Modelling and Simulation of Matching Networks For Multi-Antenna Communication Systems

Modelling and Simulation of
Matching Networks for MultiAntenna Communication

Systems
OBADA ALHAJ MOUSSA
MUHAMMAD BILAL
Signal Processing Group
Department of Signals and Systems
CHALMERS UNIVERSITY OF TECHNOLOGY
EX022/2008
ABSTRACT
Future cellular systems such as MIMO (Multiple Input Multiple Output) will
use multiple antennas, both at the transmitter and receiver ends of the
communication links to boost the performance of the communication system.
When antennas are close to each other (in terms of wave length), mutual
coupling will arise having a negative effect on the performance of the antenna
system. In addition, a mismatch between the source and transmit antenna
from one side, and the load and the receive antenna from the other side
causes an undesirable power loss.
This master thesis assesses the impact of a matching network on the mutual
coupling effects between two adjacent antennas. A method to describe a
network in terms of scattering parameters is used to represent both of the
matching network and the antenna system. This allows for a physical model
where RF Blockset is used in the implementation. The front end of the MIMO
system is implemented in another master thesis [1] and a mathematical model
of the channel is available. The three models can be connected to form the
overall MIMO system chain.
Key words: Antenna arrays, Multiple Input Multiple Output (MIMO), mutual
coupling, network theory, S-parameters, impedance matching.
CHALMERS, Signal and System, Masters Thesis EX022/2008
ACKNOWLEDGMENT
This master thesis finalizes our post graduate studies at Chalmers University
of Technology. The work was done under the supervision of professor Mats
Viberg at the department of Signals and Systems. We would like to thank
professor Mats Viberg for his guidance and help throughout the course of this
thesis work. We would also like to thank the Charmant WP 1 participants for
their valuable presentations during the Charmant WP 1 meetings. Special
thank also goes out to Patrik Persson and Ulf Carlberg for their support,
answering questions, and providing input data to this master thesis.
II
Contents
ABSTRACT
ACKNOWLEDGMENT
II
CONTENTS
ANTENNA THEORY
1.1
Motivation
1.2
Basic principles of the antenna
1.3
Definitions
1.3.1
Radiation Pattern
1.3.2
Radiation Intensity
1.3.3
Directivity
1.3.4
Gain
1.3.5
Antenna efficiency
1.3.6
Input impedance
4
4
5
5
5
5
6
1.4
Antenna Arrays
1.5
Mutual coupling
2 DESCRIBING N-PORT NETWORK BY MEANS OF SCATTERING

PARAMETERS
10
2.1
Transmission line theory
10
2.2
Network Characterization
11
2.3
S-parameters representation
2.3.1
S-parameters measurement
2.3.2
Multi-port Networks
13
16
17
2.4
19
Describing N-port networks as a 2-port network
COMPLETE CHAIN OF MIMO SYSTEM USING S-PARAMETERS

3.1
Introduction
3.2
System Components
3.2.1
Signal Source
3.2.2
Transmit Antenna
3.2.3
Channel
3.2.4
Receive antennas
3.2.5
Signal Drain
21
21
21
21
22
23
24
25
3.3
Finding the overall channel matrix
3.3.1
Merging the inner components
3.3.2
Termination with Source and Drain
4
MATCHING NETWORKS
4.1
Introduction
25
25
29
32
32
4.2
Types of matching network
4.2.1
Characteristic impedance match
4.2.2
Self impedance match
4.2.3
Input impedance match
4.2.4
Multiport conjugate match
32
33
33
34
35
4.3
Hermitian match
35
4.3.1
Introduction
35
4.3.2
Singular Value Decomposition
35
4.3.3
Derivation of the S-parameters of the Hermitian matching
network 36
5
IMPLEMENTATION IN SIMULINK
5.1
Simulink and RF Blockset
41
5.2
Introduction to our system model
41
5.3
System Description
5.3.1
Introduction
5.3.2
Two port network
5.3.3
Graphical representation of a two port network
5.3.4
Graphical representation of a four port network
42
42
42
42
45
5.4
Representing Tx Antenna by a 4-port network
48
5.5
Representing Matching Network with RF Blockset
50
5.6
Combining Antennas and Matching Network
52
5.7
Data initialization
57
5.8
Results
57
CONCLUSION AND FUTURE WORK
APPENDIX A
41
62
63
A-1 RF blocks used

A-1.1 S-Parameters Passive Network
A-1.2 Input Port
A-1.3 Output Port
63
63
65
66
A-2
70
Converting to and from Simulink Signals
APPENDIX B
B-1
Algebraic Loops
APPENDIX C
C-1
Merging Two Consecutive Networks
REFERENCES
72
72
73
73
76
Antenna Theory
1.1 Motivation
In the past few decades, technology has developed at a rapid rate in the field
of mobile communications due to the introduction of new mobile
communication networks. The number of subscribers worldwide has
drastically risen up. The requirements on the antennas needed for the ever
expanding networks are becoming continually higher. The radiation patterns
of the antennas need to be strictly defined to allow accurate network
planning. In conclusion, the antenna becomes one of the major components
that need to be efficiently utilized to compromise the rapid expansion in the
communication networks and enhance the performance of the overall system.
In this chapter we will give a general insight into antenna theory providing a
definition to the main parameters that describe the performance of an
antenna. We will also introduce antenna arrays and focus on the mutual
coupling that occurs when the antennas in the array are close to each other (in
terms of wavelength).
1.2 Basic principles of the antenna

An antenna is a transducer designed to transmit or receive electromagnetic
waves. It transforms wire propagated waves into space propagated waves. It
receives electromagnetic waves and passes them onto a receiver or transmits
electromagnetic waves which have been produced by a transmitter.
Antennas are reciprocal that is, all of the antenna parameters that are
expressed in terms of a transmission antenna can be applied to a receiving
antenna. However, impedance is not applied in an obvious way; the
impedance at the load (where the power is consumed) is most critical. For a
transmitting antenna, this is the antenna itself. For a receiving antenna, this is
at the (radio) receiver rather than at the antenna.
1.3 Definitions
1.3.1 Radiation Pattern
An antenna radiation pattern or antenna pattern is defined as a mathematical
function or a graphical representation of the radiation properties of the
antenna as a function of space coordinates. In most cases, the radiation
pattern is determined in the far-field region and represented as a function of
the directional coordinates. Radiation properties include power flux density,
radiation intensity, field strength, directivity phase or polarization [2].
4
1.3.2 Radiation Intensity

Radiation intensity in a given direction is defined as the power radiated from
an antenna per unit solid angle [2].
1.3.3 Directivity
The directivity of an antenna is defined as the ratio of the radiation intensity
in a given direction from the antenna to the radiation intensity averaged over
all directions. The average radiation intensity is equal to the total power
radiated by the antenna divided by 4. If the direction is not specified, the
direction of maximum radiation intensity is implied [2].
1.3.4 Gain
Another useful measure describing the performance of an antenna is the gain.
Although the gain of the antenna is closely related to the directivity, it is a
measure that takes into account the efficiency of the antenna as well as its
directional capabilities.
Absolute gain of an antenna (in a given direction) is defined as the ratio of the
intensity, in a given direction, to the radiation intensity that would be
obtained if the power accepted by the antenna were radiated isotropically.
The radiation intensity corresponding to the isotropically radiated power is
equal to the power accepted (input) by the antenna divided by 4 [2].
In most cases we deal with relative gain, which is defined as the ratio of the
power gain in a given direction to the power gain of a reference antenna in its
reference direction [2]. The power input must be the same for both antennas.
The reference antenna is usually a dipole, horn or any other antenna whose
gain can be calculated or is known.
1.3.5 Antenna efficiency

The total antenna efficiency is used to take into account losses at the input
terminals and within the structure of the antenna. Such losses may be due to
1- Reflections because of the mismatch between the transmission line and the
antenna.
2- I 2 R Losses (conduction and dielectric).
1.3.6 Input impedance

Input impedance is defined as the impedance presented by an antenna at its
terminals or the ratio of the voltage to current at a pair of terminals or the
ratio of the appropriate components of the electric to magnetic fields at a
point [2].
1.4 Antenna Arrays

