beams need
to be switched fast between different users. Also interference cancellation is an issue,
when many users are present. The application part is complemented by considering so
called thinned arrays for deep space communication applications.
The authors would like to interest the readers into the exciting world, visions and
opportunities of phased array antenna systems with all the options of beam steering and
forming.

Darmstadt Shun-Ping Chen

October, 2022 Heinz Schmiedel
Acknowledgements

The authors would like to thank all colleagues of the Institute of Communication Tech-
nologies, Darmstadt University of Applied Sciences h-da, for enabling and supporting
the research activities, and for many helpful discussions.
Shun-Ping Chen would like to thank Prof. Arne Jacob, former head of the Institute of
High Frequency Engineering at Technical University Hamburg, for involving him in the
investigation projects of near field beam forming and beam focusing in August 2011–
March 2012 during his research sabbatical. Since then these became one of his favorite
research topics.
The authors also appreciate the strong support of the Phase and Amplitude Con-
trol Matrix supplier Mitron Inc. Especially they like to thank its president and system
designer, Mr. Wei Li, and his developer colleagues, for rendering every assistance and
several video conferences during the first period after the implementation of the system
in our laboratory, enabling us to successfully start with the experiments.

IX
Contents

1 Introduction to Electromagnetic Wave Propagation. . . . . . . . . . . . . . . . . . . . 1

1.1 Maxwell’s Equations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Vector Algebra. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.3 Wave Propagation in Different Media. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2 Basic Antenna Elements. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.1 Isotropic Radiator. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.2 Dipole Antenna . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.3 Patch Antenna . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.4 Horn Antenna. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
3 Linear Antenna Array . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
3.1 Introduction into Phased Array Antennas . . . . . . . . . . . . . . . . . . . . . . . . . . 22
3.2 Phase Control. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
3.3 Measurement Setup and Verification. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
3.4 Far Field Characteristics of a Linear Patch Antenna Array. . . . . . . . . . . . . 31
3.5 Amplitude Weights and Side Lobes of the Radiation Pattern. . . . . . . . . . . 36
3.6 Near Field Characteristics at a Distance of 10  . . . . . . . . . . . . . . . . . . . . . 37
3.7 Near Field Characteristics at a Distance of 5.6 . . . . . . . . . . . . . . . . . . . . . 44
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
4 Planar Antenna Arrays. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
4.1 Beam Forming, Focusing and Steering of Planar Arrays. . . . . . . . . . . . . . . 50
4.2 Side Lobe Suppression. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
4.3 Simulation Results for Planar Antenna Arrays . . . . . . . . . . . . . . . . . . . . . . 56
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
5 Conformal Antenna Array. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
5.1 Beam Forming of One-Dimensional Conformal Concave Arrays. . . . . . . . 63
5.2 Beam Forming of One-Dimensional Conformal Convex Arrays. . . . . . . . . 77

XI
XII Contents

5.3 Beam Forming of 2-Dimensional Conformal Arrays . . . . . . . . . . . . . . . . . 92

5.4 Comparison of Convex, Concave and Planar Profiles. . . . . . . . . . . . . . . . . 94
5.5 Simulation Results for Conformal Antenna Arrays. . . . . . . . . . . . . . . . . . . 97
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
6 MIMO Antenna Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
7 Thinned Antenna Array. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
Reference . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
8 Applications of Phased Arrays. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
8.1 Radar, Broadcasting and Positioning. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
8.2 Smart Antenna Systems. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
8.3 5G Massive MIMO Basestation for Multi-User Coverage . . . . . . . . . . . . . 124
8.4 Deep Space Communications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
8.5 Radio Astronomy Antenna Array. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
9 Conclusions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129

Radio Frequency Bands. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131

Glossary. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
Introduction to Electromagnetic Wave
Propagation
1

1.1 Maxwell’s Equations

Maxwell’s equations are the governing equations to analyze all electromagnetic wave prop-
agation problems, from RF radio waves used in cellular mobile communications 1–2 GHz,
mm waves (frequency range of 30–100 GHz) up to optical wave propagation in the frequency
range of 200 THz (see also [1–5]). Depending on the applications, different carrier frequen-
cies are used from 1 GHz to mm-waves or even to the optical spectrum. These frequencies
have different abbreviations, standardized by different organizations like IEEE, ITU, EU
and NATO, and are listed in the appendix A Tab 1.1 shows electromagnetic field parameters
and units.
∂B
∇ ×E=− , (1.1)
∂t
∂D
∇ ×H= + J, (1.2)
∂t
∇ · D = ρ, (1.3)
∇ · B = 0. (1.4)

Table 1.1 Electromagnetic field parameters and units

Field Vector Symbol Unit
Electric field E V/m
Dielectric displacement/density D As/m2
Magnetic field H A/m
Magnetic flux density/induction B T = Vs/m2
Current density J A/m2

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 1

S.-P. Chen and H. Schmiedel, RF Antenna Beam Forming,
https://doi.org/10.1007/978-3-031-21765-4_1
2 1 Introduction to Electromagnetic Wave Propagation

1.2 Vector Algebra

By using vector algebra, the above Maxwell’s equations can be derived in different coordinate
systems depending on the problems to be investigated. For example, for rectangular metal-
lic waveguides, the Cartesian coordinate system (x, y, z, t) will be meaningful, whereas
for problems of cylindrical metallic or quartz glass optical waveguides, the cylindrical
coordinate system (r , φ, z, t) will be used. For the antenna problems the polar coordinate
system (r , θ, φ, t) will be helpful. In the following we will briefly discuss the vector algebra
and the derivation of Maxwell’s equations into the Helmholtz equations as a good approx-
imation [2, 5]. Generally a dynamic field is a physical quantity with different magnitudes
and orientations at different positions and different points in time.
Vector fields that vary with different frequencies f have always characteristic values or
magnitudes, and point into certain directions at certain positions. For example:

Electric field E(x, y, z, t) and dielectric displacement D(x, y, z, t);

Magnetic field H(x, y, z, t) and magnetic induction B(x, y, z, t);
Vector potential A(x, y, z, t);
Mechanical force F(x, y, z, t);
Velocity of solid body or fluid v(x, y, z, t).

In comparison with vector fields, scalar fields have only a value or a magnitude, and are
direction-independent. For example:

Temperature field T(x, y, z, t);

Potential Φ(x, y, z, t);
Electric charges q(x, y, z, t) or charge density ρ(x, y, z, t).

Vector algebra definitions are the following

 
∂ ∂ ∂
∇= , , , (1.5)
∂ x ∂ y ∂z
 
∂ Ex ∂ E y ∂ Ez
∇E = , , = grad E, (1.6)
∂ x ∂ y ∂z

∇ · E = ∇x E x + ∇ y E y + ∇z E z = div E, (1.7)

∇ × E = rot E = curl E, (1.8)

( ∇ × E) z = ∇x E y − ∇ y E x , (1.9)
1.3 Wave Propagation in Different Media 3

( ∇ × E) x = ∇ y E z − ∇z E y , (1.10)

( ∇ × E) y = ∇z E x − ∇x E z . (1.11)
Useful rules of vector algebra are

∇ · ( ∇T ) = ∇ 2 T , (1.12)

∇ × ( ∇T ) = 0, (1.13)

∇( ∇ · E) = a vector, (1.14)

∇ · ( ∇ × E) = 0, (1.15)

∇ × ( ∇ × E) = ∇( ∇ · E) − ∇ 2 E, (1.16)

( ∇ · ∇)E = ∇ 2 E. (1.17)

1.3 Wave Propagation in Different Media

Forces, acting on a charge q, due to electric and magnetic fields

F = q(E + v × B). (1.18)

Principle of superposition of fields in linear media

E = E1 + E2 . (1.19)
From Eqs. (1.1) and (1.2) we see that electric and magnetic fields are always related to
each other and form the electromagnetic waves transporting energy in a certain direction,
as defined by the Poynting vector

S = E × H. (1.20)
4 1 Introduction to Electromagnetic Wave Propagation

The total power, that is transported through the area S  , will be


P= E × H · ds. (1.21)
S
For the investigation of wave propagation or antenna problems, the waves are generated
by the current density J of a source, with a certain carrier frequency f , modulated with
the information data in manifold modulation schemes. These waves will be matched to the
antenna (for example dipole antenna, patch antenna, horn antenna, helix antenna or parabolic
antenna) through the so-called reactive near field, then radiated to the near field and far field
free space which is vacuum in the simplest case.
The reactive near field (r < R1 ), radiating near field (Fresnel region, R1 < r < R2 ) and
far field (Fraunhofer region, r > R2 ) are characterized by R1 and R2 [3] where D is the
diameter of the antenna or largest dimension of an antenna array

D3
R1 = 0.62 , (1.22)
λ

2 D2
R2 = . (1.23)
λ
The description of electromagnetic waves in a generally inhomogeneous dielectric
isotropic medium requires generally numerical methods, in order to solve the differen-
tial equations. In some cases, some simplified assumptions help to solve these differential
equations analytically, as will be discussed later. By considering the generally frequency-
dependent characteristics of the electromagnetic waves in a generally inhomogeneous dielec-
tric medium at the frequency f or angular frequency ω = 2π f , we have
 ∞
E(r, ω) = E(r, t)e− jωt · dt. (1.24)
−∞
This equation is the Fourier transform of the electric field. It allows us to calculate the
frequency spectrum from a given function versus time. Interpreting this equation, we see that
a signal with a high data rate, and where the signal amplitude changes rapidly with time,
has a wide frequency spectrum or needs a large bandwidth. With this complex notation,
Eqs. (1.1) and (1.2) give

∇ × H = jωD + J, (1.25)

∇ × E = − jωB. (1.26)
In generally inhomogeneous media with complex dielectric permittivity coefficients and
complex magnetic permeability coefficients, the relations between the electric field, the
1.3 Wave Propagation in Different Media 5

magnetic field, the dielectric displacement and the magnetic flux density can be described
as follows
 
D(r, ω) = (ε (ω) − jε (ω))E(r, ω), (1.27)

 
B(r, ω) = (μ (ω) − jμ (ω))H(r, ω), (1.28)
whereas the real parts correspond to the so-called dispersion, the imaginary parts correspond
to the losses caused by absorptions. Dispersion describes the frequency-dependent character-
istics, i.e. different wave propagation velocities at the different wavelengths. These relations
are described by the so-called Debye Dispersion Model which describes the delayed reaction
of the molecular dipoles to the applied electromagnetic waves. See for example [4], with
the relaxation time τ which describes the retardation of the molecular dipole in response to
the excitation field.

ε (ω) εr (0) − εr (∞)
= εr (ω) = εr (∞) + , (1.29)
ε0 1 + (ωτ )2

ε (ω) εr (0) − εr (∞)
= ωτ . (1.30)
ε0 1 + (ωτ )2
Additionally the phase shift between D and E is caused by the absorption loss. By
considering the absorption loss of the dielectric medium and the Ohm’s loss caused by the
limited conductivity κ, the total power generated by the source will be

 
1 ∗ 1
− E · J · dV = E × H∗ · dA
2 2
V

A

1 1
+ κ | E |2 · dV − (E · D∗ − B · H∗ ) · dV . (1.31)
2 V 2 V

The left hand term represents the generated power at the transmitter source, for example
by a dipole antenna with the excitation current of a certain carrier frequency f, whereas the
first term of the right hand expressions is the Poynting vector, or the radiated power from
this antenna through the medium, e.g. free space. The real part of the second and third term
represent the loss caused by the absorption, polarization loss and the limited conductivity

of the medium that is directly related to ε or tanδ .
  κ 
ε = ε − jε − j = ε (1 − j tan δ), (1.32)
ω

κ + ωε
tan δ =  . (1.33)
ωε
6 1 Introduction to Electromagnetic Wave Propagation

For homogeneous, time-invariant media (grad ε = 0), i.e. free space in the most relevant
cases of antenna propagation problems, we get the simplified Helmholtz equations by using
the complex dielectric permittivity

∇ 2 E + ω2 μεE = 0, (1.34)

∇ 2 H + ω2 μεH = 0, (1.35)

  
κ + ωε
ε =ε 1+  . (1.36)
jωε
With the definition of the so-called wave number k and wave propagation velocity v,
where v will be the free space velocity of light, if the medium in the most simple case is
vacuum (μ = μ0 , ε = ε0 ).
In many special situations the inhomogeneous media have to be taken into account,
for example electromagnetic waves in optical frequency range suffer amplitude scintillation
due to the dynamic refractive index scintillation caused by atmospheric turbulences, or radio
frequency amplitude and phase scintillation during the propagation through plasma. In the
first case the laser optical beam propagation will be disturbed [6] leading to beam scattering
and beam wander, whereas in the second case the radio frequency waves will suffer phase
shift [7]. It should be pointed out that these special applications are not being discussed
in this book. Instead we would like to treat the most general cases of free space antenna
propagation scenarios
√ ω
k = ω με = . (1.37)
v
This simplest case is also most important for wireless radio frequency communications
in vacuum (μr = 1, εr = 1, v = c), which is considered in the following chapters in this
book.
For homogeneous anisotropic dielectric media the independent permittivity coefficients
become dependent on the orientation in certain media, for example crystals, semiconduc-
tors, etc. In this case the dielectric permittivity coefficients must be described by using
a tensor matrix. A similar tensor matrix is valid for the anisotropic magnetic medium. For
inhomogeneous dielectric, isotropic, non-magnetic and lossless media, the field components
of the electromagnetic waves travelling in an inhomogeneous, dielectric, isotropic, lossless
medium, which is an ideal assumption of course, could be described by the differential equa-
tions generally depending on permittivity and permeability parameters. These differential
equations can be solved by using numerical methods. In some cases the inhomogeneous
medium can be separated into several homogenous media. By doing so, the differential
equations can be solved in each region by using analytical methods, for example in case of
step-index dielectric optical fiber consisting of core and cladding [2].
1.3 Wave Propagation in Different Media 7

Considering a homogeneous medium (ε = constant or ∇ε = 0), for example free space

which is the case for antenna problems, the differential equations can be simplified to the
so-called Helmholtz equations with the amplitude E of the electric field E and unit vector
of the polarization plane u with

u = x · ux + y · uy + z · uz , (1.38)

∂2 E ∂2 E ∂2 E ∂2 E
∇2 E = + + = με (1.39)
∂x2 ∂ y2 ∂z 2 ∂t 2
in Cartesian coordinate systems (x, y, z). Correspondingly the differential equations can be
given in cylindrical coordinate systems (ρ, φ, z) as
 
1 ∂ ∂E 1 ∂2 E ∂2 E ∂2 E
∇2 E = ρ + 2 + = με (1.40)
ρ ∂ρ ∂ρ ρ ∂φ 2 ∂z 2 ∂t 2
or in spherical coordinate systems (r , θ, φ), depending of the antenna problems under inves-
tigation as

   
1 ∂ 2∂E 1 ∂ ∂ 1 ∂2 E ∂2 E
∇ E= 2
2
r + 2 sin θ + 2 2 = με . (1.41)
r ∂r ∂r r sin φ ∂θ ∂θ r sin θ ∂φ 2 ∂t 2

Correspondingly we can derive the Helmholtz equations for the magnetic fields for homo-
geneous free space in a Cartesian coordinate system, a cylindrical coordinate system and a
spherical coordinate system

∂2 H ∂2 H ∂2 H ∂2 H
∇2 H = + + = με 2 , (1.42)
∂x 2 ∂y 2 ∂z 2 ∂t
 
1 ∂ ∂H 1 ∂2 H ∂2 H ∂2 H
∇2 H = ρ + 2 + = με 2 , (1.43)
ρ ∂ρ ∂ρ ρ ∂φ 2 ∂z 2 ∂t

   
1 ∂ 2∂H 1 ∂ ∂ 1 ∂2 H ∂2 H
∇ H= 2
2
r + 2 sin θ + 2 2 = με . (1.44)
r ∂r ∂r r sin φ ∂θ ∂θ r sin θ ∂φ 2 ∂t 2

One very useful transformation between a spherical coordinate system and a Cartesian
coordinate system can help to convert the vector potential, electric or magnetic field com-
ponents into another coordinate system, in order to enable a very efficient analysis process
⎡ ⎤ ⎡ ⎤⎡ ⎤
Ar sin θ cos φ sin θ sin φ cos θ Ax
⎣ Aθ ⎦ = ⎣ cos θ cos φ cos θ sin φ − sin θ ⎦ ⎣ A y ⎦ . (1.45)
Aφ − sin φ cos φ 0 Az
8 1 Introduction to Electromagnetic Wave Propagation

Similarly, for the extremely small variation of the relative dielectric permittivity less
than 1%, for example in case of a dielectric waveguide like quartz optical fiber, the so-called
“weekly guidance condition” is valid, where the relative gradient of the permittivity vanishes
to zero, so that the Helmholtz equation is also valid.
The wave number vector is pointing into the propagation direction of the electromagnetic
waves and c is the vacuum speed of the light. The corresponding solution can be found
similarly. The simplest solution for Maxwell’s equations or the Helmholtz equation could
be the so-called transverse electromagnetic (TEM) waves. For a TEM wave polarized in
x-direction, the power density, represented by the Poynting vector S, will be transported in
z-direction

E(r, t) = E(r, t) · ux , (1.46)

H(r, t) = H (r, t) · uy , (1.47)

S(r, t) =| E(r, t) × H(r, t) | ·uz . (1.48)

where r is a vector pointing to the observation position in space.
The ratio between the perpendicular electrical and magnetic field 120 π Ω or 377 Ω is a
constant, and is defined as the characteristic impedance of the medium which could be air
or a homogeneous dielectric medium. Helmholtz equations are valid for the far field.
The governing Maxwell’s equations can be used to explain almost all the wave prop-
agation mechanisms, both for free space wave propagation, antenna problems and optical
waveguides.

References

1. R. E. Collin, F. J. Zucker: Antenna Theory, Part 1. McGraw-Hill Book Company (1969).

2. H.-G. Unger: Planar Optical Waveguides and Fibres. Oxford Clarendon Press (1977).
3. C. A. Balanis: Antenna Theory. John Wiley & Sons, Inc. Fourth Edition (2016).
4. K. W. Kark: Antenna Antennas and Radiation Fields (in German: Antennen und Strahlungsfelder).
Springer Vieweg (2018).
5. S.-P. Chen: Fundamentals of Information and Communication Technologies. Cambridge Scholars
Publishing (2020).
6. S.-P. Chen: Investigations of free space and deep space optical communication scenarios. CEAS
Space Journal, Springer Nature (2021).
7. S.-P. Chen, J. Villalvilla: Comparison of Modified Woo’s Solar Phase Scintillation Model with
ESA’s BepiColombo Superior Solar Conjunction Measurement Data for X-Band and Ka-Band.
CEAS Space Journal, Springer Nature (2022).
Basic Antenna Elements
2

Antenna problems, free space radio frequency and optical wave propagation, multipath
propagation or scattering problems can be analyzed by using Maxwell’s equations described
in the last chapter.
Generally the transmit antenna is excited by the current density J, generating and emitting
a wave which propagates into free space in a certain defined direction, thus transmitting
information to a receiver.
By considering the current density distribution of different radiators like dipoles, patches
and horns, the vector potential A(r, t) is used to derive the magnetic field H and electric
field E at an observation point at a distance r from the source r
 
e−jk|r−r |
A(r) = J(r ) dS  , (2.1)
S 4π | r − r |

H = ∇ × A, (2.2)

1 1
E= ∇ ×H= (grad divA + k 2 A). (2.3)
jω0 jω0
Typical and often-used antenna types (Fig. 2.1) with special radiation characteristics are:

• Theoretical, hypothetical isotropic radiator

• Dipole antenna
• Patch antenna
• Horn antenna
• Helix antenna
• Parabolic antenna
• Cassegrain antenna
• Slot antenna, etc.