Usually the radiation pattern of a single element is relatively wide, and each
element provides low values of directivity (gain). In many applications it is
necessary to design antennas with very directive characteristics (very high
gains) to meet the demands of long distance communication. This can only be
accomplished by increasing the electrical size of the antenna.
Enlarging the dimensions of single elements often leads to more directive
characteristics. Another way to enlarge the dimensions of the antenna,
without necessarily increasing the size of the individual elements, is to form
an assembly of radiating elements in an electrical and geometrical
configuration. This new antenna, formed by multielements is referred to as an
array.
An important application where antenna arrays are useful is MIMO systems,
where antenna arrays are used to explore the spatial properties of the channel
and enhance the diversity order of the system. In most cases, the elements of
an array are identical. This is not necessary, but it is often convenient, simpler,
and more practical. The individual elements of an array may be of any form
(wires, apertures, etc.).
The total field of the array is determined by the vector addition of the fields
radiated by the individual elements. This assumes that the current in each
element is the same as that of the isolated element. This is usually not the case
and depends on the separation between the elements. To provide very
directive patterns, it is necessary that the fields from the elements of the array
interfere constructively (add) in the desired directions and interfere
destructively (cancel each other) in the remaining space. Ideally, this can be
accomplished, but practically it is only approached. In an array of identical
elements, there are five controls that can be used to shape the overall pattern
of the antenna. These are
1- The geometrical configuration of the overall array (linear, circular,
rectangular, spherical, etc.)
2- The relative displacement between the elements.
3- The excitation amplitude of the individual elements.
4- The excitation phase of the individual elements.
5- The relative pattern of the individual elements.
6
1.5 Mutual coupling

When the antennas in an array are placed close to each other, they will affect
each other through mutual coupling. Mutual coupling describes the
electromagnetic interaction between antenna elements [3]; it arises from the
fact that when one antenna is transmitting (or receiving) some of the radiated
energy will induce currents on the other antenna. The figure below illustrates
the mutual coupling between two antennas functioning in a receiving mode
Figure 1.1 Receiving mode coupling paths between antennas 1 and 2.

In the above figure an incident plane wave (1) strikes antenna (1) first where it
causes current flow (5). Part of the incident wave will be re-scattered into
space as (3). Some part of the scattering (3) will be directed towards antenna
(2) as (4) which will add with the incident wave (1) and then be received by
antenna(2) as (6). Antenna (1) therefore transfers some of its received energy
to antenna (2). In a similar way, antenna (2) transfers some of its received
energy to antenna (1). This phenomenon is called mutual coupling.
Mutual coupling can be described with impedance and scattering matrices.
An example of how mutual coupling between two antennas can be described
with an impedance matrix is shown below
V1 Z 11
=
V2 Z 21
Z 12 I 1

Z 22 I 2
(1.1)
where
V1 and V2 are the voltages at antenna 1 and 2 respectively

I1 and I 2 are the currents at antenna 1 and 2 respectively
Z11 and Z 22 are the self impedances on antenna 1 and 2 respectively

Z 21 and Z12 are the mutual impedances. They describe how the current on
one antenna gives rise to voltage on the other antenna.
Working out the above matrix equation we get
V1 = Z11I1 + Z12 I 2
V2 = Z 21I1 + Z 22 I 2
(1.2)
The main effects of mutual coupling are

1. Distortion of the effective radiation pattern of each antenna. An
example of how the mutual coupling distorts the radiation pattern for
two half wave dipoles in free space is shown in Figures 1.2 and 1.3 [4].
Figure 1.2 shows the radiation pattern when there is one isolated half
wave dipole in free space, while figure 1.3 shows the radiation pattern
when there are two half wave dipoles that are in close presence of each
other.
Figure 1.2 Far field radiation pattern for one isolated half wave dipole.
Figure 1.3 Far field radiation pattern for a half wave dipole when there are two half
apart.
wave dipoles placed side-by-side with distance
2
2. The mutual coupling between antennas also changes the correlation of

the signal levels at each antenna. The signal correlation describes how
much the signals at the antennas depend on each other. The more
correlated the more dependent.
3. The radiation efficiency of the antenna is also affected by the mutual

coupling. Radiation efficiency of an antenna is the ratio of the power
radiated to the total power supplied to the antenna at a given
frequency.
Describing N-port network by means

of scattering Parameters
In this chapter, we begin with a short statement of the transmission line

theory. Then we give a brief review about the traditional methods of network
characterization (H, Y, and Z-parameters) to emerge into the desired Sparameters characterization.
2.1 Transmission line theory

A transmission line is the material medium or structure that forms all or part
of a path from one place to another by directing the transmission of energy,
such as electromagnetic waves or acoustic waves as well as electric power
transmission [5].
In many electrical circuits, the length of the wires connecting the components
can for the most part be ignored. That is, the voltage on the wire at a given
time can be assumed to be the same at all points. However, when the voltage
changes in a time interval comparable to the time it takes for the signal to
travel down the wire, the length becomes important and the wire must be
treated as a transmission line. Stated another way, the length of the wire is
important when the signal includes frequency components with
corresponding wave lengths comparable to or less than the length of the wire.
For the purpose of analysis, an electric transmission line can be modelled as a
two port network as shown in the figure below
Figure 2.1 Electrical transmission line modelled as a two port network.
In the simplest case, the network is assumed to be linear (i.e. the complex
voltage across either port is proportional to the complex current flowing into
it when there are no reflections), and the two ports are assumed to be
interchangeable. If the transmission line is uniform along its length, then its
behaviour is largely described by a single parameter called the characteristic
10
impedance ( Z 0 ). This is the ratio of the complex voltage of a given wave to the
complex current of the same wave at any point on the line.
When sending power down a transmission line, it is usually desirable that as
much power as possible will be absorbed by the load and as little as possible
will be reflected back to the source. This can be ensured by making the source
and load impedances equal to Z 0 , in which case the transmission line is said to
be matched.
2.2 Network Characterization

Networks are usually characterized by their H, Y or Z-parameters. However,
at high frequencies another method of characterization (S-parameters) is
useful. We will first describe the traditional methods of characterization then
we will see how a network can be characterized in terms of its S-parameters.
Consider the network in the figure below
Figure 2.2 Two port network.

This two port network can be described by a number of parameter sets (Zparameter, Y-parameter and H-parameter). All of these network parameters
relate total voltages and total currents at each of the two ports in the following
way
H y b r id ( H P a r a m e te r s )
V 1 = H 1 1 I 1 + H 1 2V 2
(2.1)
I 2 = H 2 1 I 1 + H 2 2V 2
A dm ittance ( Y P aram eters )

I 1 = Y11V1 + Y12V 2
(2.2)
I 2 = Y21V1 + Y22V 2
11
Impedance ( Z Parameters )
V1 = Z 11 I 1 + Z 12 I 2
(2.3)
V2 = Z 21 I 1 + Z 22 I 2
These parameters can be measured by applying open circuit or short circuit to

one of the ports of the network and measuring the ratio between voltage and
current on the other port. For example, H11 can be measure by setting V2 to
zero (applying a short circuit to the output port of the network). H11 is then
the ration of V1 to I1 (the input impedance of the resulting network). Further,
H12 is determined by measuring the ratio of V1 to V2 (the reverse voltage gain
with the input port open circuited). All other parameters can be measured in
similar way.
Moving to higher and higher frequencies some problems arise:
1. Equipment is not readily available to measure total voltage and total
current at the ports of the network.
2. Short and open circuits are difficult to achieve over a broad band of
frequencies.
3. Active devices, such as transistors and tunnel diodes, will very often
not be short or open circuit stable.
Some method of characterization is necessary to overcome these problems.
The logical variables to use at high frequencies are travelling waves rather
than voltages and currents, and the network is described in terms of Sparameters rather than Z, H, or Y-parameters.
High frequency systems have a source of power. A portion of this power is
delivered to a load by means of a transmission line. The physical components
of such a system is shown in the figure below
Figure 2.3 Physical components of a high frequency system.
12
Figure 2.4 below shows the electric circuit corresponding to figure 2.3
Figure 2.4 Electric circuit corresponding to the physical components in Fig. 2.3
Voltage, current and power can be considered to be in the form of waves
travelling in both directions along this transmission line. A portion of the
wave incident on the load will be reflected. It then becomes incident on the
source, and in turn re-reflects from the source (if Z S Z 0 ), resulting in a
standing wave on the line.
The value of this total voltage at a given point along the length of the
transmission line is the sum of the incident and reflected waves at that point.
The total current on the line is the difference between the incident and
reflected voltage waves divided by the characteristic impedance of the line.
The voltage and current equations are given below
Vt = V + + V
(2.4)
V + +V
It =
Z0
(2.5)
2.3 S-parameters representation

To introduce the S-parameter representation we insert a two-port network
into the transmission line in Fig. 2.4:
13
Figure 2.5 Two port network connected into a source and a load through a
transmission line. Additional travelling waves arise at the transmission line due to
the insertion of the two port network.
We now have additional travelling waves that are interrelated. Looking at V2 ,
we see that it is made up of that portion of V2+ reflected from the output port
of the network as well as that portion of V1+ that is transmitted through the
network. Each of the other waves is similarly made up of a combination of
two waves.
The voltages and currents at the two ports are given in the equations below
V1 = V1+ + V1
(2.6)
V2 = V2+ + V2
(2.7)
V1+ V1_
I1 =
Z0
(2.8)
V2+ V2
I2 =
Z0
(2.9)
It is possible to relate these four travelling waves by some parameter set.