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 9

S.-P. Chen and H. Schmiedel, RF Antenna Beam Forming,
https://doi.org/10.1007/978-3-031-21765-4_2
10 2 Basic Antenna Elements

The performance and characteristics of the different types of antennas are characterized
by the antenna gain G which describes the increase of the power density in the wanted
direction of the antenna radiation beam in comparison with the isotropic radiator (dBi). The
isotropic antenna radiates into all directions equal power density. It is a very useful theoretical
hypothetical model. The higher the gain, the narrower the antenna radiation beam will be.
The increase in power density into the wanted direction is related to the decrease of radiated
power density into other, unwanted directions. The total radiated power stays constant. The
hypothetical isotropic antenna is considered as a reference.
The antenna radiation characteristics can be described by the so-called directivity

S(r , θ, φ)max
D(r ) = . (2.4)
Sr ,θ,φ
Since the radiation field distributions in the near field and far field are not exactly the
same, D(r ) varies slightly depending on the distance between the antenna source and the
observation point. For the far field where r  R2 , the directivity will be the same for r [4]

S(θ, φ)max
D= , (2.5)
< Sθ,φ >
Pt
< Sθ,φ >= (2.6)
4πr 2
with Pt as transmit power of the source or generator and Pr as radiated power from the
antenna, S(θ, φ)max as the maximum power density, and < Sθ,φ > as power density of the
isotropic radiator. Therefore the directivity of the antenna will be

S(θ, φ)max
D = 4πr 2 . (2.7)
Pr
The above relations do not consider a) reflexion caused by possible non-perfect matching
of the source, tapered waveguides and antenna elements, b) conductor and c) dielectric losses.
These effects can be generally taken into account by an efficiency coefficient η representing
the ratio of the effectively radiated power Pr and the source generator power Pt including
the power losses Pl , in order to calculate the antenna gain G

Pr Pr
η= = , (2.8)
Pt Pr + Pl
S(θ, φ)max
G = η D = 4πr 2 . (2.9)
Pt
The antenna gain is normally expressed in dBi with respect to the isotropic radiator

gdBi = 10 · log10 (G). (2.10)

2.1 Isotropic Radiator 11

Fig. 2.1 Often used antennas

Table 2.1 Typical antenna properties

Type Polarization Gain
Isotropic antenna 0 dBi
Dipole antenna Linear 1.5–2.3 dBi
Patch antenna Linear, circular 7–10 dBi
Horn antenna Linear 20 dBi
Helix antenna Circular 10–18 dBi
Parabolic antenna or Linear, circular 20–70 dBi
Cassegrain antenna

In Fig. 2.1 and Table 2.1 the radiation characteristics of typical or often-used antennas,
which will be utilized in this book for the discussion of antenna array beam forming, are
summarized.

2.1 Isotropic Radiator

An isotropic radiator is a theoretical or hypothetical antenna which radiates homogeneously

in all directions. The magnitude of the Poynting vector or power density S at a certain
distance r is constant
 
 S(r , θ, φ) 
  = 1. (2.11)
 S 
max
12 2 Basic Antenna Elements

The isotropic radiator is used as a reference to compare the directivities of different other
antennas. The enhanced directivity in a certain direction, in comparison with the isotropic
radiator, is also called antenna gain G > 1 (obviously the gain of the isotropic radiator is
then G = 1 or g = 0 dBi). In terms of logarithmic dB values, the gain of a generally directive
antennas will be g > 0 dBi, whereas the gain of the isotropic antenna will be g = 0 dBi.

2.2 Dipole Antenna

A perfect dipole antenna is an infinitesimally thin wire (diameter d  λ) along the z-axis,
positioned at the arbitrary point (r  , θ, φ) of the spherical coordinate system, for example
at the center of the coordinate system (Fig. 2.2). The surface integral of the vector potential
(2.1) becomes a line integral with a source current I = I0 · uz (see for example [3, 4])
   l/2 
e−jk|r−r | e−jk|r−r |
A(r) = μI(r ) dl 
= μI(r 
) dl 
l 4π | r − r | −l/2 4π | r − r |
  l/2 (2.12)
e−jk|r−r | μI0 l 
= uz μI0 
dl  = uz 
e−jk|r−r | ,
4π | r − r | −l/2 4π | r − r |

μI0 l 
Ar = A z cos θ = e−jk|r−r | cos θ, (2.13)
4π | r − r |
μI0 l 
Aθ = −A z sin θ = 
e−jk|r−r | sin θ, (2.14)
4π | r − r |

Aφ = 0. (2.15)
By using the Eqs. (2.1)–(2.3) the electric field and magnetic field components can be
generally calculated

Hr = Hθ = 0, (2.16)

 
k I0 l sin θ 1 
Hφ = · 1 + · e−jk|r−r | , (2.17)
4π | r − r | jk | r − r |
 
η I0 l cos θ 1 
Er = 
· 1 + 
· e−jk|r−r | , (2.18)
4π | r − r | jk | r − r |
 
k I0 l sin θ 1 1 
Eθ =j 
· 1+ 
− 2 
· e−jk|r−r | , (2.19)
4π | r − r | jk | r − r | k | r − r | 2

E φ = 0. (2.20)
2.2 Dipole Antenna 13

Fig. 2.2 Dipole antenna

coordinate system

Depending on the distances between the antenna and the observation point d = | r − r |
and depending on the ratio k · d, the electric field and magnetic field components behave
differently in reactive near field k · d  1, Fresnel region or radiating near field k · d > 1,
Fraunhofer region or far field region k · d  1, since the second and third term of the
equations will either dominate or vanish.
In Fig. 2.3 the radiation characteristic is shown, with the dipole oriented in z-direction.
Figures 2.4 and 2.5 show the so-called E-plane and H-plane characteristics. E-plane is the
plane where the magnetic field component H is zero, whereas H-plane is the plane where
the electric field E is zero.

Fig. 2.3 Dipole antenna

radiation characteristics
14 2 Basic Antenna Elements

Fig. 2.4 Dipole antenna

E-plane Eθ (θ) at φ = 0◦

Fig. 2.5 Dipole antenna

H-plane Eθ (φ) at θ = 90◦ or
elevation angle 0◦
2.3 Patch Antenna 15

2.3 Patch Antenna

For a microstrip patch antenna with the effective length L e , width w, height or distance h
between the microstrip patch and the groundplane, the radiated fields can be calculated in
E-plane by
    
k0 w h E 0 e−jk0 |r−r | sin( k02 h cos φ) k0 L e
E φ (φ) = j · · cos sin φ (2.21)
π | r − r | k0 h
cos φ 2
2

and in H-plane by
 
 sin k0 h k0 w 
2 sin θ 2 cos θ

k0 w h E 0 e−jk0 |r−r | sin
E φ (θ ) = j · · sin(θ ). (2.22)
π | r − r | k0 h k0 w
2 sin θ 2 cos θ

Since the coordinate system has to be defined exactly before deriving the Maxwell’s
equations, the coordinate systems for the microstrip patch antenna are defined in Fig. 2.6,
similarly like in [3] (Figs. 2.7, 2.8 and 2.9).

Fig. 2.6 Patch antenna coordinate system

16 2 Basic Antenna Elements

Fig. 2.7 Microstrip Patch antenna radiation characteristic

Fig. 2.8 Microstrip patch antenna E-plane Eθ (φ) at θ = 90◦

2.4 Horn Antenna 17

Fig. 2.9 Microstrip patch antenna H-plane Eθ (θ) at θ = 0◦

2.4 Horn Antenna

Even though the dipole antenna elements and microstrip patch antenna elements are mainly
used for the array discussions in this book, we would like to mention a horn antenna as
another important antenna type which can be used directly as a radiator or as an antenna
feed in combination with a reflector antenna or a Cassegrain antenna (see for example [3]).
Satellite TV receivers typically use parabolic reflector antennas.
Also reflector antennas are applied for space telecommunications. Especially by using a
large parabolic aperture as a reflector, an extremely large antenna gain up to 70 dBi can be
achieved, improving the horn antenna gain of 20 dBi by about 50 dB. This is very important
for deep space telecommunication applications, where the radio frequency signal is strongly
attenuated during the transmission over some astronomic units. 1 AU (astronomic unit) is
about 150,000,000 km.
Radiation characteristic, antenna directivity or the antenna gain depend on the geometrical
dimensions of the horn antenna. A typical radiation pattern is shown in Fig. 2.10 calculated
by using the 3-dimensional numerical simulation system CST Studio Suite [6] based on
the so-called FIT (Finite Integration Technique) developed by Prof. Thomas Weiland in
Darmstadt.
In Fig. 2.11 the results can be calculated by using a Matlab algorithm [7] showing high
antenna gain in boresight direction. In this example, the diameter of the parabolic reflector
18 2 Basic Antenna Elements

Fig. 2.10 Horn antenna radiation characteristic

Fig. 2.11 Reflector antenna radiation characteristic

References 19

is designed to be 1.2 m, approximately 10 λ. By using a larger diameter for the parabolic

reflector, up to 35 m or even larger, used in space telecommunications, the antenna gain
can be further increased. This is also indispensable due to the large distance between the
spacecraft and the Earth groundstation and the large attenuation.

References

1. R. E. Collin, F. J. Zucker: Antenna Theory, Part 1. McGraw-Hill Book Company (1969).

2. H.-G. Unger: Planar Optical Waveguides and Fibres. Oxford Clarendon Press (1977).
3. C. A. Balanis: Antenna Theory. John Wiley & Sons, Inc. Fourth Edition (2016).
4. K. W. Kark: Antenna Antennas and Radiation Fields (in German: Antennen und Strahlungsfelder).
Springer Vieweg (2018).
5. S.-P. Chen: Fundamentals of Information and Communication Technologies. Cambridge Scholars
Publishing (2020).
6. CST (Computer Simulation Technology): https://www.3ds.com/de/produkte-und-services/
simulia/produkte/cst-studio-suite/student-edition/.
7. Matlab Antenna Toolbox. https://www.mathworks.com/products/matlab.html.
Linear Antenna Array
3

This chapter starts with an introduction into phased array antennas, discusses required phase
shifters and then presents simulation and measurement results of one-dimensional, linear
antenna array characteristics. The measurement results are compared with simulation results.
The simulation and measurement scenario consists of a linear 1 × 8 patch antenna array
as transmitter and an identical, but single patch antenna as receiver. The idea is to validate
the simulation results obtained by an analytical method both for the far field region and the
near field. For this purpose, the amplitude weights and the phase shifts of the individual
patches are adjusted by a control matrix, taking into account the calibrated antenna coaxial
cables. Different beam steering angles have been studied, i.e. ±0◦ , ±15◦ , and ±30◦ with
respect to the forward direction, with homogeneous, binomial, or Chebyshev weights. The
measurements validate and confirm the simulation results of the analytical method.
In this chapter we would like to present both simulation [1, 2] and measurement results of
near field and far field characteristics of a linear microstrip patch antenna array designed for
the ISM frequency of 5.8 GHz with the free space wavelength λ = 0.052 m. Antenna arrays
are widely used in various applications such as mobile communications, synthetic aperture
radar, medicine, sensing, imaging, or radio astronomy [3–13] to enable fast and precise beam
forming. In some cases the systems operate in the near field region. For investigating the far
field radiation characteristics and beam forming, far field approximation is accurate enough
for antenna arrays [3–8]. The coupling effect between the array elements can be neglected
as the distances between the array elements are λ/2 and radiation is low into the direction
of the adjacent antennas. Far field and near field are commonly distinguished by the far field
distance definition
2D 2
dR = , (3.1)
λ
also called Rayleigh distance d R , with the largest dimension of the antenna array D and the
wavelength λ. For the investigations of generally conformal antenna arrays, the linear planar
array will be the best starting point. The licensefree ISM frequency f = 5.8 GHz is chosen

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 21

S.-P. Chen and H. Schmiedel, RF Antenna Beam Forming,
https://doi.org/10.1007/978-3-031-21765-4_3
22 3 Linear Antenna Array

both for the simulations and the measurements. 8 microstrip patch antenna elements with
vertical polarization are used as a linear array for the measurements at the transmit side. On
the other side one patch antenna element is used as receiver antenna. The amplitude weights
and phase shifts for the single patch antenna elements are adjusted by using a control matrix
[14]. With the number of the antenna array elements M = 8 and the distance between the
patch antenna elements dx = λ/2, the near field range (distance between the center point
of the array and the observation point) remains shorter than the Rayleigh distance 24.5 λ or
1.274 m. At the distance of 1.8 m or 35 λ we are in far field distance larger than the Rayleigh
distance, whereas at 5.6 λ or 10 λ we will be in the near field.

3.1 Introduction into Phased Array Antennas

Antenna arrays consist of many basic antenna elements, as discussed in Chap. 2. To achieve
beam steering and beam forming, these antenna elements are controlled in phase and ampli-
tude. To explain beam steering a simple arrangement is shown for far field in Fig. 3.1 and
for near field in Fig. 3.2.

Fig. 3.1 Simple phased array antenna in far field

3.1 Introduction into Phased Array Antennas 23

We define an x–y coordinate system, as shown in Fig. 3.1. The different antenna elements
1–8 are on the x axis, at locations x1 − x8 respectively. These elements are to be controlled
in phase, so that a main beam consisting of parallel beams is generated. The direction of the
main beam to be generated is into the φ0 angle, measured from the y axis. The goal is, that the
electric fields from all antenna elements in a far field point, will have identical phase, so that
the superposition of all electric fields will be maximum, to achieve high gain. From Fig. 3.1
we see, that the signal from source 2, at x2 , obviously is earlier by Δ/v = (Δd1 − Δd2 )/v
compared to source 1, at x1 , where v is the wave velocity. Δ is the path difference between
the neighboring antenna elements. For example, the signal from source 5, at x5 , is earlier by
4Δ/v with respect to source 1 respectively. These differences can be expressed in phase Δφ
also (one time period T equals 360◦ in degrees or 2π in radian). To achieve identical phases
in the far field at the far field focal point, the time differences 0, Δ/v …7Δ/v need to be
compensated for, if the source 1 is taken as reference. With increasing index i the sources
are more close to the point in the far field, so that their signals must be delayed respectively.
This can either be done by adjusting proper individual time delays, or, for one frequency
or a relatively small frequency band, by adjusting the phases of the sources, where delay is
represented by a negative phase.
Trigonometry tells us, that
sin(−φ0 ) = Δ/dx , (3.2)
where dx is the spacing between the center points of the antenna elements. Typically dx
is chosen to be dx = λ0 /2 (Remark: For dx > λ0 /2 the Shannon theorem of the Fourier
Transform is violated, cf. Sect. 3.3.), thus

Δ = λ0 /2 · sin(−φ0 ). (3.3)

Generally the reference phase could be chosen arbitrarily. The most deciding factor is to
guarantee the phase differences between the antenna elements.
Let us assume, that reference phase φr e f = 0◦ at the center point of the 1 × 8 antenna
array. The phases for the other elements, with the index i, can then be calculated as
Δdi
Δφi = (i − 9/2) · · 360◦ . (3.4)
λ0
or
Δφi = (i − 4.5) · sin(−φ0 ) · 180◦ . (3.5)
Let us consider a simple example, where the wanted steering angle be φ0 = 30◦ , and the
spacings between the centers of the antenna elements be λ0 /2. The required phases for the
antenna elements are shown in Table 3.1. For an opposite steering angle φ0 , all phases will
have the opposite sign respectively.
In case of the focus point in the near field, generally known as near field focusing (Fig. 3.2),
with a beam steering angle φ0 , the above-mentioned far field approximation will not be
applicable, so that the distances Δdi (i = 1, 2, 3, ..., 7, 8) or phase shifts Δφi must be
24 3 Linear Antenna Array

Tab. 3.1 Required phase shifts for a beam steering angle of 30◦ with the focal point in far field
Antenna element Phase shift in radian Phase shift
1 5.512 315◦
2 3.934 225◦
3 2.359 135◦
4 0.786 45◦
5 −0.785 −45◦
6 −2.354 −135◦
7 −3.920 −225◦
8 −5.483 −315◦

Fig. 3.2 Simple phased array antenna in near field

calculated individually and considered for the phase differences which define the beam
steering (Table 3.2).
Correspondingly the phases of the antenna elements would be chosen, in order to focus
the beam at the desired near field focal point, for example at 10 λ with a beam steering angle
of 30◦ . The required phase shifts are shown in Table 3.2.
3.2 Phase Control 25

Tab. 3.2 Required phase shifts for a beam steering angle of 30◦ with the focal point in near field at
d = 10 λ
Antenna element Phase shift in radian Phase shift
1 6.158 353◦
2 4.273 245◦
3 2.484 142◦
4 0.800 46◦
5 −0.770 −44◦
6 −2.219 −127◦
7 −3.526 −203◦
8 −4.712 −270◦

3.2 Phase Control

For beam steering, as mentioned in previous sections, it is important to properly control

the phases of all antenna elements involved. For beam forming, additionally, all amplitudes
need to be controlled. Phase control means to adjust the phases of all antenna elements such
that the electromagnetic waves emanating from all antenna elements will arrive coherently
at a defined focal point in space.
The same discussion is true for the reverse consideration, that is to receive a signal which
origins in that point in space. Since reprocity is true for transmit and receive antennas, the
same, individual phase shifts for all antenna elements will be needed for transmit and receive
applications. In most applications in communications there is a transmitter and a receiver.
For both modes all phases need to be controlled.
This can either be done in the radio frequency (RF) range with analog RF phase shifters
and attenuators to also control individually all amplitudes, or on the other hand this can
be achieved at an intermediate frequency (IF) with preferably digital technologies. We will
come back to the second option at IF after discussing problems in the RF range.
Typically RF-power amplifiers are needed to transmit RF signals. If there are many
antenna elements, each element may have its own medium or low power amplifier, depend-
ing on the number of antenna elements. Also passive power dividers, phase shifters and
attenuators may be used.
For receiving, also amplifiers will be needed. In this case these will be low noise ampli-
fiers. All these amplifiers inherently have phase and gain variations. Many of the critical
parameters are temperature dependent and may change over long term operation. Thus, the
antenna elements, amplifiers, phase shifters, attenuators, and the RF distribution networks,
need to be carefully calibrated for phased array applications.
Typically all RF communications signals are down converted anyway, either to a lower
intermediate frequency (IF) first, or right down to the baseband. The inverse process is of
26 3 Linear Antenna Array

Fig. 3.3 Mixer with RF, LO

and IF signals

course required for up converting signals from baseband to the RF range. Down converting
a signal from the RF range to the IF range is accomplished by a so called mixer and a stable
local oscillator (LO) signal. Ideally this mixer can be considered as a multiplication device
which multiplies the RF signal with a LO signal. Figure 3.3 shows such a mixer.
On the left hand input side, we feed a RF signal into the mixer. This RF signal has a
certain amplitude A R F and a certain phase φ R F . Its angular frequency is ω R F . This is of
course the description for a single frequency. For a complete frequency band, or a spectrum,
this spectrum would be the superposition of many of those frequency lines.

R F = A R F · cos(ω R F t + φ R F ). (3.6)

This signal, called RF, is then multiplied with a LO signal. The mixer is assumed as an
ideal multiplication stage. The LO signal is

L O = 1 · cos(ω L O t + φ L O ) (3.7)

where the amplitude is normalized to 1. The amplitude is unimportant, as a mixer typically

is a switched-type mixer, where the LO switches diodes or transistors on and off, with
low amplitude noise. Also the LO phase φ L O can be set to 0. It is however important, to
understand, that all antenna elements, in the RF range, have mixers which down convert
the individual RF signals into individual IF signals. Of course all mixers require a coherent,
identical LO signal in phase. Thus we obtain, for a single antenna element, the IF signal as
I F = R F · L O, that is

I F = R F · L O = A R F · cos(ω R F t + φ R F ) · cos(ω L O t). (3.8)

or

I F = 0.5 · A R F · [cos((ω R F + ω L O )t + φ R F ) + · cos((ω R F − ω L O )t + φ R F )]. (3.9)

After low-pass filtering, we remove the higher spectral components

3.3 Measurement Setup and Verification 27

cos((ω R F + ω L O )t + φ R F ) (3.10)

and obtain the down converted IF signal as

I F = 0.5 · A R F · cos((ω R F − ω L O )t + φ R F ). (3.11)

We see, that the original amplitude A R F and the original phase φ R F have been conserved.
This is also true for a complete spectrum, as has been mentioned. That important feature
means, that we can do the phase and amplitude control in the IF range. The IF signal is an
identical replica of the RF signal. This offers a lot of flexibility, since phase and amplitude
control, can now be done digitally. Typical digital systems are less expensive than their analog
counterparts and offer unlimited long term stability. Still the mixers and all amplifiers, if
required, will have to be calibrated and checked for long-term effects.
Above, we considered the receive case, where a RF signal is down converted to a lower IF,
where phase and amplitude control are accomplished. However a mixer can also be operated
to up convert an IF signal to a RF signal at a higher frequency. In this case all phase and
amplitude control can be accomplished also in the IF range. Again each transmit antenna
element will require its own mixer and the coherent LO signal for up converting.
It should be mentioned at this point, that all the practical measurements in this book have
been done with phase and amplitude control in the RF range with the help of a special RF
matrix “MiCable Control Matrix” [14]. This calibrated phase and amplitude matrix can be
controlled by the help of an external computer.

3.3 Measurement Setup and Verification

Ideally theory, simulations and practical measurements of a specific antenna type should
give identical results. “In theory there is no difference between theory and practice—in
practice there is.” (Richard P. Feynman used this sentence). Theory and simulations can
only be as good as the models that are being used. Often a model does not consider all the
effects existing in reality or practice. For example in our case, the coupling effects between
adjacent antenna elements is not considered in the model being used for the simulation. Also
in general, sometimes there are unforeseen or undetected flaws in the theory calculation
process or in the simulations. This is not to blame theory or simulations, but mistakes
occur… Anyway, the ultimate benchmark is practical tests and measurements. These, of
course, have their own host of problems and side effects, as we may see below.
To verify our simulation results of phased-array antenna radiation patterns we did our
practical measurements in the microwave student lab of the Department of Electrical
Engineering of University of Applied Sciences, Hochschule Darmstadt, Darmstadt, Ger-
many. This relatively simple equipment setup is sufficient to verify all our simulations
and findings for various phased-array antenna configurations. This should encourage col-
leagues, engineers and students to perform their own simple, inexpensive radiation pattern
28 3 Linear Antenna Array

experiments, even as they may not have access to a sophisticated, professional anechoic
antenna measurement chamber.
Next follows a more detailed description of the experimental setup for the measurement
of a radiation pattern of a phased-array antenna system (Fig. 3.6). A medium power signal
generator SMP22 (Rohde & Schwarz) is tuned to 5.8 GHz and the output power set up to
+20 dBm. The signal generator is pulsed in “Pulse Mode” (more, see below).
The output of the signal generator is then fed into the Phase-Amplitude Matrix (MiCable
Control Matrix NT-VPAM-1x8-5.8) [14] which has 8 output channels. Each of these channels
can be adjusted individually in phase and amplitude by an external computer.
Because of the passive power division and the losses of the internal phase shifters and
attenuators, the basic loss of the matrix is some 25 dB. These 8 outputs are then connected
by 8 coaxial cables with “equal length” (more on equal length below) to the 8 patch antennas
of the phased-array antenna setup. This is our transmitter setup.
Different antenna setups are being used, linear, convex or concave oriented patch anten-
nas. The spacing between adjacent patch antenna elements is λ0 /2. The individual antenna
elements are linearly, vertically polarized patch antennas. The reflection (S11) of a single
antenna element is better than −15 dB. The coupling (S21) between two adjacent patch
antenna elements is <−15 dB. Ideally there should be no coupling between antenna ele-
ments, as this coupling is not considered in the simulation that was used. However for
practical purposes reflection and decoupling proof to be sufficient.
The above mentioned coaxial feeder cables, specified to have an equal length of 2000 mm
mm, have a severe impact on phase. A difference of some 1 mm in length gives a phase shift
of some 10◦ . Thus the phase shift of each individual cable needs to be measured beforehand
and must afterwards, during the measurements being compensated by adjusting the phase-
amplitude matrix properly, to minimize phase errors. Also all cables should be bent in similar
shape, as the phase due to a cable changes by some degrees, when the cable is bent when
the antenna setup is rotated on the turntable.
The antenna array, to be measured, sits on the turntable of an antenna measurement
setup. This is the Leybold/Cassy system originally designed for student experiments. This
Leybold/Cassy system consists of the turntable controlled by the Cassy software running on
an external computer. The rotating antenna, in our case the phased array antenna system, is
the transmit antenna in our measurement setup. The Leybold system provides a pulsed signal
(kHz range) to operate a synchronized sensitive DC-chopper amplifier on the receive side
to amplify a DC signal from the detector diode. Thus the above mentioned signal generator
needs to be pulse-modulated with this pulse signal delivered by the Leybold system.
On the receive side there is a single patch antenna element used as receiving antenna, a
similar patch as being used in the transmit array configuration. It is also a linearly, vertically
polarized microstrip patch antenna. To improve the dynamic range of the receive side, which
is restricted by noise, the sensitivity is increased with a 5.8 GHz band-pass filter connected
to the receive antenna and followed by a low noise amplifier (noise figure NF ≈ 3 dB, gain
g ≈ 30 dB). For far field measurements (d > 35λ), a parabolic reflector is used additionally
3.3 Measurement Setup and Verification 29

to increase sensitivity. At the output of the amplifier is a detector diode which has a dynamic
range from +3 dBm to some −50 dBm. The DC video output goes to the Leybold system and
is chopper amplified internally, then processed and displayed by the Cassy system versus
rotation angle of the turntable.
The measurement of low side lobes, in a radiation pattern measurement, is always critical
because of unwanted reflections from near-by metallic devices, as are shelves, test equipment
and so on. To reduce these effects, anti-reflection absorber mats are being used effectively.
Also the parabolic reflector on the receive side, during the far-field measurements, helps to
reduce incoming unwanted, arbitrarily reflected signals from the sides.
All measurements were performed and plotted with this simple experimental setup with
high repeatability. These measurements, which have to cope with non-ideal conditions,
however nicely confirm the simulations that assume ideal conditions.
A linear patch antenna array is shown in Figs. 3.4 and 3.5, where the patch antenna
elements are allocated along the x-axis and the polarization is in z direction. For the mea-
surements the center of the transmitter patch antenna array is the coordinate origin (0, 0, 0).
The patch antenna array, connected to the control matrix, is mounted on top of a rotating
table which is controlled by the antenna measurement software of Leybold (Fig. 3.6). The
coordinate origin and the focal point F(x F , y F ) define the wanted azimuth beam steering
angle φ0 . The receiver patch antenna is positioned exactly in this focal point F(x F , y F ) for
the steering angle φ0 .
The vector dxi is pointing to the i-th array element from the coordinate origin (0, 0, 0)
which is the reference point of the antenna array, whereas d is the vector pointing to the

Fig. 3.4 Patch array with characteristic parameters

30 3 Linear Antenna Array

Fig. 3.5 5.8 GHz 1 × 8 patch antenna array transmitter and the single patch antenna receiver

Fig. 3.6 Schematic diagram of the measurement setup

focal point from the coordinate origin. Then the vector di = d − dxi can be defined between
each array element and any arbitrary observation point, or focal point, at for instance d = 10
λ (Fig. 3.4). To be more precise, the beam steering angle is the angle with respect to the main
lobe in y-direction. Neglecting the coupling effect between the array elements, the radiated
electric fields of single patch antenna array elements can be superimposed [1, 2]:
3.4 Far Field Characteristics of a Linear Patch Antenna Array 31


M
E(d, φ) = Ai · Ei (d − dxi , φ) · exp(jαi ) (3.12)
i=1

with Ai as the amplitude and αi as the properly chosen phase shift of the i-th patch antenna
element. The total phase shifts αi must include all the phase shifts φi configured in the
control matrix and possible phase variations δi of each corresponding connecting cable. For
this purpose the phase variations of all connecting cables must be measured precisely in
advance. Ei is the azimuth radiation characteristic of the single patch antenna.
Characteristics of all simulation results achieved by using our own Matlab codes and
measurement results in this book are shown with identical scales of −30 dB to 0 dB. Mea-
surement plots are genuine plots (with tiny number fonts, but identical scales).

3.4 Far Field Characteristics of a Linear Patch Antenna Array

In Figs. 3.7, 3.8, 3.9, 3.10, 3.11, 3.12, 3.13, 3.14 and 3.15 the results of the analytical
methods [1, 2] are compared pairwise with the measurement results. The logarithmic scales
for both graphs are identical. It can be clearly seen that the relatively high side lobes in
case of homogeneous amplitude weights can be significantly reduced by using binomial and
Chebyshev amplitude weights. In case of binomial amplitude weights, the beam width is
also remarkably broadened. On the other hand, narrower beams can be achieved by using a
larger number of antenna elements.

Fig. 3.7 Far field characteristics of a patch antenna array, d = 35 λ, homogeneous amplitude weights,
beam steering 0◦ . a Simulation; b measurement
32 3 Linear Antenna Array

Fig. 3.8 Far field characteristics of a patch antenna array, d = 35 λ, homogeneous amplitude weights,
beam steering 15◦ . a Simulation; b measurement

Fig. 3.9 Far field radiation characteristics of a patch antenna array, d = 35 λ, homogeneous amplitude
weights, beam steering 30◦ . a Simulation; b measurement
3.4 Far Field Characteristics of a Linear Patch Antenna Array 33

Fig. 3.10 Far field characteristics of a patch antenna array, d = 35 λ, binomial amplitude weights,
beam steering 0◦ . a Simulation; b measurement

Fig. 3.11 Far field characteristics of a patch antenna array, d = 35 λ, binomial amplitude weights,
beam steering 15◦ . a Simulation; b measurement
34 3 Linear Antenna Array

Fig. 3.12 Far field radiation characteristics of a patch antenna array, d = 35 λ, binomial amplitude
weights, beam steering 30◦ . a Simulation; b measurement

Fig. 3.13 Far field characteristics of a patch antenna array, d = 35 λ, Chebyshev amplitude weights,
beam steering 0◦ . a Simulation; b measurement
3.4 Far Field Characteristics of a Linear Patch Antenna Array 35

Fig. 3.14 Far field characteristics of a patch antenna array, d = 35 λ, Chebyshev amplitude weights,
beam steering 15◦ . a Simulation; b measurement

Fig. 3.15 Far field radiation characteristics of a patch antenna array, d = 35 λ, Chebyshev amplitude
weights, beam steering 30◦ . a Simulation; b measurement
36 3 Linear Antenna Array

We see, that there is a nice agreement between simulation and measurement results,
qualitatively and in quantity. That are shapes of the main beam, half-power beam widths,
steering angles, shape of the side lobes and nulls. So all simulations are verified by the
measurements. Looking into the details, we will see, that the half-power beam widths of
the measurements seem to be a little larger than those of the simulations. This may be
attributed to phase errors in the experiments. These errors are estimated to be less than
±10◦ . (Remark: For typical beam widths an increase in half-power beam width of 1◦ is
related to a gain reduction of some 0.1 dB.) As results, we see that the linear patch array
antenna has a half-power beam width of 13◦ –15◦ for steering angles 0◦ –30◦ . With increasing
steering angle the half-power beam width increases. This is expected, as the array broadside,
as seen from the far point at the receive antenna, gets a little smaller for a higher steering
angle. The side lobes for the linear patch antenna array, with homogeneous amplitudes, are
relatively high which is expected from Fourier Transform considerations, that is (sin x)/x,
cf. Sect. 3.5. For a steering angle of 0◦ the first side lobe is −14 dB compared to the main
lobe. For higher steering angles, the side lobes increase a little, i.e. −12 dB for 30◦ steering
angle. To effectively reduce these side lobes, amplitude weights will be introduced in the
next chapter.

3.5 Amplitude Weights and Side Lobes of the Radiation Pattern

By choosing proper amplitude weights and phases for the antenna elements beamforming
can be achieved.
Considering Eq. (2.12) for dipole source current I and the amplitude weights Ai in case of
discrete antenna array, we see that the antenna radiation pattern relates to the current density
space distribution by the Fourier Transform or Discrete Fourier Transform [15–17]. In other
words, if we know the current density space distribution, we will obtain the radiation pattern
in the far field by Fourier Transform. Vice versa we will get the current distribution, both in
amplitude and phase, by Inverse Fourier Transform from a given radiation pattern.
This simple relationship allows us some helpful insights into antenna radiation patterns,
by using simple, known Fourier Transform relationships, i.e.

• Uniform current density space distribution, or uniform illumination, is transformed into

a radiation pattern with sin(x)/x shape, resulting in strong antenna sidelobes. (Remark:
in our simulated and measured examples, all antenna elements are driven with the same
amplitude (Figs. 3.7–3.9). Observe high sidelobes and also high gain).
• Gaussian illumination will be Fourier transformed into a Gaussian shaped beam pattern
with low sidelobes as is typically preferred. (Remark: in our simulated and measured
examples all the amplitudes of the different antenna elements are not Gaussian weighted,
but binomial weighted or Chebyshev weighted).
3.6 Near Field Characteristics at a Distance of 10 λ 37

• A small antenna aperture will generate a wide antenna pattern with relatively low gain,
whereas a large antenna aperture will generate a narrow antenna beam pattern with
relatively high gain.
• When an antenna beam is steered from a linear antenna array, as in our examples, it is
easy to see, that the beamwidth will broaden, when the beam is steered from the main
direction to a side. This is because the antenna array aperture seen from a point in the
far field is smaller compared to the aperture seen in the main direction. Consequently a
broader beam width is related to a smaller antenna gain.

The similar effects can be observed for all antenna elements or antenna arrays, like linear
antenna arrays, planar antenna arrays or conformal antenna arrays which will be discussed
in the following chapters.
By using Inverse Fourier Transform, or Inverse Fast Fourier Transform with windowing
[16], wanted, also arbitrary, antenna beam patterns can be synthesized. For more specific
details, see [16, 17].
Two different amplitude weighting methods have been examined, that is binomial ampli-
tude distribution, which is close to a pseudo-Gaussian distribution, and on the other hand
a Chebyshev distribution. The phases were not changed, compared to the homogeneous
amplitude distribution from Sect. 3.2. A binomial amplitude weighting reduces the side
lobes significantly to be less than −30 dB compared to the main lobe, cf. Figs. 3.10–3.12. As
a tradeoff the half-power beam width increases from some 14◦ , for homogeneous amplitudes,
to some 23◦ for binomial amplitudes. Consequently the gain will be reduced. (Remark: The
gain reduction may be aggravated, if the total feeding power for a transmit antenna is reduced
because of lossy RF attenuators that may be used to adjust the amplitudes for antenna array
elements at the sides of the array. This is another reason for amplitude adjustment in IF
stages.) Nevertheless a binomial amplitude weighting may be of interest, if extremely low
side lobes are the design goal.
Next Chebyshev amplitude weights are examined, see Figs. 3.13–3.15. We see, that the
side lobes are also down by −30 dB. The half-power beam widths are between 17◦ for a
steering angle of 0◦ , and 19◦ for a steering angle of 30◦ respectively. So obviously Chebyshev
amplitude weights are a nice compromise with small half-power beam width and at the same
time low side lobes. (Remark: to fully exploit the relatively high gain, the amplitudes are to
be set with no RF attenuators. That is either low-loss power dividers or amplitude control
in the IF stages.)

3.6 Near Field Characteristics at a Distance of 10 λ

In Figs. 3.16, 3.17, 3.18, 3.19, 3.20, 3.21, 3.22, 3.23 and 3.24 the near field characteristics
of the analytical methods [1, 2] are compared pairwise with the measurement results for a
very short focal point distance of 10 λ. Also for the near field it can be clearly seen that the
38 3 Linear Antenna Array

Fig. 3.16 Near field characteristics of a patch antenna array, d = 10 λ, homogeneous amplitude
weights, beam steering 0◦ . a Simulation; b measurement

Fig. 3.17 Near field characteristics of a patch antenna array, d = 10 λ, homogeneous amplitude
weights, beam steering −15◦ . a Simulation; b measurement
3.6 Near Field Characteristics at a Distance of 10 λ 39

Fig. 3.18 Near field radiation characteristics of a patch antenna array, d = 10 λ, homogeneous ampli-
tude weights, beam steering −30◦ . a Simulation; b measurement

Fig. 3.19 Near field characteristics of a patch antenna array, d = 10 λ, binomial amplitude weights,
beam steering 0◦ . a Simulation; b measurement
40 3 Linear Antenna Array

Fig. 3.20 Near field characteristics of a patch antenna array, d = 10 λ, binomial amplitude weights,
beam steering −15◦ . a Simulation; b measurement

Fig. 3.21 Near field radiation characteristics of a patch antenna array, d = 10 λ, binomial amplitude
weights, beam steering −30◦ . a Simulation; b measurement
3.6 Near Field Characteristics at a Distance of 10 λ 41

Fig. 3.22 Near field characteristics of a patch antenna array, d = 10 λ, Chebyshev amplitude weights,
beam steering 0◦ . a Simulation; b measurement

Fig. 3.23 Near field characteristics of a patch antenna array, d = 10 λ, Chebyshev amplitude weights,
beam steering −15◦ . a Simulation; b measurement
42 3 Linear Antenna Array

Fig. 3.24 Near field radiation characteristics of a patch antenna array, d = 10 λ, Chebyshev amplitude
weights, beam steering −30◦ . a Simulation; b measurement

relatively high side lobes, in case of homogeneous amplitude weights, can be significantly
reduced by using binomial amplitude weights and Chebyshev amplitude weights. Narrower
beams can be also achieved by using a larger number of antenna elements.