While the derivation of this parameter set will be made for two-port
networks, it is applicable for N-ports as well. Lets start with the H-parameter
set
14
H Parameters
V 1 = H 11 I 1 + H 12 V 2
I2 = H
21
I1 + H
22
(2.10)
V2
By substituting the expressions for total voltage and total current on a

transmission line into this parameter set, we can rearrange these equations
such that the incident travelling voltage waves are the independent variables;
and the reflected travelling voltage waves are the dependent variables
V1 = F11 ( H )V1+ + F12 ( H )V 2+

V 2 = F21 ( H )V1+ + F22 ( H )V 2+
(2.11)
The functions F11 , F21 and F12 , F22 represent a new set of network parameters
relating travelling voltage waves rather than total voltages and total currents.
In this case these functions are expressed in terms of H-parameters. They
could have been derived from any other parameter set.
It is appropriate to call this new parameter set scattering parameters since
they relate those waves scattered or reflected from the network to those waves
incident upon the network. These scattering parameters are referred to as Sparameters. If we divide both sides of these equations by Z 0 , the
characteristic impedance of the transmission line, the relationship will not
change. It will, however, give us a change in variables. The new variables are
defined in the equations below
V1+
a1 =
Z0
(2.12)
V2+
a2 =
Z0
(2.13)
V1
b1 =
Z0
(2.14)
V2
b2 =
Z0
(2.15)
15
Notice that the square of the magnitude of these new variables has the
2
dimension of power. Therefore, a1 can be thought of as the incident power

2
on port one and b1 as power reflected from port one. These new waves are
called travelling power waves rather than travelling voltage waves. Using
(2.11-2.15) and doing appropriate substitutions, we see that the S-parameters
relate these four waves in this fashion
b1 = S11 a1 + S12 a 2
b2 = S 21 a1 + S 22 a 2
(2.16)
Equation 2.16 can be written in the form of a matrix equation as
b1 S11
b = S
2 21
S12 a1
S 22 a2
(2.17)
2.3.1 S-parameters measurement

We have seen how a network is described in terms of its scattering
parameters; now lets see how these scattering parameters are measured.
Figure 2.6 shows a two port network connected to a source and a load
through a transmission line where, S11 represents the input reflection
coefficient of the network, S 22 represents the output reflection coefficient, S 21
represents the transmission from port 1 to port 2 and S12 represents the
transmission from port 2 to port 1. Furthermore, a1 and a2 are the input
travelling power waves, while b1 and b2 are the output travelling power
waves. Finally, Z 0 is the characteristic impedance of the transmission line,
while Z L and Z S are the impedances of the load and source respectively. The
output travelling power waves can be written in terms of the input travelling
power waves and the scattering parameters as shown in (2.16).
Figure 2.6 Two port network connected into a source and a load through a
transmission line. The two port network is described in terms of its scattering
parameters.
16
If we terminate the output port of the network in an impedance equal to the

characteristic impedance of the transmission line ( Z L = Z 0 ), all the power
incident to the load will be totally absorbed. This is equivalent to setting
a2 = 0 . Going back to (2.16), S11 is then measured as the ratio between b1 and
a1 , while S 21 is measured as the ratio between b2 and a1 . To measure S 22 and

S12 we terminate the input port of the network by a matched load ( a1 = 0 ),
and then S 22 is measured as the ratio between b2 and a2 , while S12 is the ratio
between b1 and a2 .
S11 =
S 21 =
S 22 =
S12 =
b1
a1
b2
a1
b2
a2
b1
a2
(2.18)
a 2=0
(2.19)
a 2=0
(2.20)
a1= 0
(2.21)
a1= 0
2.3.2 Multi-port Networks

So far we have just discussed two-port networks. These concepts can be
expanded to multiple-port networks. The three port network in figure 2.7, for
example, can be characterized by the following equation
b1 S11
b = S
2 21
b3 S 31
S12
S 22
S 32
S13 a1
S 23 a 2
S 33 a3
(2.22)
17
The above matrix equation can be written out to give
b1 = S11a1 + S12 a2 + S13 a3

b2 = S 21a1 + S 22 a2 + S 23 a3
(2.23)
b3 = S 31a1 + S 32 a2 + S 33 a3
Figure 2.7 Three port network.
In a similar manner as in two port networks, S11 is measured by terminating

the second and third ports with a matched load. This again ensures
that a 2 = a3 = 0 . S11 is then given by (2.24). We could go through the
remaining S-parameters and measure them in a similar way, once the other
two ports are properly terminated.
S11 =
b1
a1
(2.24)
a 2 = a 3= 0
What is true for two and three port networks is similarly true for N-port
networks.
18
2.4 Describing N-port networks as a 2-port network

Consider the N-port network in the figure below
Figure 2.8 N port network.
This network has P ports from one side and Q ports form the other side where
P+Q=N. The scattering matrix of this network is defined by the relation
b1 S11
b S
2 21

=

bN S N 1
S12
S 22
SN 2
S1N a1
S 2 N a2

S NN a N
If we group together the ports at each side of the network, we can look at the
network as a 2-port network with the following scattering matrix equation
b1P1 S11P P
Q1 = Q P
b S
2 21
S12PQ a1P1
QQ Q1
S 22 a 2
(2.25)
Each of the scattering parameters in (2.25) is a matrix. The dimension of the

matrix is given by the superscript of the corresponding scattering parameter.
For example, S12PQ is a matrix with P rows and Q columns.
The parameters a1Px1 and a2Qx1 are the inward propagating wave vectors of port
group 1 (P ports) and 2 (Q ports) respectively, while b1Px1 and b2Qx1 are the
outward propagating wave vectors. The sub matrices S ii represent the
19
reflection of the wave vector ai to bi , whereas Sij , i j represents the

transmission from a j to bi . The voltages at the port group K is given by
VK = Z 0 (a K + bK )
(2.26)
where
Z 0 represents the diagonal characteristic impedance matrix of port group K.
Equation 2.26 is called the heavyside transformation.
From now on we will use the above representation for any network.
20
Complete Chain of MIMO system

using S-parameters
3.1 Introduction
The main components of a MIMO system are: the signal source, the transmit
antenna, the channel, the receive antenna and the signal drain. All these
components are described in terms of their scattering parameters. They are
connected together to form the complete MIMO system chain as shown in the
figure below
Figure 3.1 System model of the complete RF transmission chain. All elements are
described by scattering matrices.
We will first describe each of these components alone, then we will see how
the system components are merged and the overall channel matrix is
produced.
3.2 System Components

3.2.1 Signal Source
The signal source is depicted in the figure below
Figure 3.2 Signal source, stand for the amplifier in our model where S corresponds
to S 22 of the amplifier.
21
The signal source is the beginning of the transmission chain and determines
the transmit power distribution among the transmit antennas. It has the same
number of ports as the transmit antenna. The output impedances of the signal
source are characterized by the reflection coefficient s . The output
propagating wave vector is given as
a1TA = b0 + s b1TA
(3.1)
3.2.2 Transmit Antenna

The antenna array plays a critical role in MIMO systems. Since the antennas
will be closely spaced they will couple and interact, and they cant be
considered as separate elements. The main effects which have to be taken into
account are
1. The shape of the radiation pattern of the single antennas changes due
to the effect of the other elements.
2. Since the energy radiated by one antenna may be directly absorbed by
another antenna due to the close separation, the gain of the antenna
reduces. Thus the gain of the antenna is a property of each individual
antenna in the array and depends on the array topology and
termination of the antenna.
These effects are included in the general scattering parameter description of
the antenna.
An antenna array of M elements can be seen as 2M-port network. By grouping
the ports at the two sides of the network, the antenna array can be
represented by 2-port network as shown in the figure below
Figure 3.3 Transmit antenna described as a two port network.
22
where
Tx
A
S 11TA
= TA
S 21
S 12TA
TA
S 22
(3.2)
Ports group 1 represent the excitation ports of the antennas in the array while
ports group 2 represents the far field properties.
S11TA describes the excitation ports of the antennas and contains the scattering
parameters that represent the self (diagonal elements) and mutual coupling
(off diagonal elements) effects of the array.
TA
S 21
and S12TA describe the transmission of the signals from the excitation side to
the far field and vice versa, thus they contain information on the pattern and
gain of the antenna.
TA
S 22
contains the structural antenna scattering of the array with the excitation
ports of the antennas terminated in matched loads.
TA
TA
The elements of S 21
, S12TA and S 22
are directional, which means that they are
function of the geometry and change for different angles of arrival and
departure for different paths.
3.2.3 Channel
The communication channel is also modelled as a 2-port network as shown in
the figure below
Figure 3.4 Communication channel described as a two port network.