Fig. 3.25 Near field characteristics of a patch antenna array, d = 5.6 λ, homogeneous amplitude
weights, beam steering 0◦ . a Simulation; b measurement
3.6 Near Field Characteristics at a Distance of 10 λ 43

Fig. 3.26 Near field characteristics of a patch antenna array, d = 5.6 λ, homogeneous amplitude
weights, beam steering 15◦ . a Simulation; b measurement

Fig. 3.27 Near field radiation characteristics of a patch antenna array, d = 5.6 λ, homogeneous ampli-
tude weights, beam steering 30◦ . a Simulation; b measurement
44 3 Linear Antenna Array

3.7 Near Field Characteristics at a Distance of 5.6 λ

In Figs. 3.25, 3.26, 3.27, 3.28, 3.29, 3.30, 3.31, 3.32 and 3.33 the near field characteristics
of the analytical methods [1, 2] are compared pairwise with the measurement results for a
focal point distance of 5.6 λ. Also for the near field it can be clearly seen that the relatively
high side lobes in case of homogeneous amplitude weights can be significantly reduced by
using binomial amplitude weights and Chebyshev amplitude weights. Narrower beams can
be achieved by using a larger number of antenna elements.
Looking at all the Figs. 3.25–3.33 we see again, that the measurements validate the
simulation results nicely. Slightly higher side lobes in the measurements, compared to the
simulations may be the result of unwanted reflections within the experimental setup. For
these near field measurements no parabolic reflector was used, in order to reduce the receive
antenna aperture for more accurate measurements. Distances of 5.6 λ and 10 λ were chosen.
With proper phases, corrected for near field, and the same amplitude weights as used for far
field, we obtain very similar patterns as for the far field case. Beam widths and side lobe
distributions are almost identical. So near field applications can be designed easily with
wanted beam shapes and steering angles.
In this chapter far field and near field beam steering characteristics of one-dimensional,
linear microstrip patch antenna arrays were investigated. Simulation results by using an ana-
lytical method were compared with measurements by using an antenna measurement setup
for the ISM frequency of 5.8 GHz. Amplitude tapering with binomial amplitude weights and

Fig. 3.28 Near field characteristics of a patch antenna array, d = 5.6 λ, binomial amplitude weights,
beam steering 0◦ . a Simulation; b measurement
3.7 Near Field Characteristics at a Distance of 5.6 λ 45

Fig. 3.29 Near field characteristics of a patch antenna array, d = 5.6 λ, binomial amplitude weights,
beam steering 15◦ . a Simulation; b measurement

Fig. 3.30 Near field radiation characteristics of a patch antenna array, d = 5.6 λ, binomial amplitude
weights, beam steering 30◦ . a Simulation; b measurement
46 3 Linear Antenna Array

Fig. 3.31 Near field characteristics of a patch antenna array, d = 5.6 λ, Chebyshev amplitude weights,
beam steering 0◦ . a Simulation; b measurement

Fig. 3.32 Near field characteristics of a patch antenna array, d = 5.6 λ, Chebyshev amplitude weights,
beam steering 15◦ . a Simulation; b measurement
References 47

Fig. 3.33 Near field radiation characteristics of a patch antenna array, d = 5.6 λ, Chebyshev amplitude
weights, beam steering 30◦ . a Simulation; b measurement

Chebyshev amplitude weights efficiently reduces the side lobes over a larger beam steering
angle for the far field and near field radiation characteristics. The beamwidth increases with
binomial and Chebyshev amplitude weights and also increases with an increased beam steer-
ing angle. Depending on the beam forming requirement, in case of a wanted narrow beam,
this method can be combined with a larger number of patch antenna array elements. The
analytical and measurement methods presented in this chapter can be used to provide the
optimum parameters for dynamic beam forming problems for adaptive planar or conformal
antenna arrays.

References

1. S.-P. Chen: Improved Near Field Focusing of Antenna Arrays with Novel Weighting Coefficients.
IEEE WiVeC 2014, 6th International Symposium on Wireless Vehicular Communications (2014).
2. S.-P. Chen: An Efficient Method for Investigating Near Field Characteristics of Planar Antenna
Arrays. Wireless Personal Communications 95 (2), pp. 223–232. Springer Nature (2017).
3. R. E. Collin, F. J. Zucker: Antenna Theory, Part 1. McGraw-Hill Book Company (1969).
4. K. F. Lee: Principles of Antenna Theory. John Wiley & Sons Ltd. (1984).
5. B. D. Steinberg, H. M. Subbaram: Microwave Imaging Techniques. John Wiley & Sons, Inc.
(1991).
6. R. J. Mailloux: Phased Array Antenna Handbook. Artech House Inc. (1994).
7. L. V. Blake, M. W. Long: Antennas: Fundamentals, Design, Measurement. 3rd Edition. Scitech
Publishing, Inc. (2009).
8. D. G. Fang: Antenna Theory and Microstrip Antennas. CRC Press (2010).
48 3 Linear Antenna Array

9. W. H. Carter: On Refocusing a Radio Telescope to Image Sources in the Near Field of the
Antenna Array. IEEE Transactions on Antennas and Propagation, Vol. 37, pp. 314–319 (1999).
10. A. Badawi, A. Sebak, L. Shafai: Array Near Field Focusing. WESCNEX’97 Proceedings of
Conference on Communications, Power and Computing, pp. 242–245 (1997).
11. M. Bogosanovic, A. G. Williamsoni: Antenna Array with Beam Focused in Near Field Zone.
Electronics Letters, vol. 39, pp. 704–705 (2005).
12. J. Grubert: A Measurement Technique for Characterization of Vehicles in Wireless Communi-
cations. PhD Thesis (in German) of Technical University Hamburg-Harburg, Cuvillier Verlag
Goettingen (2006).
13. S. Ebadi, R. V. Gatti, L. Marcaccioli, R. Sorrentinoi: Near Field Focusing in Large Reflector
Array Antennas Using 1-bit Digital Phase Shifters. Proceedings of the 39th European Microwave
Conference, pp. 1029–1032 (2009).
14. MiCable Inc.: http://en.micable.cn.
15. C. Balanis, Antenna Theory Analysis and Design, John Wiley, 1997.
16. F. Harris, “On the Use of Windows for Harmonic Analysis with the Discrete Fourier Transform”,
Proc. of IEEE, vol. 66, No. 1, pp. 51–83, 1978.
17. X. Wang, Y. Zhong, and Y. Wang, “An Improved Antenna Array Pattern Synthesis Method Using
Fast Fourier Transform”, International Journal of Antennas and Propagation, 2015, 1–9, 2015.
Planar Antenna Arrays
4

Near field focusing of 2-dimensional planar antenna arrays is analyzed by using efficient
analytical methods, which were validated by comparing the simulation results with the
measurement results of one-dimensional linear patch antenna array in the last chapter, instead
of time consuming numerical methods. This is especially important for the design and
optimization process for dynamic beam forming and other near field applications.
Antenna arrays show increased side lobes at larger beam steering angles in the near field.
Conventional techniques such as inhomogeneous but symmetrical amplitude illuminations,
to assure a certain side lobe suppression level, do not reduce the side lobes completely
or sufficiently for the near field case, especially in the back fire direction. In this chapter,
asymmetrical amplitude weighting coefficients in combination with Dolph Chebyshev are
used to further improve the side lobe suppression over a larger angular range.
Antenna arrays are widely used in various applications such as mobile communications,
synthetic aperture radar, medicine, sensing, imaging or radio astronomy [1–11] to enable
fast and precise beam forming. Generally a high resolution is required for beam forming
which leads to relatively large antenna arrays and complex signal processing systems. In
some cases the systems should operate in the near field region.
The far field radiation characteristics and beam forming methods for antenna arrays have
been thoroughly investigated and discussed [1–6]. The coupling effect between the array
elements can often be neglected for the distances between the array elements larger than
or equal to λ/2, if radiation from an antenna element is mimimum into the direction of the
neighoring antenna elements.
Far field and near field are commonly distinguished by the far field distance definition
rmin = 2D 2 /λ with the largest dimension of the antenna array D and the wavelength λ.
For the investigations in this chapter, i.e. for the number of the antenna array elements
M, N ≤ 10 and dx , dz = λ/2 (Fig. 4.1–4.3), the near field range (distance between the
center point of the array and the observation point) remains shorter than 32λ, even though
the principles are valid for arbitrary antenna arrays with large numbers M and N [12].

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 49

S.-P. Chen and H. Schmiedel, RF Antenna Beam Forming,
https://doi.org/10.1007/978-3-031-21765-4_4
50 4 Planar Antenna Arrays

4.1 Beam Forming, Focusing and Steering of Planar Arrays

A planar antenna array is illustrated schematically in Fig. 4.1, where xm and z n are the x-
and z-coordinates of the single array elements. The center of the array and the focal point
F(x, y, z) form the beam steering angle θ0 or φ0 . r − rmn is the vector between the array
element (xm , ymn , z n ) and the focal point F(x, y, z). rmn is the vector pointing to the array
element from the coordinate origin (0, 0, 0) as reference point of the antenna array. To be
more precise, the beam steering angle is the angle with respect to the main lobe in y-direction.
Neglecting the coupling effect between the array elements, the planar array can be further
investigated by superimposing two orthogonal linear arrays parallel to x axis and z axis (see
for example Fig. 4.2 and 4.3).
For simplicity the planar antenna arrays are analyzed first, i.e. ymn = 0, even though the
proposed method is generally valid for conformal antenna arrays.
The near field characteristics of this type of plannar antenna arrays or modified conformal
antenna arrays can be analyzed either by using numerical methods or by using the analytical
method by superimposing all the radiation fields resulting from all the array elements located
along the x − z plane. The total radiation pattern of the arbitrary array at the focal point or
point of observation F can be obtained by


M 
N
E(r,θ .φ) = amn · Emn (r − rmn , θ, φ) · exp(jα mn ) (4.1)
m n

Fig. 4.1 Planar antenna M×N

array in the x-z plane with the
F(x, y. z)
4.1 Beam Forming, Focusing and Steering of Planar Arrays 51

Fig. 4.2 Linear array along

x-axis, side by side, with the
focal point F(x,y,z), projection
on the x-y plane for z-oriented
dipoles

Fig. 4.3 Linear array along

z-axis, side by side, with the
focal point F(x,y,z) projection
on the y-z plane for z-oriented
dipoles

with amn as amplitude, and αmn as properly chosen phase shift of the corresponding antenna
element located at (xm , 0, z n ) to control the desired beam steering. The elements can be
Hertzian dipoles, patches or other radiators. For focusing the radiation of the antenna array,
the side lobes should be reduced to a minimum. The desired narrow beam width can be
achieved by using a large number of array elements M × N .
For simplicity, and in case of a planar antenna array located in the x − z plane, ymn = 0 is
valid. Emn (r − rmn , θ, φ) is the basic function describing the field radiation characteristics
52 4 Planar Antenna Arrays

of each antenna element. r is the vector pointing at the focus F from the coordinate system
origin (0, 0, 0).
For the far field radiation characteristics investigations some simplifications are possible,
i.e. the E-fields are reciprocally proportional to the distance r between the source and the
observation point or focal point in the first approximation, but for the near field case either the
simplified isotropic radiator or the exact near field pattern of each single radiator (dipoles
or patches) located at (xm , 0, z n ) are to be considered. For realistic cases, especially by
investigating the planar antenna arrays, Hertzian dipoles or patch antenna elements along
the z-axis are investigated (cf. [6]).
For planar arrays the beam can be formed and steered by properly choosing the amplitudes
amn and phase shifts αmn in a one-dimensional manner. The 2-dimensional problem of beam
forming can be simplified, if we neglect the coupling effect between the array elements and
assume some amplitude and phase distributions which allow the separation of the double
sum into the product of simple sums equations (4.2) – (4.4):

E(r, θ. φ) = EM (r − rm , θ, φ) · EN (r − rn , θ, φ), (4.2)


M
EM (r, θ .φ) = am · Em (r − rm , θ, φ) · exp(jαx (m)), (4.3)
m


N
EN (r, θ. φ) = an · En (r − rn , θ, φ) · exp(jαz (n)). (4.4)
n
For this case the planar array can be simply replaced by two othogonal linear arrays
parallel to x− and z−axis separately. The analyses become much simpler, especially if the
focus is on the near field focusing.
By doing so, the beam forming tasks to achieve φ0 can also be done for instance by
properly choosing the amplitude paramenters am and phase shifts αm first in x-y plane
(Fig. 4.2) and then in x-z plane to achieve θ0 by properly choosing the amplitude coefficients
an and phase shifts αn (Fig. 4.3).
To steer the beam to a desired direction in the near-field, i.e. if the focus F is near the
sources, say in a distance < 30λ, the connecting lines between the focal point and the array
elements are no more parallel like in far field case, so that the easy far field model leads
to wrong estimation of the phase shifts of the antenna elements. The phases from the array
elements to the focal point F must be calculated by considering the exact distance between
the array element and the focal point, not only the steering angle like in the far field case
[10]: 
αx (m) = k0 Rm 2 + (md )2 − 2R · m · d · cos(φ ),
x m x 0 (4.5)

αz (n) = k0 Rn2 + (ndz )2 − 2Rn · n · dz · sin(θ0 ). (4.6)
4.2 Side Lobe Suppression 53

With the distances Rm , Rn and the steering angles θ0 , φ0 , the phase differences between
the elements, especially the difference with respect to the origin, can be simply calculated
as Δα = αm − k0 · r . Fig. 4.2 illustrates the relationship.
In all simulations in this chapter, the phase shifts will be based on this exact near field
assumption.
In order to simulate the near field focusing problems quickly, an analytical method is
prefered. To ensure this, the analytical method was compared with commonly known sophis-
cated, but time-consuming numerical simulation techniques like CST and HFSS in [12]. The
relatively small differences between these solutions confirm the low coupling effect for the
parameter settings also used in this chapter, and justify the usage of the proposed analytical
method to investigate the near field focusing problems as a good approximation, especially
for patch antennas.
In [10, 12] it could be shown that the side lobe level increases significantly in near
field, when the beam steering angle becomes larger, if homogeneous amplitude weights are
applied.

4.2 Side Lobe Suppression

Firstly a 10 dipole array is analyzed for different beam steering angles 15◦ – 45◦ with the
focal point designed to be at a distance of 10λ to show the typical radiation characteristics
with and without additional asymmetrical amplitude weights.
There are many well-known steering techniques for beam forming and side lobe sup-
pression, such as binomial and Chebyshev illumination coefficients, defined as weighting
function W (m) and W (n) (e. g. [4, 5]). By using the binomial distribution the side lobe
level can be reduced without ripple, but the beam width is also increased at the same time.
Chebyshev amplitude weights lead to better results both in terms of narrow beam width and
reduced side lobe level which are eminently important for near field focusing.
The side lobe suppression is improved by using Chebyshev illumination coefficients, but
the major side lobe increases for the beam steering angles of 15◦ to 45◦ [10, 12], especially
for near field cases.
The significant increase of the side lobes is disadvantageous for signal transmission,
sensing or imaging applications. These could lead to errors and interferences due to multipath
propagation effects. In [10] a simple but effective way was proposed to reduce these side
lobes further by using an additional asymmetrical amplitude illumination, which was also
further improved in [12] for θ0 and the array parallel for x-axis (see definitions in equations
(4.7)–(4.10), the same equations are valid for φ0 and the array parallel to z-axis)

(180 − θ0 )
ns = · M, (4.7)
180

aasym (i) = 1 + s · (i − 1); for 1 <= i <= n s , (4.8)

54 4 Planar Antenna Arrays

(M − i) · n s
− aasym (i) = 1 + s · ; for n s <= i <= M, (4.9)
(M − n s )

a(i) = W (i) · aasym (i) exp(j · α(i)). (4.10)

The results for a linear array in Fig. 4.4–4.6 show the improvement of beam focusing by
using additional amplitude weights more comprehensively.
Therefore in this chapter we use this method to optimize the side lobe suppression also
for the near field applications of the 2-dimensional planar antenna arrays. The asymmetrical
amplitude weights are used both parallel to x- and z-axis.
The amplitude weight a(i) is divided into conventional coefficients W (m) (either homo-
geneous, binomial or Chebyshev) and an additional asymmetrical weighting function
aasym defined in equations (4.7)–(4.10). The phase shift of the array element is given by
α(i) = αx (i) or α(i) = αz (i) depending on the allocation of the array elements. Adjusting
the phases at three different distances according to equation (4.5) leads to the patterns shown
in Fig. 4.5 and 4.6. Parameter s is an optimization parameter used individually to minimize
the side lobe level. Depending on the array size, distances between the array elements and
distances to the focal point, this parameter s can vary differently [10]. In Fig. 4.6 the parame-
ter s was varied to achieve different side lobe suppression effects. The parameter s is obtained

Fig. 4.4 Near field radiation pattern of a z-oriented dipole array along x-Axis, with homogeneous
amplitude weights, for a beam steering angle of 15◦ with the focal point at a distance of 10λ.
4.2 Side Lobe Suppression 55

Fig. 4.5 Near field radiation pattern of a z-oriented dipole array, with Dolph Chebyshev amplitude
weights, for a beam steering angle of 15◦ with the focal point at a distance of 10λ, θ = 90◦

Fig.4.6 Normalized radiation pattern of a z-oriented dipole array, with Chebyshev amplitude weights
and additional asymmetrical amplitude weights (s = 0.15–0.35), for beam steering angle of 15◦ ,
with the focal point at a distance of 10λ, θ = 90◦
56 4 Planar Antenna Arrays

empirically and iteratively. The results by using the simple asymmetrical amplitude weight
of equations (4.8) and equation (4.9) with s = 0.15 − 0.35 are shown in Fig. 4.6.
A further substantial reduction of the dominant side lobe at φ = 180◦ is obtained when
combining the Chebyshev weights with the proposed asymmetrical amplitude weights [12].
The results have shown that the asymmetrical weighting coefficients lead to a remarkable
further reduction of side lobe levels over a larger range. Especially the backward side lobes
(“back fire”) which can not be sufficiently reduced for near field beam steering, by only
using Chebyshev coefficients, can be minimized to a lower level.

4.3 Simulation Results for Planar Antenna Arrays

The 2-dimensional simulation results are displayed in Fig. 4.7, 4.8, 4.9 and 4.10. It is obvious
that the relatively high side lobes level of homogeneous amplitude weights can be reduced
almost completely by binomial amplitude weights, but the beam width is increased signif-
icantly, whereas by using Chebyshev with additional asymmetrical amplitude weights, the
beam width remains narrow and the side lobes are suppressed efficiently. By doing so, the
best near field focusing results can be achieved.

Fig. 4.7 Near field radiation pattern of a 10×10 Hertzian dipole array, with homogeneous amplitude
weights, for beam steering angles of 90◦ − θ0 = 30◦ and 90◦ − φ0 = 15◦ , with the focal point at a
distance of 10λ
4.3 Simulation Results for Planar Antenna Arrays 57

Fig. 4.8 Near field radiation pattern of a 10×10 Hertzian dipole array, with binomial amplitude
weights, for beam steering angles of 90◦ − θ0 = 30◦ and 90◦ − φ0 = 15◦ , with the focal point at a
distance of 10λ

Fig. 4.9 Normalized radiation pattern of a 10×10 Hertzian dipole array, with Dolph Chebyshev
amplitude weights, for beam steering angles of 90◦ − θ0 = 30◦ and 90◦ − φ0 = 15◦ , with the focal
point at a distance of 10λ
58 4 Planar Antenna Arrays

Fig. 4.10 Normalized radiation pattern of a 10×10 Hertzian dipole array, with Dolph Chebyshev
amplitude weights and additional asymmetrical amplitude weights (s = 0.15) for beam steering angles
of 90◦ − θ0 = 30◦ and 90◦ − φ0 = 15◦ , with the focal point at a distance of 10λ

In this chapter the beam steering characteristics of planar antenna arrays in the near-field
has been investigated by using a very efficient analytical method based on a new asymmet-
rical amplitude weighting function [12] in combination with Chebyshev coefficients.
Amplitude tapering by using the asymmetrical weighting coefficients reduces the side
lobes over a larger angular range efficiently. Depending on the beam forming requirement,
this method can be combined with the well-known far-field technique (binomial or Cheby-
shev) to achieve certain radiation characteristics. The parameter s can be determined indi-
vidually on different antenna array design parameters, such as the number of elements and
the near field focal point, and can be used to optimize the overall side lobe suppresion.
The method presented in this chapter can be used to quickly provide the optimum parame-
ters for dynamic beam forming problems for the adaptive planar or conformal antenna arrays.