Port 1 represent the far field ports of the transmitting antenna array while
port 2 represents the far field ports of the receiving antenna array. The
relationship between these ports is given by the scattering matrices in (3.3)
23
b 1C
C
b
2
S 11C
= C
S
21
S 12C
S 22C
a 1C
C
a
2
(3.3)
Assuming that only far-field effects are of interest, the coupling disappears by
definition. Note that the coupling between single antennas is included in the
scattering matrices for the antennas. Additionally, it is assumed that there is
C
no reflection from the far fields. Thus, the submatrices S11C and S 22
can be set
to zero.
Moreover, the back transmission of signals through the physical channel is
subjected to the channel attenuation; thus, the power reradiated by the
receiver and received by the transmitter is twice as strongly attenuated as the
signals at the receiver. Therefore, it is justified to neglect the back
transmission and set S12C to zero.
C
In summary, the communication channel is modelled by S 21
alone, and (3.3)
C
C
C C
can be written as b1 = 0 , b2 = S 21a1
3.2.4 Receive antennas

The receive antennas can be modelled as
Figure 3.5 Receive antenna described as a two port network.

where
Rx
A
S 11RA
= RA
S 21
S 12RA
S 22RA
(3.4)
S11RA contains the structural antenna scattering of the array with the excitation
ports of the antennas terminated in matched loads.
24
S22RA describes the excitation ports of the antennas and contains the scattering
parameters that represent the self (diagonal elements) and mutual coupling
(off diagonal elements) effects of the array.
S12RA and S 21RA describe the transmission of the signals from the excitation side
to the far field and vice versa.
3.2.5 Signal Drain

The signal drain is similar to the signal source and is described by the
reflection coefficient D .
3.3 Finding the overall channel matrix

Figure 3.1 shows the five system components attached together. The overall
channel matrix of the system in figure 3.1 is found in two steps as shown in
sections 3.3.1 and 3.3.2.
3.3.1 Merging the inner components

The transmit antenna, channel and receive antenna are merged into one
network as shown in figure 3.6. The scattering parameters of the new network
are function of the scattering parameters of the 3 merged components.
Figure 3.6 Transmit antenna, communication channel and receive antenna are
merged into one network.
The transmit antenna is first connected with the communication channel as

shown in figure 3.7
25
Figure 3.7 Transmit antenna connected to the communication channel.
Following the same procedure as in Appendix C, the above two networks can
be merged into one network with new scattering parameters as shown in the
below
Figure 3.8 Network resulted from merging the Transmit antenna and the
Communication channel.
Using equations (C-6 C-9) in Appendix C we get

TA
S11TAC = S11TA + S12TA ( I S11C S 22
) S11C S21TA
1
= S (I S S
C TA 1
11 22
S12C
(3.6)
TAC
C
TA C
TA
S21
= S21
( I S22
S11 ) S21
(3.7)
TAC
C
C
TA C
TA C
S22
= S22
+ S21
( I S 22
S11 ) S22
S12
(3.8)
TAC
12
TA
12
26
(3.5)
C
Recalling that from section 3.2.3 S11C = S 22
= 0 and substituting in the above
equations we get
S11TAC = S11TA
(3.9)
S12TAC = S12TA S12C
(3.10)
TAC
C TA
S21
= S21
S21
(3.11)
TAC
C TA C
S22
= S21
S22 S12
(3.12)
The resulting network is then connected with the receive antennas as shown
in figure below
Figure 3.9 Receive antenna connected to the network resulted from merging the
transmit antenna and the communication channel.
The above two networks can also be merged into one new network as shown
in the figure below
Figure 3.10 Network resulted from merging the Transmit antenna, communication
channel and the Receive antenna.
27
Using (C-6 C-9) in Appendix C we get

TAC
S11H = S11TAC + S12TAC ( I S11RA S 22
) S11RA S 21TAC
1
=S
TA
11
+ S S (I S
TA
12
C
12
RA
11
C
21
TA
22
S S S
TAC
S 12H = S 12T A C ( I S 11R A S 22
C
12
RA
11
C
21
S S
(3.13)
TA
21
S 12R A
C
TA C
= S 12T A S 12C ( I S 11R A S 21
S 22
S 12 ) S 12R A
1
TAC RA
TAC
S 21H = S 21RA ( I S 22
S11 ) S 21
(3.14)
=S
RA
21
(I S S S S
C
21
TA
22
C
12
RA 1
11
S S
C
21
(3.15)
TA
21
TAC RA
TAC RA
S 22H = S 22RA + S 21RA ( I S 22
S11 ) S 22
S12
1
=S
RA
22
+S
RA
21
(I S S S S
C
21
TA C RA 1
22 12 11
S S S S
C
21
(3.16)
TA C RA
22 12 12
Recalling that from section 2.3.3 S12C = 0 and substituting it in the above
equations we get
S11H = S11TA
(3.17)
S12H = 0
(3.18)
C TA
S 21H = S 21RA S 21
S 21
(3.19)
S22H = S22RA
(3.20)
The scattering parameters matrix of the resulting network is then given by
28
=
S
S 11TA
RA
21
C
21
0
S
TA
21
RA
22
(3.21)
3.3.2 Termination with Source and Drain

The source and the Drain are connected to the network in figure 3.10 as
shown in the figure below
Figure 3.11 Transmit antenna, communication channel and Receive antenna are
merged into one network. Source and Drain terminate the network.
Applying equations (C-14 C-17) in Appendix C we get
a1TA = ( I s S11H ) b0, s

1
(3.22)
a2RA = ( I D S 22H ) D S 21H ( I s S11H ) b0, s
(3.23)
b1TA = ( I S11H s ) S1H1 b0, s
(3.24)
b2RA = ( I S 22H D ) S 21H + S 21H s ( I S11H s ) S11H b0, s

1
(3.25)
The overall channel matrix expresses the ratio of the voltages at the receive
antennas to the voltages at the transmit antennas as
V RA = H
ext
V TA
(3.26)
29
where V is given by the heavyside transformation as
V = Z 0 (a + b )
(3.27)
Substituting in (3.26) we get
Z 0, RA a2RA + b2RA = H ext a1TA + b1TA
Z 0,TA
(3.28)
Using (3.22-3.25) and performing appropriate substitutions, the overall

channel matrix is found as
H ext = Z0,RA (I + D ) I S22H D
S21H I + S11H
Z0,TA
(3.29)
C TA
The term S 21H = S 21RA S 21
S 21 describes the transmission of the signal from the
input ports of the transmit antennas to the output ports of the receive
antennas. Using the heavyside transformation it can be expressed as the ratio
of the voltages V at the mth transmit antenna and nth receive antenna as
S 21H
nm
Z 0, m Vn
Z 0, n Vm
Km
(3.30)
ak =0
where Z 0 is the characteristic impedance of the scattering parameters.

Vn
can take different forms depending on the transmission channel
Vm
representation. In all cases it is mainly a function of the gain and pattern of
the antennas in use.
The ratio
S11H = S11TA and S 22H = S 22RA contain the self and mutual coupling effects of the
transmit and receive antennas, while D is the reflection coefficient of the

signal drain.
In this master thesis work, it was decided that the effects of S 22H , S11H and D
can be modelled physically using the RF toolbox in Simulink, while S 21H can be
conveniently represented by a mathematical model. The availability of
30
Simulink boxes that serve as an interface between a physical and a

mathematical model makes our decision feasible.
Since mutual coupling depends on the termination of transmit and receive
antennas, a matching network can be inserted between the signal source and
transmit antenna from one side and the receive antenna and the signal drain
from the other side.
The overall system will then look like
Figure 3.12 Complete RF transmission chain with matching networks inserted

between the source and transmit antenna from one side and the receive antenna and
the load from the other side.
The matching network effects are described in the next chapter.
31
Matching networks
4.1 Introduction
The Matching network becomes one of the main components of a multiple
input multiple output (MIMO) system for many reasons. A MIMO system
makes use of multiple antennas at both the transmitter and receiver ends to
exploit the spatial channel for increasing the capacity. Correlation of the
signals at the different antenna elements can considerably decrease the
capacity of a MIMO system [6]. Such correlation occurs particularly when the
separation between the antennas is small [7]. In addition, for a small
separation, the effect of mutual coupling between the antennas becomes
important. Mutual coupling distorts antenna patterns and therefore modifies
the correlation results [8], [9]. The change in input impedances of the antennas
is another consequence of mutual coupling, and it results in greater mismatch
between the antennas and their corresponding source and load impedances
[10].
The matching network has a significant influence on the performance of
multiple antenna system when the antennas are close to each other. The
purpose of the matching network is to minimize the effect of mutual coupling
and improve the efficiency of the antennas by maximizing the power transfer
to the load.
The use of S-parameter representation to model an entire Multi input Multi
output (MIMO) communication system was proposed in the previous
chapter. Using this approach, a more diverse range of matching networks can
be studied.
4.2 Types of matching network

This section introduces four different types of impedance matching networks
with their respective S-parameter representations. The S-matrix of the
matching network is represented by
32
S 11M
= M
S 21
S 12M
M
S 22
(4.1)
in the following , we will discuss some different options for choosing S M .
4.2.1 Characteristic impedance match