References

1. R. E. Collin, F. J. Zucker: Antenna Theory, Part 1. McGraw-Hill Book Company (1969).

2. K. F. Lee: Principles of Antenna Theory. John Wiley & Sons Ltd. (1984).
3. B. D. Steinberg, H. M. Subbaram: Microwave Imaging Techniques. John Wiley & Sons, Inc.
(1991).
4. R. J. Mailloux: Phased Array Antenna Handbook. Artech House Inc. (1994).
References 59

5. L. V. Blake, M. W. Long: Antennas: Fundamentals, Design, Measurement. 3rd Edition. Scitech

Publishing, Inc. (2009).
6. D. G. Fang: Antenna Theory and Microstrip Antennas. CRC Press (2010).
7. W. H. Carter: On Refocusing a Radio Telescope to Image Sources in the Near Field of the
Antenna Array. IEEE Trans. Antennas Propagat., Vol. 37, pp. 314–319 (1999).
8. A. Badawi, A. Sebak, L. Shafai: Array Near Field Focusing. WESCNEX’97 Proceedings of
Conference on Communications, Power and Computing, pp. 242–245 (1997).
9. M. Bogosanovic, A. G. Williamsoni: Antenna Array with Beam Focused in Near Field Zone.
Electronics Letters, vol. 39, pp. 704–705 (2005).
10. J. Grubert: A Measurement Technique for Characterization of Vehicles in Wireless Communi-
cations. PhD Thesis (in German) of Technical University Hamburg-Harburg, Cuvillier Verlag
Goettingen (2006).
11. S. Ebadi, R. V. Gatti, L. Marcaccioli, R. Sorrentinoi: Near Field Focusing in Large Reflector
Array Antennas Using 1-bit Digital Phase Shifters. Proceedings of the 39th European Microwave
Conference, pp. 1029–1032 (2009).
12. S.-P. Chen: Improved Near Field Focusing of Antenna Arrays with Novel Weighting Coefficients.
IEEE WiVeC 2014, 6th International Symposium on Wireless Vehicular Communications (2014).
Conformal Antenna Array
5

The near field characteristics of conformal antenna arrays are analyzed by using a very
efficient analytical method, instead of time consuming numerical methods. This is especially
important for the design and optimization process for dynamic beam forming in many near
field applications. In order to optimize the near field characteristics, in terms of nominal
beam width changes and side lobe suppression, typical conformal antenna arrays which
are distributed on convex, concave and planar surfaces are compared with each other. This
efficient analytical method can be further developed for analyzing generally volumetric
antenna array problems.
Antenna arrays are widely used in various applications such as 5G mobile communica-
tions MIMO radio access networks, synthetic aperture radar, medicine, sensor applications,
imaging or radio astronomy [1–11] to enable fast and precise beam forming. Generally a
high resolution is required for beam forming which leads to relatively large antenna arrays
and complex signal processing systems. In some cases the antenna array should operate in
the near field region.
The far field radiation characteristics and beam forming methods for antenna arrays have
been thoroughly investigated and discussed in [1–6]. The coupling effect between typical
array elements can often be neglected.
Far field and near field are commonly distinguished by the far field distance definition
rmin = 2D 2 /λ with the largest dimension of the antenna array D and the wavelength λ. For
the investigations in this book, i.e. for the number of the antenna array elements M, N ≤ 10
and dx , dz = λ/2 (Fig. 4.1, 5.1 and 5.2), the near field range (distance between the center
point of the array and the observation point) remains shorter than 32λ, even though the
principles are valid for arbitrary antenna arrays with large numbers M and N [12].
The one-dimensional conformal antenna arrays can be designed as schematically illus-
trated in Fig. 5.1 and 5.2. The curvature of the conformal array profile, on which the single
antenna elements are positioned, can be defined by a curvature radius rc with corresponding
center points at yc (concave) or −yc (convex).

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 61

S.-P. Chen and H. Schmiedel, RF Antenna Beam Forming,
https://doi.org/10.1007/978-3-031-21765-4_5
62 5 Conformal Antenna Array

In Fig. 5.1 the one-dimensional conformal concave array is schematically illustrated with
the parameter +yc . In Fig. 5.2 the one-dimensional conformal convex array is schematically
illustrated with the parameter −yc .

yc = d 2 − xm,max 2 . (5.1)

Fig. 5.1 One-dimensional conformal concave array

Fig. 5.2 One-dimensional conformal convex array

5.1 Beam Forming of One-Dimensional Conformal Concave Arrays 63

The coordinate ym of the array element and distance from each single array element to the
observation point P(x, y) at r = 10λ rs = r − rmn can then be calculated

ym = ±(yc − rc 2 − xm 2 ). (5.2)

5.1 Beam Forming of One-Dimensional Conformal Concave Arrays

The concave patch array in Fig. 5.3 was investigated with all possible parameters. In Figs. 5.4
5.5, 5.6, 5.7, 5.8, 5.9, 5.10, 5.11, 5.12, 5.13, 5.14, 5.15, 5.16, 5.17, 5.18, 5.19, 5.20, 5.21, 5.22,
5.23, 5.24, 5.25, 5.26, 5.27, 5.28, 5.29 and 5.30 the results of the analytical methods [12, 13],
applied to a one-dimensional, concave patch antenna array, are compared pairwise with the
measurement results. The logarithmic scales for both graphs are identical. The maximum
in the radiation pattern is normalized to 0 dB. The minimum in the diagrams is −30 dB. It
can be clearly seen that the relatively high side lobes, in case of homogeneous amplitude
weights, can be significantly reduced by using binomial and Chebyshev amplitude weights.
In case of binomial amplitude weights, the beam width is also remarkably broadened. On the
other hand, narrower beams can be achieved by using a larger number of antenna elements.
To increase the sensitivity, and thus the dynamic range of the receiver, a parabolic reflector
is added to the receive antenna for the far field measurements.
Again in Figs. 5.4–5.30 measurement results confirm the simulation results of concave
arrays. Typically the measurements show somewhat higher side lobes than the simulations.
This is assumed to be due to small phase errors in the experimental setup and also reflection
from the antenna mount as well as reflection from the side not suppressed sufficiently. As a
main conclusion, we see, that the results of the concave arrangement is pretty much the same
as for the linear array. (Remark: For a change the designed steering angle is opposite to that
of the linear array diagrams.) So any curvature, on which antenna elements can be placed,
may be used, as long as the phases are corrected accordingly. An interesting special case is
for d = 10 λ, since the radius of the circle section on which the patch antenna elements are
placed is also 10 λ. So the receive antenna is placed exactly into the focal point. With equal

Fig. 5.3 One-dimensional conformal concave array

64 5 Conformal Antenna Array

Fig. 5.4 Far field characteristics of a patch antenna concave array, d = 35 λ, homogeneous amplitude
weights, beam steering 0◦ . a) Simulation; b) measurement

Fig. 5.5 Far field characteristics of a patch antenna concave array, d = 35 λ, homogeneous amplitude
weights, beam steering −15◦ . a) Simulation; b) measurement
5.1 Beam Forming of One-Dimensional Conformal Concave Arrays 65

Fig. 5.6 Far field characteristics of a patch antenna concave array, d = 35 λ, homogeneous amplitude
weights, beam steering −30◦ . a) Simulation; b) measurement

Fig. 5.7 Far field characteristics of a patch antenna concave array, d = 35 λ, binomial amplitude
weights, beam steering 0◦ . a) Simulation; b) measurement
66 5 Conformal Antenna Array

Fig. 5.8 Far field characteristics of a patch antenna concave array, d = 35 λ, binomial amplitude
weights, beam steering −15◦ . a) Simulation; b) measurement

Fig. 5.9 Far field characteristics of a patch antenna concave array, d = 35 λ, binomial amplitude
weights, beam steering −30◦ . a) Simulation; b) measurement
5.1 Beam Forming of One-Dimensional Conformal Concave Arrays 67

Fig. 5.10 Far field characteristics of a patch antenna concave array, d=35 λ, Chebyshev amplitude
weights, beam steering 0◦ . a) Simulation; b) measurement

Fig. 5.11 Far field characteristics of a patch antenna concave array, d = 35 λ, Chebyshev amplitude
weights, beam steering −15◦ . a) Simulation; b) measurement
68 5 Conformal Antenna Array

Fig. 5.12 Far field characteristics of a patch antenna concave array, d=35 λ, Chebyshev amplitude
weights, beam steering −30◦ . a) Simulation; b) measurement

Fig. 5.13 Near field characteristics of a patch antenna concave array, d = 10 λ, homogeneous ampli-
tude weights, beam steering 0◦ . a) Simulation; b) measurement
5.1 Beam Forming of One-Dimensional Conformal Concave Arrays 69

Fig.5.14 Near field characteristics of a patch antenna concave array, d=10 λ, homogeneous amplitude
weights, beam steering −15◦ . a) Simulation; b) measurement

Fig. 5.15 Near field characteristics of a patch antenna concave array, d = 10 λ, homogeneous ampli-
tude weights, beam steering −30◦ . a) Simulation; b) measurement
70 5 Conformal Antenna Array

Fig. 5.16 Near field characteristics of a patch antenna concave array, d = 10 λ, binomial amplitude
weights, beam steering 0◦ . a) Simulation; b) measurement

Fig. 5.17 Near field characteristics of a patch antenna concave array, d = 10 λ, binomial amplitude
weights, beam steering −15◦ . a) Simulation; b) measurement
5.1 Beam Forming of One-Dimensional Conformal Concave Arrays 71

Fig. 5.18 Near field characteristics of a patch antenna concave array, d = 10 λ, binomial amplitude
weights, beam steering −30◦ . a) Simulation; b) measurement

Fig. 5.19 Near field characteristics of a patch antenna concave array, d = 10 λ, Chebyshev amplitude
weights, beam steering 0◦ . a) Simulation; b) measurement
72 5 Conformal Antenna Array

Fig. 5.20 Near field characteristics of a patch antenna concave array, d = 10 λ, Chebyshev amplitude
weights, beam steering −15◦ . a) Simulation; b) measurement

Fig. 5.21 Near field characteristics of a patch antenna concave array, d = 10 λ, Chebyshev amplitude
weights, beam steering −30◦ . a) Simulation; b) measurement
5.1 Beam Forming of One-Dimensional Conformal Concave Arrays 73

Fig. 5.22 Near field characteristics of a patch antenna concave array, d = 5.6 λ, homogeneous ampli-
tude weights, beam steering 0◦ . a) Simulation; b) measurement

Fig. 5.23 Near field characteristics of a patch antenna concave array, d = 5.6 λ, homogeneous ampli-
tude weights, beam steering −15◦ . a) Simulation; b) measurement
74 5 Conformal Antenna Array

Fig. 5.24 Near field characteristics of a patch antenna concave array, d = 5.6 λ, homogeneous ampli-
tude weights, beam steering −30◦ . a) Simulation; b) measurement

Fig. 5.25 Near field characteristics of a patch antenna concave array, d = 5.6 λ, binomial amplitude
weights, beam steering 0◦ . a) Simulation; b) measurement
5.1 Beam Forming of One-Dimensional Conformal Concave Arrays 75

Fig. 5.26 Near field characteristics of a patch antenna concave array, d = 5.6 λ, binomial amplitude
weights, beam steering −15◦ . a) Simulation; b) measurement

Fig. 5.27 Near field characteristics of a patch antenna concave array, d = 5.6 λ, binomial amplitude
weights, beam steering −30◦ . a) Simulation; b) measurement
76 5 Conformal Antenna Array

Fig. 5.28 Near field characteristics of a patch antenna concave array, d = 5.6 λ, Chebyshev amplitude
weights, beam steering 0◦ . a) Simulation; b) measurement

Fig. 5.29 Near field characteristics of a patch antenna concave array, d = 5.6 λ, Chebyshev amplitude
weights, beam steering −15◦ . a) Simulation; b) measurement
5.2 Beam Forming of One-Dimensional Conformal Convex Arrays 77

Fig. 5.30 Near field characteristics of a patch antenna concave array, d = 5.6 λ, Chebyshev amplitude
weights, beam steering −30◦ . a) Simulation; b) measurement

phases for all antenna elements we obtain almost exactly the same radiation pattern as for
the linear array with equal phases in far field.

5.2 Beam Forming of One-Dimensional Conformal Convex Arrays

The convex patch array in Fig. 5.31 was investigated with all possible parameters. In
Figs. 5.32, 5.33, 5.34, 5.35, 5.36, 5.37, 5.38, 5.39, 5.40, 5.41, 5.42, 5.43, 5.44, 5.45, 5.46,
5.47, 5.48, 5.49, 5.50, 5.51, 5.52, 5.53, 5.54, 5.55, 5.56, 5.57 and 5.58 the results of the
analytical methods [12, 13], applied to a one-dimensional, convex patch antenna array, are
compared pairwise with the measurement results. The logarithmic scales for both graphs are
identical. It can be clearly seen that the relatively high side lobes in case of homogeneous
amplitude weights can be significantly reduced by using binomial and Chebyshev ampli-
tude weights. In case of binomial amplitude weights, the beam width is also remarkably
broadened. On the other hand, narrower beams can be achieved by using a larger number
of antenna elements. To increase the sensitivity, and thus the dynamic range of the receiver,
a parabolic reflector is added to the receive antenna for the far field measurements. For low
gains the noise floor can be seen in the measured characteristics.
78 5 Conformal Antenna Array

Fig. 5.31 One-dimensional conformal convex array

Fig. 5.32 Far field characteristics of a patch antenna convex array, d = 35 λ, homogeneous amplitude
weights, beam steering 0◦ . a) Simulation; b) measurement

Figures 5.32–5.58 show the simulation and measurement results of convex arrays. Again
the simulations are validated by the measurements. Small differences between simulation
and measurement results may be attributed to similar reasons as discussed before for concave
arrays. Also similar to concave arrays the results are pretty much identical to the linear array.
This demonstrates, that antenna elements may be placed on a convex circle on an antenna
tower, to allow for communication with a complete steering angle within 0◦ − 360◦ . That
is many convex sectors, similar to the sector we investigated, to complete a full circle.
5.2 Beam Forming of One-Dimensional Conformal Convex Arrays 79

Fig. 5.33 Far field characteristics of a patch antenna convex array, d = 35 λ, homogeneous amplitude
weights, beam steering 15◦ . a) Simulation; b) measurement

Fig. 5.34 Far field characteristics of a patch antenna convex array, d = 35 λ, homogeneous amplitude
weights, beam steering 30◦ . a) Simulation; b) measurement
80 5 Conformal Antenna Array

Fig. 5.35 Far field characteristics of a patch antenna convex array, d = 35 λ, binomial amplitude
weights, beam steering 0◦ . a) Simulation; b) measurement

Fig. 5.36 Far field characteristics of a patch antenna convex array, d = 35 λ, binomial amplitude
weights, beam steering 15◦ . a) Simulation; b) measurement
5.2 Beam Forming of One-Dimensional Conformal Convex Arrays 81

Fig. 5.37 Far field characteristics of a patch antenna convex array, d = 35 λ, binomial amplitude
weights, beam steering 30◦ . a) Simulation; b) measurement

Fig. 5.38 Far field characteristics of a patch antenna convex array, d = 35 λ, Chebyshev amplitude
weights, beam steering 0◦ . a) Simulation; b) measurement
82 5 Conformal Antenna Array

Fig. 5.39 Far field characteristics of a patch antenna convex array, d = 35 λ, Chebyshev amplitude
weights, beam steering 15◦ . a) Simulation; b) measurement

Fig. 5.40 Far field characteristics of a patch antenna convex array, d = 35 λ, Chebyshev amplitude
weights, beam steering 30◦ . a) Simulation; b) measurement
5.2 Beam Forming of One-Dimensional Conformal Convex Arrays 83

Fig.5.41 Near field characteristics of a patch antenna convex array, d = 10 λ, homogeneous amplitude
weights, beam steering 0◦ . a) Simulation; b) measurement

Fig.5.42 Near field characteristics of a patch antenna convex array, d = 10 λ, homogeneous amplitude
weights, beam steering −15◦ . a) Simulation; b) measurement
84 5 Conformal Antenna Array

Fig.5.43 Near field characteristics of a patch antenna convex array, d = 10 λ, homogeneous amplitude
weights, beam steering −30◦ . a) Simulation; b) measurement

Fig. 5.44 Near field characteristics of a patch antenna convex array, d = 10 λ, binomial amplitude
weights, beam steering 0◦ . a) Simulation; b) measurement
5.2 Beam Forming of One-Dimensional Conformal Convex Arrays 85

Fig. 5.45 Near field characteristics of a patch antenna convex array, d = 10 λ, binomial amplitude
weights, beam steering −15◦ . a) Simulation; b) measurement

Fig. 5.46 Near field characteristics of a patch antenna convex array, d = 10 λ, binomial amplitude
weights, beam steering −30◦ . a) Simulation; b) measurement
86 5 Conformal Antenna Array

Fig. 5.47 Near field characteristics of a patch antenna convex array, d = 10 λ, Chebyshev amplitude
weights, beam steering 0◦ . a) Simulation; b) measurement

Fig. 5.48 Near field characteristics of a patch antenna convex array, d = 10 λ, Chebyshev amplitude
weights, beam steering −15◦ . a) Simulation; b) measurement
5.2 Beam Forming of One-Dimensional Conformal Convex Arrays 87

Fig. 5.49 Near field characteristics of a patch antenna convex array, d = 10 λ, Chebyshev amplitude
weights, beam steering −30◦ . a) Simulation; b) measurement

Fig. 5.50 Near field characteristics of a patch antenna convex array, d = 5.6 λ, homogeneous ampli-
tude weights, beam steering 0◦ . a) Simulation; b) measurement
88 5 Conformal Antenna Array

Fig. 5.51 Near field characteristics of a patch antenna convex array, d = 5.6 λ, homogeneous ampli-
tude weights, beam steering 15◦ . a) Simulation; b) measurement

Fig. 5.52 Near field characteristics of a patch antenna convex array, d = 5.6 λ, homogeneous ampli-
tude weights, beam steering 30◦ . a) Simulation; b) measurement
5.2 Beam Forming of One-Dimensional Conformal Convex Arrays 89

Fig. 5.53 Near field characteristics of a patch antenna convex array, d = 5.6 λ, binomial amplitude
weights, beam steering 0◦ . a) Simulation; b) measurement

Fig. 5.54 Near field characteristics of a patch antenna convex array, d = 5.6 λ, binomial amplitude
weights, beam steering 15◦ . a) Simulation; b) measurement
90 5 Conformal Antenna Array

Fig. 5.55 Near field characteristics of a patch antenna convex array, d = 5.6 λ, binomial amplitude
weights, beam steering 30◦ . a) Simulation; b) measurement

Fig. 5.56 Near field characteristics of a patch antenna convex array, d = 5.6 λ, Chebyshev amplitude
weights, beam steering 0◦ . a) Simulation; b) measurement
5.2 Beam Forming of One-Dimensional Conformal Convex Arrays 91

Fig. 5.57 Near field characteristics of a patch antenna convex array, d = 5.6 λ, Chebyshev amplitude
weights, beam steering 15◦ . a) Simulation; b) measurement

Fig. 5.58 Near field characteristics of a patch antenna convex array, d = 5.6 λ, Chebyshev amplitude
weights, beam steering 30◦ . a) Simulation; b) measurement
92 5 Conformal Antenna Array

5.3 Beam Forming of 2-Dimensional Conformal Arrays

A planar antenna array, which is a special case of a conformal antenna array, is illustrated
schematically in Fig. 5.59 and 5.60, where xm and z n are the x- and z-coordinate of the
single array element. The center of the array and the focal point F(x F , y F , z F ) form the
beam steering angle θ0 or φ0 . rF − rmn is the vector between the array element (xm , ymn , z n )
and the focal point F(x F , y F , z F ). rmn is the vector pointing to the array element from the
coordinate origin (0, 0, 0) as reference point of the antenna array. Similarly the vector can
be defined for a distance between the array element and any arbitrary observation point at 10
λ. To be more precise, the beam steering angle is the angle with respect to the main lobe in
y-direction. Neglecting the coupling effects between the array elements, conformal arrays
can be investigated by simply superimposing radiation patterns from all the array elements
[12].
If ymn  = 0, we will have the conformal antenna arrays (Fig. 5.61 and 5.62). The curvature
of the conformal array profile, on which the single Hertzian dipoles or patch antenna elements
are positioned, can be defined by a curvature radius rc with corresponding center points at
yc (concave) or −yc (convex)

yc = rc 2 − xm,max 2 − z n,max 2 . (5.3)

The coordinate ymn of the array element and distance from each single array element to the
observation point P(x, y, z) at r = 10 λ rs = r − rmn can be calculated then as

Fig. 5.59 Planar antenna M×N array in the x-z plane with the focus point F(x F , y F , z F )
5.3 Beam Forming of 2-Dimensional Conformal Arrays 93

Fig. 5.60 Planar antenna M×N array with ymn (xm , z n ) = 0, cf. Fig. 5.59

Fig. 5.61 Convex conformal antenna M×N array with ymn (xm , z n ) > 0


ymn = ±(yc − rc 2 − xm 2 − z n 2 ), (5.4)

xr = r · sin(θ ) · cos(φ), (5.5)

yr = r · sin(θ ) · sin(φ), (5.6)
zr = r · cos(θ ), (5.7)

rs = (xr − xm )2 − (yr − ymn )2 − (zr − z n )2 . (5.8)
94 5 Conformal Antenna Array

Fig. 5.62 Concave conformal antenna M×N array with ymn (xm , z n ) < 0

The near field characteristics of this type of conformal antenna arrays can be analyzed either
by using numerical methods or by using the analytical method by superimposing all the
radiation fields resulting from all the array elements located on the conformal array surface,
like the planar antenna arrays described in details in the last chapters.