The characteristic impedance match is the simplest form of impedance
matching, where the antennas are terminated with a load of characteristic
impedance Z 0 , which is usually taken as 50 ohm. In other words, there is no
matching network. This can be modelled either by removal of the matching
network or by setting S11M = S 22M = 0 and S 21M = S12M = I . The degree of mismatch
depends on the difference between the antenna impedances and the
characteristic impedance.
4.2.2 Self impedance match

The Self-impedance match refers to terminating each antenna with a load
( ).
A
equivalent to the conjugate of the antenna self-impedance that is, Z L = Z sf
In S-parameter form this match can be expressed as
M
11
= Z S
diag Z A *
( sf )
(4.2)
where Z S indicates the transformation from Z to S parameter and Z sfA is the

self impedance of a receiving antenna.
For an isolated antenna, the self-impedance match is also known as the
complex conjugate match [11]. It facilitates maximum power transfer to the
load when there is no mutual coupling that is, the array antennas are
infinitely far apart. However at finite antenna separations, the goodness of the
match depends on the behavior of the mutual impedance which is not taken
into account.
33
4.2.3 Input impedance match

While the self-impedance match only takes into account the self-impedance of
the antenna, the input impedance match also takes into account mutual
coupling [12], [13]. The input impedance match attempts to conjugate-match
the antenna pair individually (one at a time) that is, there is a separate
matching network for each port, and the matrix of the matching network is
diagonal.
For a two antenna array, the input impedance of antenna 1 can be expressed
as
Z in = Z 11
Z 12 Z 21
Z 22 + Z L
(4.3)
where Z L is the load on the antennas. For input impedance match, Z L
= Z in*
or
Z Z
Z L = Z 11 12 21
Z 22 + Z L
(4.4)
The above equation is iterated to find the desired matching impedance Z L . A

closed form solution of the above equation can also be found, using some
straightforward algebraic manipulations. For the simple symmetrical twodipole case ( Z11 = Z 22 , Z12 = Z 21 ), this solution is given by
ZL =
+d
c 2d
c
a2
2
cd
+ j
b
a
(4.5)
where
Z 11 = a + jb
Z 12 = c + jd
34
(4.6)
Thus, S11 of the matching network can, in the two-antenna case, be written as
Z
S 11M = Z S L
0
Z L
(4.7)
4.2.4 Multiport conjugate match

Like the input impedance match, the so-called multiport conjugate match [14]
(also called Hermitian match) also takes into account the mutual coupling
among the antenna ports. However, unlike the input impedance match, it
allows the interconnections between all ports on the two sides of the network.
The Hermitian match condition will be discussed in more details in the next
section.
4.3 Hermitian match

4.3.1 Introduction
The Hermitian match is a simplified proof of the multiport conjugate match
which states that , for an antenna array with mutual couplings terminated by
a load, the maximum power can be transferred to the load if Z L = Z A , where
Z L is the impedance matrix of the load and Z A is the conjugate of the
impedance matrix of the antenna array [15]-[18]. The Hermitian match makes
use of singular value decomposition to provide the proof which is stated in
sections 4.3.3.1 and 4.3.3.2.
4.3.2 Singular Value Decomposition

The singular value decomposition (SVD) is an important factorization of a
rectangular real or complex matrix, with several applications in signal
processing and statistics. The spectral theorem [19] says that normal matrices,
which by definition must be square, can be unitarily diagonalized using a
basis of eigenvectors. The SVD can be seen as a generalization of the spectral
theorem to arbitrary, not necessarily square, matrices.
35
Suppose M is an m-by-n matrix whose entries come from the field K, which is
either the field of real numbers or the field of complex numbers. Then there
exists a factorization of the form
1
2
M = U V H
(4.8)
where
U is an m-by-m unitary matrix over K which contains a set of orthonormal

output basis vector directions for M
1
2 is an m-by-n matrix with nonnegative numbers on the diagonal and zeros

off the diagonal containing the singular values, which can be though of as
scalar gain controls by which each corresponding input is multiplied to give
a corresponding output.
V H denotes the conjugate transpose of V , an n-by-n unitary matrix over K
that contains a set of orthonormal input basis vector directions for M.

Such a factorization is called a singular-value decomposition of M.
4.3.3 Derivation of the S-parameters of the Hermitian matching

network
The figure below depicts a receiving condition of a MIMO system, with the
matching network inserted between the receiving antennas and the load
Figure 4.1 Coupled receiving antenna array connected to a multiport matching

network and individual loads.
36
For high-frequency systems such as mutually coupled antenna networks, the

S-parameter matrix representation provides a convenient analysis framework.
Each element of the coupled array is characterized by a generator whose
signal passes through the matrix S RA with a block representation
S RA
S12RA
RA
S 22
S11RA
= RA
S 21
(4.9)
where 1 and 2 refer to input and output ports, respectively.

The input and output reflection coefficients in Fig. 4.1 can be expressed as
in = S11M + S12M ( I S L S 22M ) S L S 21M

1
= S + S SL ( I SL S
M
11
M
12
M
22
(4.10)
M
21
out = S22M + S21M ( I S22RA S11 ) S22RA S12M

1
( )
If the matching network is lossless, S M
(S )
M
11
(S )
M
11
(S )
M
12
(4.11)
S M = I , which results in
S11 + ( S 21M ) S 21 = I
(4.12)
S12 + ( S 21M ) S22 = 0
(4.13)
S12 + ( S 22M ) S22 = I
(4.14)
1
2
ij
Substitution of the SVD of the subblocks S ij = U ij Vij into the above

H
equations yields
V21 21 = V11
(4.15)
V12 12 = V22
(4.16)
37
21 = I 11
(4.17)
12 = I 22
(4.18)
where 21 and 12 are diagonal phase shift matrices with arbitrary complex
elements of unit magnitude. This operation also produces the condition
1
2
11
U U12 (I 22 ) 12 = (I 11 ) U U 22
1
2
H
11
H
21
1
2
H
21
1
2
22
(4.19)
The SVD proposed above facilitates a simple proof of the fact that multiport
( )
conjugate match S11M = S 22RA
results in maximum power transfer to the load.
In the next section we will provide an abbreviated version of the proof.
4.3.3.1 Load Match

To begin the derivation we remove the source coupling block in Fig. 4.1 to
arrive at the network in the figure below
Figure 4.2 Set up for the load matching problem. To ensure maximum power transfer,
we must find S M such that in = 0 .
The power delivered to the load will be maximized for any possible excitation
( a1 ) if and only if we can find S M for a lossless network such that in = 0 . We
need to show that S 22M = S LH is sufficient to get in = 0 for arbitrary excitation.
The below SVD representations for S 22M and S LH are used for this purpose
38
S 22M = S LH
M
11
= U L L2 VL
(4.20)
1
H
2
11 11
(4.21)
= U 11 V
Using the constrains in (4.15-4.18) coupled with (4.10) leads to

1
1
1
1
in = U 11112 + U 12 (I L ) 2 L2 12VLH U 21 (I 11 ) 2 21 V11H
(4.22)
(4.19) can be re-written for this problem as
(I L )
1
2
1
2
L
1
2
11
V U 21 (I 11 ) 21 = U U11
1
2
H
12 L
H
12
(4.23)
which upon substitution into (4.22) results in
in = U 11 11 U 12 U 12H U 11 11 V 11H = 0
(4.24)
Therefore, choosing S 22M = S LH is sufficient to ensure that in = 0 for arbitrary

excitation. Next we show that in = 0 results in S 22M = S LH , indicating that this
later condition is also necessary for maximizing power transfer. Multiplying
( )
4.10 by S12M
assuming in = 0 and making appropriate substitutions from
the lossless conditions leads to
( S 22M ) + I ( S 22M ) S 22M S L ( I S 22M S L )

H
) S
M
21
=0
(4.25)
) M S
=0
(4.26)
Let
M = ( S22M ) + I ( S 22M ) S22M S L ( I S22M S L )

H
M
21
39
Since for a lossless matching network S 21M must be full rank [16], the above
equation can be only satisfied if M = 0 , which after simplification yields to
S 22M = S LH .
4.3.3.2 Source Match

We now consider the problem of matching the arbitrary coupled source
element in Fig. 4.1 to a set of uncoupled identical loads of impedance Z 0 . In
this case S L = 0 such that a 2 = 0 and in = S11M . If we collapse the matching
network and load network into a single network block, then our equivalent
network is in the form of the figure below
Figure 4.3 Set up for the source matching problem. To ensure maximum power
transfer, S11M must equal to ( S 22RA ) .
H
Based upon our work above and since the matching network is lossless, we
know that all available power will be transferred to the loads if and only if
( )
S11M = S 22RA
[20].
Knowing S11M and S 22M , S12M and S 21M can be obtained from (4.12) and (4.14) using
singular value decomposition. There are infinitely many solution of S12M and
S 21M that satisfy (4.12) and (4.14). One possibility is to use the Cholesky
factorization [21].
40
Implementation in Simulink
5.1
Simulink and RF Blockset
Simulink is a software for modelling, simulating, and analyzing dynamic

systems. With Simulink, we can easily build models from scratch, or modify
existing models to meet our needs. Simulink supports linear and nonlinear
systems, modelled in continuous time, sampled time, or a hybrid of the two.
The MathWorks provides several additional products that extend the
capabilities of Simulink among them are toolboxes for different fields. In our
model we used the RF Blockset, which provides a library of blocks for
modelling RF systems that include RF filters, transmission lines, amplifiers,
and mixers[1]. RF Blockset is used to represent the components of our RF
system in a Simulink model. The blockset provides several types of
component representations using network parameters (S, Y, Z, ABCD, H, and
T format), mathematical descriptions, and physical properties. With RF
Blockset we can visualize the network parameters of the blocks using plots
and Smith charts. RF Blockset can also be used in conjunction with Real-Time
Workshop to automatically generate embeddable C code for real-time
execution.
5.2 Introduction to our system model

Recall that from chapter 2 that the overall MIMO system looks like
Figure 5.1 Complete RF transmission chain with matching networks inserted

between the source and transmit antenna from one side and the receive antenna and
the load from the other side.
The Source is composed of a series of components (mixer, amplifier, etc...)
and is implemented in another master thesis work [1].
In the present master thesis, the implementation of the matching network and
the antenna system are presented.
A mathematical model of the channel is also available. The receiving side (Rx
Antenna, Match and Drain) is implemented similarly to the transmitting side.
41
Finally, the above three implementations can be put together to represent the
overall system chain.
5.3
System Description
5.3.1
Introduction
Since we are representing the system components (antenna and matching

network) by their scattering parameters, a physical model will be put into
considered. The physical model is created in Simulink by using the RF
Blockset. The blocks available in RF Blockset (S-parameters passive network)
can only represent a 2-port network. Since the antenna network represents an
array with two elements, we need a block that represents 4-port network.
The main RF blocks that are used in this implementation (input port, output
port and S-parameters passive network) are described in Appendix A.
5.3.2
Two port network
A two port network is simply represented by the S-parameter passive

network block in the RF Blockset. This block takes the scattering parameters
of the two port network as an input and it calculates the overall transfer
function from the input port (port 1) to the output port (port 2). Refer to
Appendix A for more details about how this block functions. For more than
two port network, no similar blocks are available and the representation of
the network has to be done in a different way. This is done by describing each
network graphically by its scattering parameters, and then use the Sparameters passive network to represent each of the scattering parameters.
The procedure is illustrated in the next section.
5.3.3
Graphical representation of a two port network
For a two port network, the output travelling waves can be given in terms of
the scattering parameters of the network and the input travelling waves as in
(2.16), which can be represented graphically as shown in figure 5.2
42

Figure 5.2 Graphical representation of the scattering matrices of a two port network.
where
S 21 represents the response from point a1 to point b2 (transmission

from port 1 to port 2).
S11 represents the response from point a1 to point b1 (the reflection
from port 1).
S12 represents the response from point a2 to point b1 (transmission
from port 2 to port1).
S 22 represents the response from point b2 to point a2 (reflection from
port 2).
Recall that from Appendix A that the transfer function of the S-parameters
passive network is given, at a certain frequency f, by
H( f ) =
S 21 * (1 + l ) * (1 s )
2* (1 S 22 * l )(1 in * s )
(5.1)
where
l =
Zl Z0
Zl + Z 0
s =
Z s Z0
Z s + Z0
l
in = S 11 + S12 * S 21 *
(1 S22 * l )
43
If we set S11 , S12 and S 22 equal to zero, knowing that Z L = Z S = Z 0 , the transfer
function of the S-parameters passive network will be given by
H(f )=
S 21
2
(5.2)
Going back to figure 5.2, we can replace each of the boxes by an Sparameters passive network with S11 = S12 = S 22 = 0 and S 21 = S ij . For
example S11 can be replaced by an S-parameter passive network with the
following description
Figure 5.3 S-parameters array. All scattering parameters are zeros except for S 21 .
An input port/output port is needed before/after the S-parameters passive
network to serve as an interface between the mathematical and the physical
models (see Appendix A). This interaction introduces a 6dB loss (see
Appendix A), thus a gain of 2 is added before the input port to compensate
for the loss.
The representation of S11 in terms of RF blocks is shown in the figure below
Figure 5.4 Representation of S11 by RF blocks.
44
The overall representation of a 2 port network will then look like
Figure 5.5 Two port network model using RF blocks.

While this representation could be replaced by an S-parameters passive
network from the RF blockset, a 4-port network representation could not be
replaced by just one block and the above method is used.
5.3.4
Graphical representation of a four port network
For a four port network, the output travelling waves can be given in terms of
the scattering parameters of the network and the input travelling waves as
b1 = S11a1 + S12 a2 + S13 a3 + S14 a4

b2 = S 21a1 + S 22 a2 + S 23 a3 + S 24 a4
b3 = S31a1 + S32 a2 + S33 a3 + S34 a4
(5.3)
b4 = S 41a1 + S 42 a2 + S 43a3 + S 44 a4
45
The above equations can be represented graphically, in a similar way as in the

two port network; then each of the scattering parameters is replaced by its
corresponding representation in terms of RF Blocks as shown in figure 5.4.
The resulting 4-port network is shown in the figure below
S Parm
1
a1
S31
4
b3
S Parm
S11 S Parm
S32
S13
S Parm S 33
S Parm S34
S Parm
b1
a3
S14
S Parm
S 43 S Parm
S Parm S12
S Parm
S41
3
b4
2
a2
S Parm
S22 S Parm
S21 S Parm
b2
S42
S Parm S 44
a4
S23
2
S Parm
S24
S Parm
Figure 5.6 Four port network model using RF blocks.
The above figure is a bit complicated, and if we go for 6 or 8 port networks, it

will be a complete mess. However, there is an option in Simulink where we
can create a subsystem that hides the complication of the original system.
Then the subsystem could be masked so that the user can define all the
variables of the subsystem from outside without going through the individual
blocks in the subsystem [22]. After creating a subsystem for the above 4-port
network, the user will see the box in figure 5.7.
46

Figure 5.7 Sub-system representing the four port network model.
If you double click on the figure above, a dialog Box will open where you can
set all the variables in the subsystem. The box is shown in the figure below
Figure 5.8 Function block parameters of a 4 port network.
Each of the terms in the above dialog box is described below

FIRlength
Desired length of the basebandequivalent impulse response for the
physicalmodels.
Centerfrequency(Hz)
Centerofthemodelingfrequencies.
47
Sampletime(s)
Timeintervalbetweenconsecutivesamplesoftheinputsignal.
Sourceimpedance(Ohms)
Sourceimpedance ofthe RFnetwork described in the physical model

towhichitisconnected.
LoadImpedance(Ohms)
LoadimpedanceoftheRFnetworkdescribedinthephysicalmodelto
whichitisconnected.
Referenceimpedance
Referenceimpedanceofthenetwork.
The above 4-port network can easily be modified to fit our requirements. In
the upcoming sections we will see how we use this 4-port network to
represent the antenna and the matching network.
5.4
Representing Tx Antenna by a 4-port network
The representation of a Tx antenna in terms of its scattering parameters is

given in chapter 2 section 2.3.2 as
Tx
A
S 11TA
= TA
S 21
S 12TA
TA
S 22
(5.4)
For the case of 2 antenna array, the antenna system is a 4-port network. Each
of the above scattering matrices is a square matrix. As we mentioned earlier,
we are interested in the self and mutual coupling effects (represented by S11TA ).
The effects of S12TA and S 21TA are implicitly contained in the channel model.
Thus, the 4-port network antenna system is represented as
48
S11
S 21
1
0
S12
S 2 2
0
0
0
0
0
1
0
0
0
0
(5.5)
Notice that we have a lot of zeros scattering parameters. Using the 4-port
network created in the previous section is a possibility, but an error message
(dividing by zero) will appear, and this will result in the block not functioning
properly. To get rid of the error message, a special 4-port network is built for
the 2 antenna array which removes all the RF blocks that have all zero
scattering parameters.
The resulting 4-port antenna is depicted in the figure below
1
a1
4
b3
S Parm
S11_antenna
b1
1
S Parm
S 12 _antenna
2
a2
3
b4
S Parm
S22_antenna
S Parm
S21_antenna
b2
2
Figure 5.9 4-port antenna network.
49
5.5
Representing Matching Network with RF Blockset
We have studied different types of matching networks in chapter 4. The

network of interest is the one called Hermitian match. We have seen how
the scattering parameters of the Hermitian matching network can be derived
from the scattering parameters of the adjacent networks (amplifier and
antenna). Knowing the scattering parameters of the matching network, we
can represent it by a 4-port network as shown in the figure below
S Parm
1
a1
S31 _match
4
b3
S Parm
S 11_match S Parm
S32 _match
S13 _match
S Parm S33 _match
S Parm
b1
a3
3
S14 _match
S Parm
S Parm
S41 _match
3
b4
2
a2
S Parm
S22 _match S Parm
S42 _match
b2
S Parm S44 _match
S23 _match
a4
4
2
S Parm
S24 _match
S Parm
Figure 5.10 4-Port Hermitian Matching Network.