5.4 Comparison of Convex, Concave and Planar Profiles

Firstly different 10×10 conformal Hertzian dipole arrays (concave, convex, and planar)
are analyzed for different beam steering angles, i.e. θ0 = 75◦ and φ0 = 60◦ with the
focal point designed to be at a distance of 10λ to show the typical radiation characteristics
with homogeneous amplitude weights, binomial and with asymmetrical weighting function
proposed in [10, 12, 13] in combination with Chebyshev in this book to provide an optimum
side lobe suppression especially for the asymmetrical beam steering (Fig. 5.63, 5.64, 5.65
and 5.66). The parameter s [12, 13] can be determined individually on different antenna
array design parameters, such as the number of elements and the near field focal point, and
can be used to optimize the overall side lobe suppression.
The results for a linear array in last chapter show the improvement of beam focusing by
using additional asymmetrical amplitude weights in comination with Chebyshev amplitude
weights more comprehensively.
Therefore in this chapter we use this method to optimize the side lobe suppression also for
the near field applications of the 2-dimensional conformal antenna arrays. The asymmetrical
amplitude weights are used both parallel to the x- and z-axis.
The amplitude weight a(i) is divided into conventional coefficients W (m) (either homo-
geneous, binomial or Chebyshev) and an additional asymmetrical weighting function
5.4 Comparison of Convex, Concave and Planar Profiles 95

Fig. 5.63 Near field radiation pattern of a convex 10×10 Hertzian dipole array, with homogeneous
amplitude weights, for beam steering angles of 90◦ − θ0 = 15◦ and 90◦ − φ0 = 30◦ , with the focal
point at a distance of 10λ, rc = 9λ.

Fig. 5.64 Near field radiation pattern of a convex 10×10 Hertzian dipole array, with binomial ampli-
tude weights, for beam steering angles of 90◦ − θ0 = 15◦ and 90◦ − φ0 = 30◦ , with the focal point
at a distance of 10λ, rc = 9λ
96 5 Conformal Antenna Array

Fig. 5.65 Normalized radiation pattern of a convex 10×10 Hertzian dipole array, with Dolph Cheby-
shev amplitude weights, for beam steering angles of 90◦ − θ0 = 15◦ and 90◦ − φ0 = 30◦ , with the
focal point at a distance of 10λ, rc = 9λ

Fig. 5.66 Normalized radiation pattern of a convex 10×10 Hertzian dipole array, with Dolph Cheby-
shev amplitude weights and additional asymmetrical amplitude weights (s=5) for beam steering angles
of 90◦ − θ0 = 15◦ and 90◦ − φ0 = 30◦ , with the focal point at a distance of 10λ, rc = 9λ
5.5 Simulation Results for Conformal Antenna Arrays 97

aasym defined in equations (4.7)–(4.10). The phase shift of the array element is given by
α(i) = αx (i) or α(i) = αz (i) depending on the allocation of the array elements. Adjust-
ing the phases at the different distances according to equations (4.5) and (4.6) leads to the
patterns shown in Fig. 4.7–4.10. Parameter s is an optimization parameter used individu-
ally to minimize the side lobe level. Depending on the array size, distances between the
array elements and distances to the focal point, this parameter s can vary differently [10].
In Fig. 4.10 the parameter s was varied to achieve different side lobe suppression effects.
The parameter s is obtained emperically and iteratively. The results by using the simple
asymmetrical amplitude weight of equations (4.7)–(4.10) with s = 0.15–0.35 are shown in
Fig. 4.10.
A further substantial reduction of the dominant side lobe at φ = 180◦ is obtained when
combining the Chebyshev weights with the proposed asymmetrical amplitude weights [12].
The results have shown that the asymmetrical weighting coefficients lead to remarkable
further reduction of side lobe levels over a larger range. Especially the backward side lobes
(„back fire“) which can not be sufficiently reduced for near field beam steering by using
Chebyshev coefficients, can be minimized to a lower level.

5.5 Simulation Results for Conformal Antenna Arrays

The relatively high side lobe level of homogeneous amplitude weights can be reduced almost
completely by binomial amplitude weights, but the beam width is increased significantly,
whereas by using Chebyshev with additional asymmetrical amplitude weights, the beam
width remains relatively narrow and the side lobes are suppressed efficiently. By doing so,
the best near field focusing results can be achieved.
Figure 5.67 and 5.68 show the near field radiation characteristics of different conformal
antenna arrays, with beam steering angles φ = θ = 0◦ .
Figure 5.69 and 5.70 show the near field radiation characteristics of different conformal
antenna arrays, if the beam steering angle is non zero only in one plane, i.e. in either φ plane
or θ plane. The side lobes increase for all conformal array profiles especially in back fire
direction (Fig. 5.71 and 5.72).
Figure 5.73 and 5.74 show the near field radiation characteristics of different conformal
antenna arrays, with beam steering angles φ = 60◦ and θ = 75◦ . The side lobes are
suppressed further with asymmetrical amplitude weights for all antenna array profiles, even
though the beam width increases at the same time. The desired beam width can be achieved,
if correspondingly more array elements are used.
In this book the beam steering characteristics of conformal antenna arrays in the near-
field have been investigated and compared by using a very efficient analytical method [10,
12, 13].
Amplitude tapering by using the proposed asymmetrical weighting coefficients reduces
the side lobes over a larger angular range efficiently. Depending on the beam forming
98 5 Conformal Antenna Array

Fig.5.67 Comparison of normalized radiation patterns of different conformal 10×10 Hertzian dipole
arrays in φ plane, with Dolph Chebyshev amplitude weights for beam steering angles of 90◦ −θ0 = 0◦
and 90◦ − φ0 = 0◦ , with the focal point at a distance of 10λ, rc = 9λ

Fig.5.68 Comparison of normalized radiation patterns of different conformal 10×10 Hertzian dipole
arrays in θ plane, with Dolph Chebyshev amplitude weights for beam steering angles of 90◦ −θ0 = 0◦
and 90◦ − φ0 = 0◦ , with the focal point at a distance of 10λ, rc = 9λ
5.5 Simulation Results for Conformal Antenna Arrays 99

Fig.5.69 Comparison of normalized radiation patterns of different conformal 10×10 Hertzian dipole
arrays in φ plane, with Dolph Chebyshev amplitude weights for beam steering angles of 90◦ − θ0 =
15◦ and 90◦ − φ0 = 0◦ , with the focal point at a distance of 10λ, rc = 9λ

Fig.5.70 Comparison of normalized radiation patterns of different conformal 10×10 Hertzian dipole
arrays in θ plane, with Dolph Chebyshev amplitude weights for beam steering angles of 90◦ −θ0 = 0◦
and 90◦ − φ0 = 30◦ , with the focal point at a distance of 10λ, rc = 9λ
100 5 Conformal Antenna Array

Fig.5.71 Comparison of normalized radiation patterns of different conformal 10×10 Hertzian dipole
arrays in φ plane, with Dolph Chebyshev amplitude weights for beam steering angles of 90◦ − θ0 =
15◦ and 90◦ − φ0 = 30◦ , with the focal point at a distance of 10λ, rc = 9λ

Fig.5.72 Comparison of normalized radiation patterns of different conformal 10×10 Hertzian dipole
arrays in θ plane, with Dolph Chebyshev amplitude weights for beam steering angles of 90◦ −θ0 = 15◦
and 90◦ − φ0 = 30◦ , with the focal point at a distance of 10λ, rc = 9λ
5.5 Simulation Results for Conformal Antenna Arrays 101

Fig.5.73 Comparison of normalized radiation patterns of different conformal 10×10 Hertzian dipole
arrays in φ plane, with Dolph Chebyshev amplitude weights and additional asymmetrical amplitude
weights (s=5) for beam steering angles of 90◦ − θ0 = 15◦ and 90◦ − φ0 = 30◦ , with the focal point
at a distance of 10λ, rc = 9λ

Fig.5.74 Comparison of normalized radiation patterns of different conformal 10×10 Hertzian dipole
arrays in θ plane, with Dolph Chebyshev amplitude weights and additional asymmetrical amplitude
weights (s=5) for beam steering angles of 90◦ − θ0 = 15◦ and 90◦ − φ0 = 30◦ , with the focal point
at a distance of 10λ, rc = 9λ
102 5 Conformal Antenna Array

requirement, this method can be combined with the well-known far-field technique (bino-
mial or Chebyshev) to achieve certain radiation characteristics and desired beam widths by
using a larger number of array elements. The method presented in this book can be used
to quickly provide the optimum parameters for dynamic beam forming problems for the
adaptive planar or conformal antenna arrays. The same method could be used to analyze
generally volumetric antenna array problems.

References

1. R. E. Collin, F. J. Zucker: Antenna Theory, Part 1. McGraw-Hill Book Company (1969).

2. K. F. Lee: Principles of Antenna Theory. John Wiley & Sons Ltd. (1984).
3. B. D. Steinberg, H. M. Subbaram: Microwave Imaging Techniques. John Wiley & Sons, Inc.
(1991).
4. R. J. Mailloux: Phased Array Antenna Handbook. Artech House Inc. (1994).
5. L. V. Blake, M. W. Long: Antennas: Fundamentals, Design, Measurement. 3rd Edition. Scitech
Publishing, Inc. (2009).
6. D. G. Fang: Antenna Theory and Microstrip Antennas. CRC Press (2010).
7. W. H. Carter: On Refocusing a Radio Telescope to Image Sources in the Near Field of the
Antenna Array. IEEE Trans. Antennas Propagat., Vol. 37, pp. 314-319 (1999).
8. A. Badawi, A. Sebak, L. Shafai: Array Near Field Focusing. WESCNEX’97 Proceedings of
Conference on Communications, Power and Computing, pp. 242-245 (1997).
9. M. Bogosanovic, A. G. Williamsoni: Antenna Array with Beam Focused in Near Field Zone.
Electronics Letters, vol. 39, pp. 704-705 (2005).
10. J. Grubert: A Measurement Technique for Characterization of Vehicles in Wireless Communi-
cations. PhD Thesis (in German) of Technical University Hamburg-Harburg, Cuvillier Verlag
Goettingen (2006).
11. S. Ebadi, R. V. Gatti, L. Marcaccioli, R. Sorrentinoi: Near Field Focusing in Large Reflector
Array Antennas Using 1-bit Digital Phase Shifters. Proceedings of the 39th European Microwave
Conference, pp. 1029-1032 (2009).
12. S.-P. Chen: Improved Near Field Focusing of Antenna Arrays with Novel Weighting Coefficients.
IEEE WiVeC 2014, 6th International Symposium on Wireless Vehicular Communications (2014).
13. S.-P. Chen: An efficient Method for Investigating Near Field Characteristics of Planar Antenna
Arrays. Global Conference of Wireless and Optical Communications, Malaga, Spain (September
2016). Wireless Personal Communications, 95(2), 223-232, 2017.
MIMO Antenna Systems
6

In the previous chapters, linear phased arrays, planar phased arrays and conformal phased
arrays were investigated, by applying certain amplitude weights and phase shifts between
the antenna elements, in order to form the beam and focus the beam to a certain receiver.
Many researchers have investigated intensively within the last decades to find different
sophisticated methods and algorithms to optimize the SINR (Signal to Interference and
Noise Ratio) and the channel capacities, by using MIMO (Multiple Input and Multiple
Output) antenna arrays, used in many wireless and mobile communication systems like
LTE, 5G and future 6G [1, 2].

MIMO Channel
Normally the MIMO channel can be described generally by the model in Fig. 6.1 (see also
similar schematic model diagrams in [1, 2]) by

y = H · x + n, (6.1)
with H as channel matrix describing the relation between the transmit signal vector x con-
taining the transmitted data and received signal prior to additive white Gaussian noise n,
so that receiver receives the disturbed signal y. The received signal vector y is disturbed by
noise n and interference is caused by multipath propagation.
 T
x = x1 x2 ... x M , (6.2)

 T
y = y1 y2 ... y N , (6.3)

 T
n = n 1 n 2 ... n N , (6.4)

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 103
S.-P. Chen and H. Schmiedel, RF Antenna Beam Forming,
https://doi.org/10.1007/978-3-031-21765-4_6
104 6 MIMO Antenna Systems

with the channel transfer matrix or channel transfer function

⎡ ⎤
h 11 h 12 h 13 ... h 1M
⎢ h 21 h 22 h 23 ... h 2M ⎥
H=⎢ ⎣ ... ... ... ...
⎥.
⎦ (6.5)
h N 1 h N 2 h N 3 ... h N M

If several consecutive vectors will be transmitted, the transmitted and received data vectors
can be arranged each in a matrix [3]
 
X = x1 x2 ... xM (6.6)

 
Y = y1 y2 ... yN (6.7)

 
N = n1 n2 ... nN (6.8)

Y = HX + N (6.9)
If the multipath propagation effect can be neglected, then the open loop beam forming
can be easily done, as discussed in the last chapter with amplitude weights and phases. The
MIMO channel can be described with the model in Fig. 6.1 with M transmit antennas and
N receive antennas. The channel transfer matrix H contains the matrix element h i j with
i = 1...N and j = 1...M and the spatially additive white circular Gaussian random noise
n = CN(0; σ 2 I) as additive white Gaussian noise.
The special cases for SISO (Single Input Single OUTPUT), SIMO (Single Input Multiple
Output) and MISO (Multiple Input Single Output) could described by using the same model,
if the parameters are set as

SISO: M = N = 1,
SIMO: M = 1,
MISO: N = 1.

Generally the transmit antenna for forming narrow beams should be an antenna array,
either linear, planar or a conformal antenna array, i.e. multiple antenna elements. The receiver
generally can also consist of multiple antenna elements, commonly known as MIMO antenna
systems [1]. One special case for MIMO antenna system is MISO. In this case a receiver
consists of only one single antenna element.
By using MIMO either Single-User-MIMO (SU-MIMO) or Multiple-User-MIMO (MU-
MIMO) could be implemented. In case of SU-MIMO, the receive antenna arrays with ampli-
tude weights and phase shifts form a prefered receiving beam pointing exactly into the direc-
tion of the transmitter. In case of MU-MIMO, i.e. for N users, N different MIMO receive
6 MIMO Antenna Systems 105

Fig. 6.1 MIMO model

antenna arrays will be implemented at the site of the users, each correspondingly like the
SU-MIMO case. The transmit antenna arrays are precoded with the amplitude weights and
phase shifts in such a way, that multiple beams will be formed pointing at these N users.

Wave Propagation, Fading

In case of an ideal LOS (line of sight) wave propagation environment, like wave propagation
in the rural region without multiple reflections caused by the buildings, the classical beam
forming techniques can be used with the calculated amplitude weights and phase shifts,
like discussed in the previous chapters. The path loss is the dominating factor for the wave
propagation
λ
Pr = Pt · · Gr · G t (6.10)
2·π ·d
with Pr , Pt as received and transmitted power, G r and G t as transmit and receive antenna
gain, d as transmission distance and λ as RF carrier wavelength.
If however, the different reflected beams (e.g. due to the multipath propagation) from
a transmitter antenna cannot be completely neglected, these will be superimposed either
106 6 MIMO Antenna Systems

constructively or destructively, leading to dynamic changing received power at the receiver

antennas, with or without LOS wave propagation of the desired beam.
Different fading effects can be classified as follows (see for example [4]):

a) macroscopic fading caused by shadowing effects;

b) microscopic fading consisting of rapid fluctuations of the received signal caused by
multipath propagation, when LOS propagation is not possible;
c) Doppler spread caused by movement of transmitter and receivers with velocity v leading
to Doppler frequency shifts at the carrier frequency f c ± f D . The corresponding coherence
time Tc indicates how fast the channel changes in the time domain and thus defines the
symbol duration limit to avoid the disturbance caused by the Doppler spread;
d) Delay spread due to multipath leading to an average delay spread of τ . The corresponding
coherence bandwidth Bc = 1/τ defines the channel or subchannel bandwidth in order
to avoid or minimize the disturbance of the multiple propagation. When the spectral
bandwidth for the transmission B > Bc , the N multi-subcarriers are used with the single
subcarrier bandwidth less than the coherence bandwidth Bc

B
Bn =  Bc . (6.11)
N
This type of robust multi-subcarrier-system, like OFDM modulation techniques, was
proposed for many digital wireline and mobile communication systems, like DMT in xDSL,
DAB, DVB, WLAN and LTE, in order to ensure the optimized SINR and total channel
capacities C (see [3, 4])
1
C = B log2 I + 2 HPHH , (6.12)
σ
with H as channel gain matrix, B as channel bandwidth, and P as covariance of the input
signal x
P = E{xx H }. (6.13)
In case of the multi-subcarrier system OFDM with Si as received signal at the i-th
subcarrier, Bn as subcarrier band width, which is constant in most cases, N0 as spectral
noise density and σ as variance of the additive white Gaussian noise. With respect to the
channel model in equation (6.1), the channel capacity can then be written as
N
Si
C= Bn log2 1 + . (6.14)
N0 Bn
i=1

When the LOS propagation dominates the other paths, this type of flat fading channel
is called Rician channel, or macroscopic fading. Due to the dominating LOS beam, the
variation along the propagation path is more large scale that is rather slow. The beam forming
and beam steering at the receiver can be designed and implemented like described in the
previous chapters, with the required amplitude weights and phase shifts precoded based on
6 MIMO Antenna Systems 107

the knowledge of the channel. At the receiver side, the received signals will be postprocessed
or estimated by some algorithms based on the knowledge of the channel. This technique is
also called open loop beam forming. Since the channel changes are slow, for example for
4th Generation of Mobile Communications LTE, the so-called code-books for the precoding
matrix could be applied to form the beams at the transmitter [5, 6].
In fast fading, or frequency selective fading in multipath propagation environment, the
channel suffers constructive and destructive superposition of the multipath beams leading
to fast fading or deep fade. Therefore the channel is no more deterministic. In worst case the
line of sight LOS propagation is not even possible. This channel is called Rayleigh channel.
Rayleigh channels are mostly in urban areas, where many buildings reflect the transmitted
beams, behave statistically and dynamically changing.
The Rician fading and Rayleigh fading could be described by using a unique formula by
assuming the perfect LOS channel or direct path HLOS and HNLOS selective fading channel
or the sum of all scattered paths H N L O S [4], then a fading effect can be described by using
one single formula
K 1
H= HLOS + HNLOS (6.15)
1+K 1+K
with K = 0 for non LOS (Rayleigh fading), 0 < K < 1 for Ricean fading and K = 1
corresponds to the perfect LOS mobile communication channel.