It is noteworthy to mention that there is an ambiguity in finding the scattering
parameters of the Hermitian matching network, namely S 21 and S12 where
infinite number of solutions are possible. However, if we consider no
interaction between the ports at the two sides of the matching network, which
means that we conjugate match each port individually, then S 21 and S12 will
equal to zero. Further, S11 and S 22 are determined in the same way as in
chapter 4 section 4.3.3. The corresponding matching network is depicted in
figure 5.11.
50
Unit Delay 3
1
3
z
a1
2
b3
S Parm
S11_match
S Parm S33 _match
S Parm S1_amp
Unit Delay 1
a3
1
z
Unit Delay 2
1
4
a2
1
b4
S Parm
S22 _match
S Parm S44 _match
Unit Delay
S Parm S2
a4
2
Figure 5.11 4-Port Matching Network (Individual Port Match).
If we take into account that the amplifier is already matched, S11 of the
matching network is not needed anymore, and we remain with S 22 of the
matching network. This is equivalent to applying the input impedance match
which only takes into account the self and mutual coupling effects. The
corresponding matching network is depicted in figure 5.12.
51
3
a1
2
b3
S Parm S33 _match

Unit Delay 1
a3
1
z
4
a2
1
b4
S Parm S44 _match
Unit Delay
a4
2
z
Figure 5.12 Input Impedance Match.
5.6
Combining Antennas and Matching Network
We have described the antenna and the matching network models in the
above sections. These models are masked and connected to each other as
shown in the figure below
a3
b4
a4
a1
b2
a1
a2
b1
b3
4-Port Matching Network
b4
a2
b3
4-Port Tx Antenna
Figure 5.13 Connections between the antenna and the matching network models.
52
The ports, a1 and a2 of the 4-Port Matching Network are the input ports
through which the signals from the source is fed to the two transmitting
antennas. Further, b4 and b3 of the 4-Port Tx Antenna are the output signals
from the two transmitting antennas to be passed through the channel
whereas, b1 and b2 of the (4-port Matching Network) are the reflected signals
from the matching network to the amplifier.
If we look under the mask we can see in more detail how all the ports of the
two networks are connected. Figure 5.14 shows the connections for the
Hermitian match where all the scattering parameters of the network are
included. Figure 5.15 shows the connections when we consider no interaction
between the ports at the two sides of the matching network. And finally,
figure 5.16 shows the connections when the amplifier is already matched.
53
1
a1
S Parm S11_match
S13 _match
S Parm
S31_match
S Parm
S32_match
S Parm S33_match
Unit Delay
S Parm
Unit Delay 1
S Parm
S14_match
S Parm
S42_match
S Parm
S41_match
1
z
S Parm S22_match
S23_match
S Parm
S Parm S44_match
Unit Delay 3
S Parm S22_amp
2
a2
S Parm S22_amp 1
Unit Delay 2
1
z
S11_antenna S Parm
S Parm S12_antenna
S22_antenna S Parm
S21_antenna S Parm
2
b6
1
b5
54
S24_match
S Parm
Figure 5.14 Flow chart for the case of a 4-port Hermitian match
1
a1
2
a2
S Parm
S Parm
S11_match
S22 _match
Unit Delay 1
S Parm S33 _match
S Parm S44 _match
Unit Delay
S Parm S11_antenna
S Parm S 12 _antenna
S Parm S22 _antenna
S21 _antenna S Parm
2
b3
1
b4
55
S Parm S22 _amp
Unit Delay 3
1
z
S Parm S22 _amp 1
Unit Delay 2
Figure 5.15 Flow chart for the case of individual port match.
The block S22_amp in the above two figures represent the reflection
coefficient of the amplifier. We mentioned earlier that this system will be
connected to the front end (source) model from the master thesis in [1]. In
their model, they consider no backward transmission (reflection). Thus, there
is only one input line at the interface with the matching network. However, in
our model there are two lines for each port (input to the port, reflection from
the port). In the above model, the input port a1 takes the signal from the first
amplifier while the reflected signal is fed through the block that represents the
reflection coefficient of the amplifier to be added to the input signal. In this
way we considered the reflection from the amplifier. The input port a2 takes
the signal from the second amplifier.
1
a1
2
b3
S Parm
S11_antenna
S Parm S33 _match

Unit Delay 1
1
z
S Parm S12 _antenna
2
a2
1
b4
S Parm S44 _match
S Parm
S22 _antenna
Unit Delay
1
S21 _antenna S Parm
Figure 5.16 Flow chart of the individual port match taking into account that the
amplifier is already matched.
In all the above figures, the unit delay box is used to avoid algebraic loops.
Refer to Appendix B to learn more about algebraic loops.
56
5.7
Data initialization
Recall that from section 5.2.4 after creating the subsystem, a box is available
whereallthevariablesofthesubsystem(centerfrequency,samplingtime)
can be defined. What remains is how to define the scattering parameters of
eachblock.ThisisdonebyaMatlabfilethattakesthescatteringparameters
of the antenna system (Touchstone s2p format) and the reflection coefficient
oftheamplifier(S22_amp)asaninput,andthenitdeterminesthescattering
parameters of the corresponding matching network. The Matlab file also
changes the format of the scattering parameters from Touchstone s2p to the
formatthatissuitabletobeusedbytheRFblocks.AfterrunningtheMatlab
file, all the scattering parameters in the Simulink boxes will be defined and
theusercanrunthesimulationtogettheresults.
5.8
Results
To test our system, the below configuration is used
Random
Source
Digital
Filter
a3
Filter 2
a1
b4
a4
a2
a1
b4
a2
4-Port Matching Network
b3
B-FFT
ouput Signal 1
4-Port Tx Antenna
Digital
Filter
Random
Source 1
b1
b2
b3
Filter 1
B-FFT
output Signal 2
a1
b4
a2
b3
B-FFT
input signal 1
4-Port Tx Antenna 1
B-FFT
input signal 2
Figure 5.17 Simulation Setup.
57
The main model is depicted in the upper part of figure 5.17, where we have
the matching network connected to the antenna. To see the effect of mutual
coupling on the antennas, we created another model where we didnt include
the matching network as can be seen in the lower part of figure 5.17. The two
systems are driven by a random source connected to a Digital filter which
makes the signal band limited. The Spectrum Scope (B-FFT) is used to plot the
spectrum of input and output signals verses frequency for each of the two
antennas.
The antennas system under test is a two different dipole antenna array that
has the following configuration
Length of antenna one: 0.3
Length of antenna two: 0.5
Distance between antennas: 0.4
Radius of the antenna: 0.003
The scattering parameters corresponding to this antenna array are shown in
the figure below
Figure 5.18 S-Parameters plots for 2 Different Length Parallel Dipoles.
Since our focus in this master thesis is the effect of the matching network on
the self and mutual coupling of antenna arrays, we used the configuration
shown in figure 5.12 for the matching network. The scattering parameters of
the matching network are derived from the scattering parameters of the two
antennas.
58
For pedagogical reasons, we feed the two antennas by two signals that
operate at different frequencies. The spectra of the two input signals are
shown in the two figures below
Figure5.19(a)
Figure5.19(b)
Figure 5.19 Input signal spectra (a) for first antenna and (b) for second antenna.
To see the effect of mutual coupling we have used the lower part of Fig. 5.17
where the spectra of the output signals for each of the two antennas are
plotted verses the spectra of the input signals and the results are shown in
figures 5.20 (a) and 5.20 (b)
59
Figure 5.20 (a)
FromFigure5.20(a),weseethattheoutputsignaliscomposedoftwoparts.
Thefirstpartshowstheeffectoftheselfimpedanceontheinputsignalwhere
weseeadropofaround5dB.Thesecondpartrepresentsthemutualcoupling
fromthesecondantenna.Thesameeffectscanbeseenintheoutputsignalof
thesecondantenna.
Figure 5.20 (b)

Figure 5.20 Coupling and Reflections effects (a) for First Antenna and (b) for Second
Antenna.
60
Figures 5.21 (a) and 5.21 (b) below show the output signals after connecting
thematchingnetworktothetransmitantennas
Figure 5.21 (a)
If we compare Fig. 5.20 (a) with Fig. 5.21 (a), we can see that the matching
network has reduced the effect of the self impedance, and also significantly
reduced the effect of the mutual coupling from the second antenna on the first
antenna. The same results are obtained for the second antenna, and are shown
in the figure below
Figure 5.21 (b)