Open Loop Beam Forming, Closed Loop Beam Forming

In open loop beam forming, channel spatial multiplexing and diversity gain can be achieved,
even though these are not the optimized ways to estimate the received signals.
The channel characteristics should be estimated either semi-blind or by using the pilot
symbols or training symbols. The received signals can be obtained, if the channel character-
istics are either known as CSI (Channel State Information) or can be estimated by some cost
function optimization algorithms, like Zero Forcing (ZF), Minimum Mean Square Errors
(MMSE) or Maximum Likelihood (ML). In this case, the inverse channel matrix will be
multiplied with the received signals y and the channel signal-to-interference-and-noise ratio
(SINR) and the channel capacities will be maximized. By using this information the pre-
coding matrix for the transmitter “or combining matrix for the receiver” can be utilized to
optimize the SINR and therefore the total channel capacity.
By using a large number of the transmit antennas and receive antennas, diversity gain
will be achieved to improve the SINR.
Another method is to define a pilot signal grid spanning over the time domain and over the
subchannels. By using the well known pilot symbols transmitted over the multipath chan-
nels, the characteristics of the fast fading channels can be determined for all resource blocks
(RB), by applying some selected pilot symbols at different selected sub-carriers and by inter-
polation between them. For a dynamically changing multipath propagation environment, the
channel measurements should be performed continuously and feedback information will be
updated correspondingly. These provide the precise channel measurement data, allowing
108 6 MIMO Antenna Systems

exact determination of the precoding and combining matrix for MIMO systems, but on the
other hand lead to delay and overhead.
If we use predefined pilot symbols at different selected sub-carriers, the received signals,
together with the channel noise and interference, can be used to determine the channel
characteristics by performing the channel measurements. The channel measurement results
can be used to determine the transmit antenna array amplitude weights and phase shifts,
the so-called precoding matrix. The precoding matrix can be used to control the transmit
powers and phases of the single antenna elements. On the receiver site, in case of MU-MIMO
antenna arrays, similarly the receive antenna array amplitude weights and phase shifts will
be applied to combine the received signals and to achieve the best SINR or channel capacity.
This process is called closed loop beam forming. The most powerful method is the Maximum
Likelihood (ML) criterion.
In 5G mobile communications, the concept of massive MIMO is applied [2], where a
huge number of antenna elements is used for the transmit and receive antenna. This is to
achieve an efficient beam forming, either for SU-MIMO or MU-MIMO applications in a
base station coverage area (Fig. 6.2). Here the single user or multiple users could also be
considered as single hotspot of mobile users, or multiple hotspots of mobile users in the cell.
If the multipath propagation effect can be neglected in some situations, i.e. the wave
propagation environment is predominantly LOS, for example in the rural region, the antenna
arrays for SU-MIMO, discussed typically in the last chapters, can be used to do the beam
forming.
It can be considered as multiple MISO in downlink and multiple SIMO in uplink where
there is no correlation between the user equipments (UEs). The extension to the multi-user
MIMO is massive MIMO and it is preferred since it saves the UE a lot of burdens and loads
of the complexity of the eNodeBs.

Fig. 6.2 Multi-user MIMO system.

6 MIMO Antenna Systems 109

Fig. 6.3 Antenna pattern of a 1×16 or open-loop SU-MIMO system with steering angle 45◦ .

Generally MU-MIMO can be implemented in two ways: complete channel measurement

and channel estimation for multiple users, either pseudo-dynamically or statically, for exam-
ple for some fixed pico-cells by using 4×16 arrays (4 rows and 16 columns are necessary).
If multiple users or hotspots should be covered, for examples 2 hotspots of users, then a very
simple and straightforward method is to use subarrays of a planar antenna array that can
be used to form the single beams, for example 4 times 4×16 or 4×64 antenna arrays, each
for one single beam with sufficient side lobe suppression, in order to avoid any interference
with the other beams.
Obviously this method is more flexible, less complex and requires less computation time
for the channel measurement and channel estimation. If open loop beam forming can be
applied, then the different beams can be formed very quickly. Figure 6.3 shows a planar
antenna array of 1×16 antenna elements with a steering angle of 45◦ . For the simulation,
the 2.5 GHz ISM band is used for 1×16 elements.
In Fig. 6.4 an example of open loop sub-array beam forming (steering angles 45◦ and
105◦ ) is illustrated, by using two sub-arrays of 2×16 antenna elements.
In classical open loop beam forming, the multiple transmit Nt and receive antennas Nr
are used to improve the diversity to achieve the maximum ratio or the optimized SINR. The
signal s0 will be transmitted Nt times in parallel to receive the best diversity. On the other
110 6 MIMO Antenna Systems

Fig. 6.4 Antenna pattern of a 2×16 or open-loop MU-MIMO system with steering angles 45◦ and
105◦ .

hand, the channel efficiency will be significantly reduced, if the number of the parallel signal
data streams is higher. This method is called Maximum Ratio Receive Combining (MRRC).
The most efficient method to guarantee diversity without reducing the efficiency is pro-
posed by S. Alamouti [7] by using more than 2 antennas (frequency domain) and 2 subsequent
time slots (time domain) to transmit orthogonal complex symbols. The Alamouti scheme is
also called STBC (Space-Time Block Coding). The diversity remains similar in comparison
with MRRC. The channel for MRRC could be described as follows

r0 = h0 s0 + n0 , (6.16)

r1 = h1 s0 + n1 . (6.17)
To illustrate the Alamouti model and compare the Alamouti with MRRC comprehen-
sively, two channels are used to transmit two complex symbols so and s1 with the channel
response ho and the channel response h1 with the additive noises no and n1

r0 = h0 s0 + h1 s1 + n0 , (6.18)
6 MIMO Antenna Systems 111

r1 = −h0 s1∗ + h1 s0∗ + n1 . (6.19)

MIMO is the most prominent technology to achieve the 5G requirements for high channel
capacity and efficiency. 5G strives to increase the throughput of the wireless channel and in
order to do that it requires the utilisation of all possible radio access technologies (Multi-
ple RATs) and therefore the multiple subcarrier technique Orthogonal Frequency Division
Multiple Access (OFDMA).
Modern 4G mobile communication technology LTE-A and also 5G are able to achieve
the best possible utilization of time domain and frequency domain multiple access through
OFDM combined with carrier aggregation and also manage to reduce the effect of ISI by
using cyclic prefixes. By performing a channel measurement and channel estimation with
pilots, in combination with zero-forcing algorithm, the precoding matrix can be determined
in a multi-path propagation environment. Channel estimation plays an important part in
an OFDM system. It is used to increase the capacity of OFDMA systems by improving
the system performance in terms of bit error rate (BER). To facilitate the estimation of the
channel characteristics, LTE uses cell-specific reference signals (pilot symbols) inserted in
both time and frequency [5]. In the 3GGP Technical Specifications [6] the channel measure-
ments are performed by using so-called multi-layer, i.e. up to 8 antenna ports. These pilot
symbols provide an estimate of the channel at given locations within a sub-frame. Through
interpolation, it is possible to estimate the channel across an arbitrary number of sub-frames.
The pilot symbols in LTE-A are assigned at certain positions within a sub-frame depending
on the cell identification number and the transmit antenna. In [5, 6] so-called codebooks
for precoding/weighting of the antenna arrays are defined und used in the beam forming
depending on the channel measurement results. UE-specific demodulation reference signals
(DM-RS) for demodulation of Physical Downlink Shared Channel (PDSCH) depending on
the channel conditions by using QPSK, 16 QAM or 64 QAM. CSI-RS allows the cell-specific
UE downlink channels state information (CSI) measurements.
Generally the channels are measured by considering multipath propagation or fading
effect, interferences and noises. These are then estimated. The input signals x are pre-coded
after the channel measurement and channel estimation, so that the received signals y will
be exactly as desired d, i.e. the MIMO beams will exactly point at the user terminals (user
1, user 2 and user 3). Generally speaking the users could be single users or devices, or also
pico-cell base stations which cover a hot-spot with many mobile users.
Generally the MIMO channel transfer matrix H in equation (6.1) can be decomposed to
(see also Fig. 6.5)
H = USV H (6.20)
with
ỹ = U H y = U H H x + U H n (6.21)
and
x = Vx̃. (6.22)
112 6 MIMO Antenna Systems

Fig. 6.5 MIMO channel model.

We obtain the new channel matrix equation, if the channel is known to the transmitter [4]

ỹ = U H HVx̃ + U H n = U H USV H Vx̃ + U H n = Sx̃ + ñ. (6.23)

The decomposed matrices S, U H and V represent the diagonal eigenvalue matrix, the
combining or postcoding matrix at the receiver, and precoding matrix at the transmitter
respectively. The matrix element si j corresponds to the transfer function or the power of the
corresponding orthogonal sub-carrier, resource block, and the trace of S is the total transmit
power of the M transmit antennas. The channel estimation algorithm extracts the precoding
matrix for the transmit antenna array and combining matrix for the receive antenna array
from the channel measurement data achieved by using selected pilot signals at different
sub-carriers and different time-slots. Depending on the channel characteristics the pilot grid
can be chosen with strong oversampling and large overhead, or appropriate pilot and low
overhead with additional interpolation between the pilot signals.

Matched filters
Matched filters are optimized to maximize the SNR in the context of the received signals,
when the impulse response is known and the noise is additive white noise. However the
matched filters do not address ISI (intersymbol inferference) [8].
Another methods are the zero-forcing ZF algorithm [2, 8] and the maximum ratio algo-
rithm.

Zero-Forcing
Zero-forcing eliminates the interferences among the multiplexed signals, whereas the max-
imum ratio method tries to maximize the total channel capacity. Zero-forcing equalizers
minimize the ISI but ignore the impact of the noise on the system [8]. Nevertheless zero-
forcing can enhance additive noise to a certain extend as it suppreses the ISI.
6 MIMO Antenna Systems 113

Maximum Ratio Method

The maximum ratio method tries to maximize the total SNR or/and total channel capacity of
all sub-carriers. This can be done by adaptively mapping the data symbols to the sub-carriers
with good SNR and dropping such sub-carriers with low SNR. One famous method is the
so-called Water-Filling algorithm [9–11].
This method also improves the peak to average power ratio (PAPR), in order to achieve
the best linear back-off for the linearity of the amplifiers.

Minimum Mean Square Error

The least squares estimate or Minimum Mean Square Error (MMSE) criterion of the channel
transfer function can be estimated [1, 2] by considering both ISI and additive white noise.
In this method the error is defined as cost function J M S E with the target that the mean
square error MSE should be minimized, with the error  [1]

 k = dk − w H · x k , (6.24)

J M S E (E{k2 }) = E{(dk − wH · xk )2 } = dk2 − 2wH E{dk xk } + wH E{xk xkH }w. (6.25)

The cost function J M S E becomes minimum, if the partial derivative

∂ J M S E (E{2k })
=0 (6.26)
∂w
or
w = E{xk xkH }−1 E{dk xk }. (6.27)
This solution is also commonly known as Wiener solution.
In case of multicarrier modulation OFDM, the sum of sub-carrier channel capacity should
be maximum. For this purpose, the sub-carriers with low SNR (Signal Noise Ratio) will
be dropped or let unused, only the sub-carriers with good SNR values will be adaptively
modulated with signal symbols (for example BPSK, QPSK, 16QAM, 64QAM, 256QAM) to
achieve the best total channel capacity. This procedure is also called Water-Filling algorithm
[9–11].
We would like to additionally emphasize that the above-mentioned techniques could
also be combined with the most advanced modulation technology probabilistic amplitude
modulation PAM to approach the Shannon channel capacity limit.
The least squares errors (LSE) are then averaged to reduce any unwanted noise from
the pilot symbols since the average of noise is zero. That is fast and easy to do in the case
of SISO but once the system is expanded and becomes Multi User MIMO (Fig. 6.2) the
channel estimation requires an adaptive filter to take many iterations to get a solid channel
estimation. The coordination for the pilots also becomes more complicated to ensure that
they remain orthogonal and unique in case of zero forcing, where different interfering beams
should be minimized at the desired receiver.
114 6 MIMO Antenna Systems

The uplink and downlink capacity CU L and C DL can be estimated also by using Shannon’s
formula
M
C= Bi · log2 1 + S N Ri . (6.28)
i=1

Hybrid Beam Forming

Up to now open loop beam forming and closed loop beam forming can involve time domain
resource elements (RE) and frequency domain sub-carriers in OFDM involving x and y
with all the REs. This leads to mathematical calculations with extremely large sizes of the
matrices.
5G New Radio (NR) wireless communication systems use MIMO beamforming technol-
ogy for signal-to-noise ratio (SNR) enhancement and spatial multiplexing to improve the
data throughput in scatterer-rich environments. In a scatterer-rich environment, line-of-sight
(LOS) paths between the transmit and receive antennas are not always present. In [12] the
design of MIMO phased arrays, including complex subarray structures, so-called hybrid
beam forming is applied to combine RF and Baseband beam forming.
To gain the required throughput, MIMO beamforming implements precoding on the
transmitter side and combining on the receiver side to increase SINR and to separate spatial
channels. However, a full digital beamforming structure requires each antenna to have a
dedicated RF-to-baseband chain, which can increase the overall hardware cost and drive the
power consumption higher. As a solution, hybrid beamforming is used to employ fewer RF-
to-baseband chains. With deliberate selection of the weights for precoding and combining,
hybrid beamforming can achieve a level of performance that is comparable to that of full,
all-digital beamforming [12].

References

1. C. A. Balanis: Antenna Theory. John Wiley & Sons, Inc. Fourth Edition (2016).
2. Th. Marzetta, E. G. Larsson, J. Yang, H. Q. Ngo: Fundamentals of Massive MIMO. Cambridge
University Press (2016).
3. E. G. Larsson, P. Stoica: Space-Time Block Coding for Wireless Communications. Cambridge
University Press (2003).
4. A. Paulraj, R. Nabar, D. Gore: Introduction to Space-Time Wireless Communications. Cambridge
University Press (2003).
5. Bernhard Schulz: Rohde & Schwarz Whitepaper: LTE Transmission Modes and Beamforming
(2016).
6. Technical Specification Group Radio Access Network: Physical Channels and Modulation,
Release 10; 3GPP TS 36.211 V 12.5.0, March (2015).
7. S. Alamouti: A Simple Transmit Diversity Technique for Wireless Communications. IEEE
JOURNAL ON SELECT AREAS IN COMMUNICATIONS, VOL. 16, NO. 8 (1998).
8. Ph. Golden, H. Dedieu, K. Jacobsen: Fundamentals of DSL Technology. Auerbach Publications,
Taylor & Francis Group. Boca Raton Florida, New York (2006).
References 115

9. G. Muenz, S. Pfletschinger, J. Speidel: An Efficient Waterfilling Algorithm for Multiple Access

OFDM (2002). IEEE Global Telecommunications Conference 2002 (Globecom ’02), Taipei,
Taiwan (2002).
10. L. G. Li, G. Stueber: Orthogonal Frequency Division Multiplexing for Wireless Communications.
1st Edition. New York : Springer Science+Business, Media Inc. (2006).
11. M. Kuhn: Advanced Modulation. Lecture Notes. Darmstadt University of Applied Sciences (July
31, 2021).
12. Matlab Whitepapaer: Exploring Hybrid Beam Forming Architecture for 5G Systems (2015).
Thinned Antenna Array
7

In previous chapters we investigated linear antenna arrays, planar antenna arrays, and con-
formal antenna arrays with different amplitude weights and phase shifts, but always with the
same distances between the neighbouring antenna elements with d = λ/2. Different ampli-
tude weights and phase shifts were chosen, in order to form and steer the beams on one
hand, and on the other hand to suppress the side-lobes.
Besides the equidistant allocation of the antenna elements, sometimes either larger dis-
tances (d > λ/2), or non-equidistant antenna elements (d = λ/2, 2 λ/2, 3 λ/2, 4 λ/2, ... etc.)
could be used, generally with increasing distance from the array center to the array edges.
Also prime number times λ/2 could be chosen as distances between the antenna array ele-
ments.
In Fig. 7.1 for simplicity we consider different 1-dimensional, linear antenna array types:
(a) 8-element λ/2 array with aperture width of 7×λ/2; (b) 8-element λ array with aperture
width of 7×λ; (c) 8-element non-equidistant array with aperture width of 13×λ/2; (d)
15-element λ/2 array with aperture width of 7×λ/2, the same size of (b).
First we compare the radiation characteristics of 8-element λ/2 array with 8-element λ
array (Fig. 7.2). It can be clearly seen, that the side-lobe of the 8-element λ array is slightly
raised, i.e. the directivity is reduced correspondingly.
Similarly the radiation characteristics of an 8-element non-equidistant array shows raised
side-lobes in comparison with the reference 8-element λ/2 array (Fig. 7.3), i.e. the directivity
is reduced correspondingly.
However, if we compare the 8-element λ array with 15-element λ/2 array (Fig. 7.4), where
both arrays have the same aperture width, it can be clearly seen, that the side-lobes of the
15-element λ/2 array are suppressed in comparison with the 8-element λ array. At the same
time, the radiation beam width is narrowed, i.e. the directivity is increased correspondingly.

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 117
S.-P. Chen and H. Schmiedel, RF Antenna Beam Forming,
https://doi.org/10.1007/978-3-031-21765-4_7
118 7 Thinned Antenna Array

Fig. 7.1 Different antenna array types: (a) equidistant λ/2 array; (b) thinned, equidistant λ array;
c) thinned, non-equidistant array; (d) equidistant λ/2 array with 15 elements and same aperture size
like (b)

Fig. 7.2 Thinned, equidistant λ array

7 Thinned Antenna Array 119

Fig. 7.3 Thinned, non-equidistant array

Fig. 7.4 Comparison of equidistant 8-element λ array with 15-element λ/2 array, corresponding to
the same aperture width

Reference

1. R. E. Collin, F. J. Zucker: Antenna Theory, Part 1. McGraw-Hill Book Company (1969)

Applications of Phased Arrays
8

Antenna arrays can be widely used for various applications, like space communications,
radio broadcasting, radar techniques and mobile communications.

8.1 Radar, Broadcasting and Positioning

Phased array antennas come in a variety of many, many different applications. There are
fixed constellations, where the single antenna elements are connected to a fixed network
providing the signals to all the involved antenna elements. Often equal phase is applied to
all elements. When these elements are located in a plane, the main direction of radiation is
of course perpendicular to this plane.
In order to improve the efficiency of radar applications, higher transmitting antenna gain
is desired, to resolve small radar cross section RCS [3], either by using high gain antenna
elements like horn antennas or parabolic antennas, or the phased antenna arrays, firstly to
achieve higher gain, and secondly to enable beam steering by adaptively choosing the phase
shifts of the antenna elements.
This type of fixed arrays is often found in simple RADAR units where the group of array
elements is used to enlarge the overall antenna aperture to increase the gain or reduce the
half-power beamwidth vice versa.
The same type of a fixed array is for example employed on the GPS satellites. In this
case a constellation of 15 helical antennas is being used on the block II satellites. For GPS
the beamwidth is such that the whole earth is being illuminated, more or less evenly. Also
the helical antenna elements facilitate the radiation of circularly polarized waves.
Besides these fixed constellations there are variable, steerable constellations where all
elements are controlled in phase and amplitude. Individual phase shifters and attenuators
control phase and amplitudes of all individual antenna elements for transmitting signals into
a wanted direction in space. Vice versa the antenna element’s phases and amplitudes are

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 121
S.-P. Chen and H. Schmiedel, RF Antenna Beam Forming,
https://doi.org/10.1007/978-3-031-21765-4_8
122 8 Applications of Phased Arrays

controlled for receiving a wanted signal from a defined direction in space. Evaluating the
incoming phases also allows for calculating the angle of arrival (AoA), The phase shifts
and amplitude adjustments for transmit and receive may be adjusted in the operating radio
frequency (RF) range.
Often it is more convenient to adjust phase and amplitude at a lower intermediate fre-
quency (IF) with digital techniques. When an RF signal is down converted with the help of
an ideal mixer and a fixed local oscillator (LO), the signal at the IF will be an exact replica
of the RF signal. It has the same information as phase and amplitude. This is also true for
up converting an IF signal to an RF signal, as described in Chap. 3.
We will briefly describe typical and well known examples for steerable phased array
antennas, i.e. the PATRIOT system and the IRIDIUM system. The PATRIOT system [1] is a
surface-to-air defense system that was first operational in the 1980s There is a sophisticated
antenna array consisting of 5161 array antenna elements. The size is about 2.5 m. Its purpose
is to track incoming rocket targets, control its own rockets and perform identification of
friends and foe (IFF). The search angle is 90◦ . The system is able to track up to 100 targets
simultaneously. Simultaneously, in this context, means sequential in a very fast sequence.
The PATRIOT system can launch and control up to 9 defense rockets to intercept incoming
missiles. The Russian counterpart, which has a similar capability, is the SA-10 system.
The IRIDIUM system [2] was developed as a world-wide mobile telephone system. The
satellite system consists of 66 satellites in low-earth orbit (LEO) covering the whole surface
of the earth. Each satellite has three phased array panels with 165 antenna elements each
located in 10 columns. Each panel covers 16 cells on the earth’s surface with individual
beams. So one satellite covers 48 cells simultaneously for data communication. The antenna
elements are single patch radiators. Every antenna element has a transmit/receive (T/R)
module. These beams are controlled permanently. When a satellite flies out of sight of a cell,
the next, adjacent satellite resumes the data communication by a sophisticated hand-over
procedure.
A similar satellite system for worldwide Internet access, however with a huge number of
many thousands of satellites, is the Starlink System provided by SpaceX. Also this system
uses extensively phased array antennas for the satellites and also the subscriber ground
stations.
Another state-of-the art application for steerable antenna arrays is the 5G mobile system
with its MIMO technology.
Most applications discussed so far employ antenna elements located in a plane or oriented
linearly and transmitting into far field or receiving from far field. It is also possible to put
antenna elements on a plane with a defined curvature. There are concave arrays, where
the beam is directed in a converging configuration. Also this arrangement can be used for
irradiating a target in near field. The opposite arrangement is convex. That is all antenna
elements are placed on a curvature which seems to diverge all antenna beams. Since the
phases can be adjusted, also this arrangement can be used to control a beam (transmit or
8.1 Radar, Broadcasting and Positioning 123

receive) in all directions. Preferably an arrangement like this can be used to operate from a
tower (almost like a lighthouse) to rotate a beam in full circle.

8.2 Smart Antenna Systems

By using the theory of the antenna arrays and by applying continuously developed Dig-
ital Signal Processing (DSP) techniques and Application Specific Integrated Circuits
(ASICs)[3], smart antenna systems can focus one or more beams to the desired users or
hotspots in a coverage area and achieve the best signal to interference and noise ratio (SINR),
thus optimizing the channel capacity.
In order to avoid interference, adaptive antenna array systems try to optimize the radiation
pattern, in real-time, to maximize the beam to the desired user or hotspot and to minimize
the beams from the direction of arrival (DOA) of the interferers [3]. This is also the basic
idea of the Zero-Forcing algorithm in the channel estimation in Massive MIMO, cf. Chap. 6.

Fig. 8.1 3-sector basestation antenna by using a Massive MIMO antenna array
124 8 Applications of Phased Arrays

Fig. 8.2 Basestation antenna by using a Massive MIMO antenna array as a cylindrical conformal
antenna array

8.3 5G Massive MIMO Basestation for Multi-User Coverage

Either the base station cells consist of classical 3 sectors with 3 antenna array panels (Fig. 8.1),
or a cylindrical conformal antenna array (Fig. 8.2) consisting of a very large number of
antenna elements, that can be used to provide multi-user or multi-hotspot coverage, in order
to improve the spectrum efficiency [4].

8.4 Deep Space Communications

In space communications, extremely large parabolic antenna apertures of 26 m, 34 m or even

70 m diameter are used. This is due to the extremely low signals because of the large distance
between the spacecrafts and the Earth groundstation [5–7]. Generally the antenna arrays are
thinned antenna arrays with distances between the single parabolic antennas much larger
than λ/2, as have been discussed in the previous chapters.
8.4 Deep Space Communications 125

Fig. 8.3 An example of a 2×2 deep space parabolic antenna array

In Fig. 8.3 one of these phased arrays is schematically illustrated. In principle these
phase arrays are thinned antenna arrays with extremely increased antenna gain for deep
space radio frequency communications. As mentioned before, the diameter of a parabolic
reflector antenna dish could be, for example, 34 m. The distances dx and d y could be 100 m
to some 1000 m, depending on the design criteria. In order to focus the beam, the parabolic
dishes are simultaneously steered to a certain direction. At the same time the phase shifts
between the different parabolic antennas should be precisely controlled so that all individual
signals add up in phase to and from a spacecraft.
Of course, depending on the missions, a larger number of parabolic antennas could be
applied, an example are the 14 operational antennas of Jet Propulsion Laboratory (JPL) in
Southern California, in order to track, send commands to, and receive the scientific data
from faraway spacecraft (Fig. 8.4).
One very famous antenna system is the Deep Space Network (DSN) of National Astro-
nomic Science Agency (NASA). To be accurate, the main task of a DSN is not a phased
array beam forming or steering, but rather to provide a redundant communications link con-
nection to the spacecrafts permanently during the rotation of the Earth. The DSN consists of
three antenna facilities spaced at equal distances from each other (about 120 degrees apart
in longitude) around the world, operated through the Network Operations Control Center at
JPL [6]:

• The Goldstone Deep Space Communications Complex near Barstow, California;

• The Madrid Deep Space Communications Complex near Madrid, Spain;
• The Canberra Deep Space Communication Complex near Canberra, Australia.
126 8 Applications of Phased Arrays

8.5 Radio Astronomy Antenna Array

Besides the deep space network for mission control, tracking and communications, also
a radio astronomy observatory antenna array will be used at National Radio Astronomy
Observatory in San Agustin in New Mexico [7]. It consists of up to 28 parabolic antennas
with a dish size of 25 m. By using this Very Large Antenna (VLA) with a spectral range
from 1.0 GHz – 50 GHz, cosmic radio waves can be detected. Different from the DSN this
is a phased array system with all parabolic antennas as antenna array elements.
ALMA is another powerful telescope for studying the Universe at submillimeter and
millimeter wavelengths, on the boundary between infrared light and the longer radio waves
[8]. However, ALMA does not resemble many people’s image of a giant telescope.
ALMA comprises 66 antennas, 54 of them with 12-meter diameter dishes, and 12 smaller
ones, with a diameter of 7 m each.
The dish surfaces are carefully controlled and the antennas can be steered very precisely
and pointed to an angular accuracy of 0.6 arcseconds (one arcsecond is 1/3600 of a degree).
This is accurate enough to pick out a golf ball at a distance of 15 km.

References

1. www.radartutorial.eu
2. A. B. Rohwer, D. H. Derosiers, W. Bach, H. Estavillo, P. Makridakis and R. Hrusovsky: Irid-
ium Main Mission Antennas – A phased array success story and mission update. 2010 IEEE
International Symposium on Phased Array Systems and Technology, pp. 504–511.(2010).
3. C. A. Balanis: Antenna Theory. John Wiley & Sons, Inc. Fourth Edition (2016).
4. Th. L. Marzetta, E. G. Larsson, H. Yang, H. Q. Ngo: Fundamentals of Massive MIMO. Cambridge
University Press (2016).

Fig. 8.4 An example of a 4×4 very large deep space parabolic antenna array
8.5 Radio Astronomy Antenna Array 127

5. NASA JPL: NASA Adds Giant New Dish to Communicate With Deep Space Missions. https://
www.jpl.nasa.gov/news/nasa-adds-giant-new-dish-to-communicate-with-deep-space-missions
6. NASA JPL: Deep Space Network. https://www.jpl.nasa.gov/missions/dsn
7. National Radio Astronomy Observatory in San Augustin in New Mexico. https://public.nrao.edu/
telescopes/vla/
8. European Southern Observatory. https://www.eso.org/public/ireland/teles-instr/alma/antennas
Conclusions
9

After an introduction into general antenna theory, based on Maxwell’s equations and the
vector algebra, general properties of antennas were derived. Then various different antenna
elements were described. Next phased array antennas were introduced. After these basic
explanations the focus was on simulation of radiation characteristics and their verification
by measurements of a linear array of patch antennas.
A comprehensive, systematic catalogue shows radiation characteristics for linear arrays
for various steering angles. Binomial and Chebyshev amplitude weights were employed to
perform beam forming. All these constellations are presented for far field and near field also.
It was shown, that an increasing steering angle generates higher side lobes and increases the
beam width a little and reduces the gain, as a minor effect, a little respectively.
In a next chapter 2-dimensional arrays were investigated and simulated. Antenna radiation
patterns are shown for different wanted beam steering angles. Also beam forming with
homogeneous and Chebyshev amplitude weighting, to achieve proper beam forms, that is
suppression of unwanted side lobes, was presented. For higher steering angles an asymmetric
amplitude weighting was explained and used successfully. Again radiation characteristics
are presented for far field and various near field constellations. Consequently conformal
arrays were considered next. First the results of a 1-dimensional array were considered,
that is concave and convex arrangements of patch antennas. Again there is a comprehensive
catalogue of the simulation and measurement results of antenna radiation patterns for various
steering angles and amplitude weights. Pattern results are given for far field, near field and
the special case, where the point of interest is in the focus point of the concave array. 1-
dimensional arrays are followed by 2-dimensional arrays. Again radiation characteristics
and properties were shown for various constellations.
All these arrays allow to steer and form a beam into a wanted direction in space. Concave
arrays may be used for precise irradiation of specimen whereas convex arrays are the ideal
foundation of 360◦ steering angle configurations, e.g. for communication systems. Then a
complete chapter was dedicated to MIMO antenna systems. To understand the implications

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 129
S.-P. Chen and H. Schmiedel, RF Antenna Beam Forming,
https://doi.org/10.1007/978-3-031-21765-4_9
130 9 Conclusions

and requirements for the antenna systems for different mobile communication systems,
communications technology aspects were explained comprehensively. Eventually thinned
arrays and their properties have been discussed with an emphasis on space and astronomical
applications.
Hopefully this book will contribute and help engineers, and students as future engineers,
to design new antenna array systems for given design specifications for new technical chal-
lenges.
Radio Frequency Bands

A.1 IEEE Frequency Bands

In the following tables, the often used IEEE frequency bands with special abbreviations,
frequency bands and wavelength ranges are shown (Tab. A.1).

Table A.1 IEEE Frequency bands

Abbreviation Frequency band Wavelength range
HF 0.003 GHz – 0.03 GHz 10 m – 100 m
VHF 0.03 GHz – 0.3 GHz 1 m – 10 m
UHF 0.3 GHz – 1 GHz 300 mm –1000 mm
L Band 1 GHz – 2 GHz 150 mm – 300 mm
S Band 2 GHz – 4 GHz 75 mm – 150 mm
C Band 4 GHz – 8 GHz 37.5 mm – 75 mm
X Band 8 GHz – 12 GHz 25 mm – 37.5 mm
Ku Band 12 GHz – 18 GHz 16.7 mm – 25 mm
K Band 18 GHz – 27 GHz 11.1 mm – 17 mm
Ka Band 27 GHz – 40 GHz 7.5 mm – 11.1 mm
V Band 40 GHz – 75 GHz 4 mm – 7.5 mm
W Band 75 GHz – 110 GHz 2.7 mm – 4 mm
mm Waves 100 GHz – 300 GHz 1 mm – 2.7 mm
THz Band 300 GHz – 3 THz 0.1 mm – 1 mm
IR Band 3 THz – 480 THz 0.625 µm –100 µm
Visible Light 480 THz – 680 THz 0.44 µm – 0.625 µm

© The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer 131
Nature Switzerland AG 2023
S.-P. Chen and H. Schmiedel, RF Antenna Beam Forming,
https://doi.org/10.1007/978-3-031-21765-4
132 Radio Frequency Bands

The IEEE frequency bands and wavelength ranges of the most used electromagnetic
waves in information and communication technologies, extended in this table with the ter-
ahertz THz band, infrared IR-band and visible light.

A.2 ITU Frequency Bands

In the following tables, the ITU frequency bands with special abbreviations, frequency bands
and wavelength ranges are shown (Tab. A.2).
ITU frequency bands and wavelength ranges of the most used electromagnetic waves in
information and communication technologies with the special abbreviations, the frequency
bands and wavelength ranges.

Table A.2 ITU Frequency bands

ITU No. Abbreviation Frequency Bands Wavelengths
1 ELF 3–30 Hz 100,000–10,000 km
2 SLF 30–300 Hz 10,000–1,000 km
3 ULF 300–3,000 Hz 1,000–100 km
4 VLF 3–30 kHz 100–10 km
5 LF 30–300 kHz 10–1 km
6 MF 300–3,000 kHz 1,000–100 m
7 HF 3–30 MHz 100–10 m
8 VHF 30–300 MHz 10–1 m
9 UHF 300–3,000 MHz 1–0.1 m
10 SHF 3–30 GHz 100–10 mm
11 EHF 30–300 GHz 10–1 mm
12 THz, THF 300–3,000 GHz 1–0.1 mm
Infrared 3 THz – 480 THz 0.625 µm – 100 µm
Visible Light 480 THz – 680 THz 0.44 µm – 0.625 µm
A.3 EU NATO Frequency Bands 133

A.3 EU NATO Frequency Bands

In the following tables, the EU NATO frequency bands with special abbreviations, frequency
bands and wavelength ranges are shown (Tab. A.3).

Table A.3 EU NATO bands

IEEE Band EU/NATO Band
L 1–2 GHz D 1–2 GHz
S 2–4 GHz E 2–3 GHz
C 4–8 GHz F 3–4 GHz
X 8–12 GHz G 4–6 GHz
Ku 12–18 GHz H 6–8 GHz
K 18–27 GHz I 8–10 GHz
Ka 27–40 GHz J 10–20 GHz
V 40–75 GHz K 20–40 GHz
W 75–110 GHz L 40–60 GHz
mm 110–300 GHz M 60–100 GHz
Terahertz 0.3–3 THz Terahertz 0.3–3 THz
Infrared 3–480 THz Infrared 3–480 THz
Visible Light 480–680 THz Visible Light 80–680 THz
Glossary

16QAM Quadrature Amplitude Modulation with 16 symbols

32QAM Quadrature Amplitude Modulation with 32 symbols
64QAM Quadrature Amplitude Modulation with 64 symbols
256QAM Quadrature Amplitude Modulation with 256 symbols
2G GSM, Global System of Mobile Communications
3G UMTS, Universal Mobile Telecommunication System
3GPP Third Generation Project Partnership, International Standardization Organization for
Mobile Communications
4G LTE, Long Term Evolution
5G 5th Generation Mobile Communication Networks
6G 6th Generation Mobile Communication Networks
AU Astronomical unit, distance between earth and sun, about 150 million km
BER Bit Error Rate
C Band 4 GHz – 8 GHz band
CSI Channel State Information, Cell-specific channel state measurement information
DAB Digital Audio Broadcasting
DL Downlink
DM-RS Demodulation Reference Signal
DMT Discrete Multi-Tone Technique
DSOC Deep Space Optical Communications
DSP Digital Signal Processing
DSN Deep Space Networks
DVB Digital Video Broadcasting
ELF Band Extremely Low Frequency Band
eNodeB evolved NodeB, LTE/5G Base Station
FDMA Frequency Division Multiple Access
ISI Inter-Symbol Interference
© The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer 135
Nature Switzerland AG 2023
S.-P. Chen and H. Schmiedel, RF Antenna Beam Forming,
https://doi.org/10.1007/978-3-031-21765-4
136 Glossary

ISM Band Open Access Frequency Bands for Industry, Science, and Medicine
ISO International Standardisation Organization
ITU International Telecommunication Union
ITU-T ITU Recommendations for Telecommunication Systems and Applications
ITU-R ITU Radiocommunication Standards
K Band 18-21 GHz band
Ka Band 27-40 GHz band
Ku Band 12-18 GHz band
L Band 1-2 GHz band
LF Band Low Frequency Band
LOS/LoS Line of Sight propagation condition or environment
LSE Least Square Errors
LTE Long Term Evolution, 4th Generation Mobile Communication Networks
LTE-A Long Term Evolution Advanced, 4th Generation Mobile Communication Networks
MBWA Mobile Broadband Wireless Access
MCS Modulation and Coding Scheme
MF Band Medium Frequency Band
MIMO Multiple Input Multiple Output
MISO Multiple Input Single Output
ML Maximum Likelihood
MMSE Minimum Mean Square Errors
M-RAT 5G Multiple Radio Access Technologies
MRRC Maximum Ratio Receive Combining
MSE Mean Square Errors
MU-MIMO Multiple User Multiple Input Multiple Output
NLOS Non Line of Sight propagation condition
PAPR Peak to Average Power Ratio
OFDM Orthogonal Frequency Division Multiplex
OFDMA Orthogonal Frequency-Division Multiple Access, LTE Downlink
PDSCH Physical Downlink Shared Channel
QAM Quadrature Amplitude Modulation
QPSK Quadrature Phase Shift Keying modulation technique
RB Resource Block
RE Resource Element
RS Reference Signal
S Band 2-4 GHz band
SC-FDMA Single Carrier Frequency Division Multiple Access, LTE Uplink
SHF Band Super High Frequency Band
SINR Signal to Interference and Noise Ratio
SISO Single Input Single Output
SIMO Single Input Multiple Output
Glossary 137

SLF Band Super Low Frequency Band

SNR Signal to Noise Ratio
STBC Space-Time Block Coding
SU-MIMO Single User Multiple Input Multiple Output
TDMA Time Division Multiple Access
THF Band Tremendously High Frequency Band
TMN Telecommunications Management Network
UHF Band Ultra High Frequency Band
UL Uplink
ULF Band Ultra Low Frequency Band
UMTS Universal Mobile Telecommunication System, 3rd Generation Mobile Communi-
cation Networks
UV Ultraviolet
VHF Band Very High Frequency Band
VLF Band Very Low Frequency Band
WLAN Wireless Local Area Network, also called WiFi
xDSL Digital Subscriber Line
ZF Zero Forcing algorithm for channel estimation