Figure 5.21 Reduction in Coupling and Reflections effects (a) for First Antenna and
(b) for Second Antenna
61
Conclusion and future work
In this master thesis, we have shown how to model physical MIMO antennas
with matching network connected to the source (RF frontend) from one side
and to the antenna array from the other side. We have presented a way to
implement this in simulink using the RF Blockset where the scattering
parameters representation of a network is considered. The resulting model
was used to check the effect of the matching network on the mutual coupling
and antenna efficiency. The results show that the matching network has
significantly reduced the effect of mutual coupling and thereby increased the
antenna efficiency.
The lack of a more than 2-port network block in the RF Blockset imposes
limitations for this kind of applications where the networks are represented in
terms of their scattering parameters. Due to this limitation, the proposed
approach to implement the matching network and antenna model is
impractical for MIMO systems bigger than 2x2. For longer systems it is better
to compute the resulting scattering parameters in impeded Matlab file, and
then implement the all system using just one compound block in Simulink.
As for the future work, the whole system (Source (RF frontend), transmit
antenna, channel, receive antenna and Drain) has to be put together. The
design of a full conjuage match has to be clarified, implemented and tested
in Simulink. It is also proposed to implement an interface in Matlab that can
handle bigger MIMO systems.
62
Appendix A
A-1
RF blocks used
A-1.1
S-Parameters Passive Network
The S-parameter Network models a two port passive network in terms of its
S-parameters, frequencies and impedances of these S-parameters.
The S-parameters are given in the form of a 2x2xM matrix, where M
represents the number of frequencies for which measurements are taken. The
following figure shows the correspondence between the S-parameters array
and the vector of frequencies
Figure A.1 Correspondence between the S-parameters array and the vector of
frequencies.
Each of the terms in the dialog box is described below

S-Parameters
S-parameters for a two-port passive network in a 2-by-2-by-M array. M
is the number of frequencies.
Frequency (Hz)
Frequencies of the S-parameters as an M-element vector. All
frequencies must be positive.
Reference impedance (Ohms)
Reference impedance of the network.
63
Interpolation method
The method used to interpolate the network parameters. Table A.1 lists
the available methods and the description of each one
Method
Description
Linear Interpolation
Linear
(default)
Cubic Spline Interpolation
Spline
Piecewise Cubic Hermite
Cubic
Interpolation
Table A.1
Figure A.2 Dialog Box for S-Parameter Passive Network.

Visualization Tab
It is used to plot Magnitude, Angle, Real or Imaginary parts of the SParameters.
64
A-1.2
Input Port
The input port is used as a connection port from Simulink environment to RF

physical blocks. It also calculates modelling frequencies and basebandequivalent impulse response for the physical subsystem. The input port has
the following parameters
FIRlength
Desired length of the baseband-equivalent impulse response.
Centrefrequency(Hz)
Centre of the modelling frequencies.
Sampletime(s)
Time interval between consecutive samples of the input signal.
Sourceimpedance(Ohms)
Source impedance of the RF network described in the physical model
to which it is connected.
Addnoise
If this parameter is selected, noise data in the RF physical is considered.
Initialseed
You specify initial seed value for random number generator.
Figure A.3 Dialog Box for Input Port.

65
A-1.3
Output Port
The Output Port block produces the baseband-equivalent time-domain

response of an input signal travelling through a series of RF physical
components. It performs the following functions
a. Partitioning of RF physical components into linear and nonlinear
subsystems.
b. Extracts the complex impulse response of the linear subsystem for
baseband-equivalent modelling of the RF linear system.
c. Extracts the nonlinear AMAM/AMPM modelling for the RF
nonlinearity.
A-1.3.1
Linear Subsystem
For a linear subsystem, the output port uses input port parameters and
interpolated S-parameters calculated by physical blocks to find the basebandequivalent impulse response. Specifically, it
a. Determines the modelling frequencies f , which is a function of the
centre frequency fc, the sample time ts, and the finite impulse response
filter length N, which are given as Inputport parameters.
The nth element of f, fn, is given by
f n = f min +
n 1
, n = 1, ...., N
ts N
(A-1)
Where
f m in = f c
1
2ts
b. Calculates the passband transfer function for the frequency range as
H ( f ) =
VL ( f )
VS ( f )
Where VS and VL are the source and load voltages and f represents the
modelling frequencies. More specifically
H ( f ) =
66
S 2 1 * (1 + l ) * (1 s )
2 * (1 S 2 2 * l ) (1 i n * s )
(A-2)
where
l =
Zl Z0
Zl + Z0
s =
Zs Z0
Zs + Z0
l
in = S 11 + S 1 2 * S 2 1 *
(1 S 2 2 * l
and
Z S , Z L are source and load impedances.
Sij are the S-parameters of a two-port network.
RF Blockset finds the passband transfer function from the Input Port
parameters as shown in the following figure
Figure A.4 Representation of Input Port Parameters.

c. Translates the passband transfer function to baseband as H ( f f c ) ,
67
The baseband transfer function is shown in the following figure
Figure A.5 Representation of Input Port Parameters.
d. Obtains the baseband-equivalent impulse response by calculating the

inverse FFT of the baseband transfer function.
Each of the terms in the dialog box is described below
LoadImpedance(Ohms)
LoadimpedanceoftheRFnetworkdescribedinthephysicalmodelto
whichitis connected.
VisualizationTab
This tab shows parameters for creating plots if you display the Output
Port mask after you perform one or more of the following actions
68
Run a model with two or more blocks between the Input Port
block and the Output Port block.
Click the Update Diagram button to initialize a model with two

or more blocks between the Input Port block and the Output
Port block.

Figure A.6 Dialog Box for Output Port.
69
A-2
Converting to and from Simulink Signals
When we simulate an RF model, RF Blockset must convert the mathematical

Simulink signal to and from the RF Blockset physical modelling environment.
To do this, RF Blockset interprets the Simulink signal that enters the Input
Port block, Sin, as the voltage VS across the source impedance Z S .
Figure A.7 RF Blockset interpretation of Input and Output signals. The input signal
is represented as a voltage across the source impedance. The output signal is
represented as a voltage over the load Impedance.
RF Blockset interprets the output Simulink signal as the voltage VL over the
load impedance Z L . The RF Blockset interpretation of the input Simulink
signal as the source voltage, VS , produces different results than the
interpretation that the input Simulink signal is the input voltage, Vin. When
the source and load impedances are the same, the former interpretation
produces 6 dB of loss compared to the latter.
The following figure shows the equivalent circuit model when the input
Simulink signal, Sin , is taken to be the input voltage shown in the previous
diagram
+
Sin
VL
70
The load power is
PL =
S2
V L2
= in
R
R
(A-3)
In decibels, the load power is
S2
1 0 lo g ( PL ) = 1 0 lo g in
R
(A-4)
The following figure shows the equivalent circuit model when the input
Simulink signal, Sin, is taken to be the source voltage shown in the previous
diagram
R
+
Sin
VL
The load power is
S in
2
2
VL
PL =
=
R
R
(A-5)
In decibels, the load power is
S2
S2
1 0 lo g ( PL ) = 1 0 lo g in = 1 0 lo g in 6 .0 2
4R
R
(A-6)
71
Appendix B
B-1
Algebraic Loops
Some blocks have direct feedback, meaning that the output cannot be
computed without knowing the values entering the block at input port. When
an input port with direct feedback is driven by the output of the same block
or by a feedback path from other blocks algebraic loops occur.
u(t)
SUM
y(t)
Figure B.1 y (t ) = u (t ) y (t ) .
Figure B-1 shows a simple example [23] where a block is summing two
inputs. Feeding the output y (t ) as input to the SUM block creates an algebraic
loop. On detection of this problem, Simulink uses iterative loops and ends up
solving it correctly i.e. y (t ) = u (t ) . Simulink was able to solve this simple
2
loop but there are situations where the internal solver in Simulink does not
work. For these situations, Simulink provide another solution.
Solution
Simulink has an Algebraic Constraint block which is a convenient way to
model algebraic equations and specify initial guesses. The Algebraic
Constraint block constrains its input signal to zero and outputs an algebraic
state. This block outputs the value necessary to produce a zero at the input.
The output must affect the input through some feedback path. Initial guess
can be provided for the algebraic state value in the block's dialog box to
improve algebraic loop solver efficiency.
Another way to get rid of these loops is to introduce a small delay [23] in your
model at its highest level. You have to be careful where you put this unit
delay as this can change your output considerably. We have introduced unit
delay in our matching network and antenna to solve algebraic loops.
72
Appendix C
C-1
Merging Two Consecutive Networks
Figure C-1 Two connected multiport networks with different numbers of outwardand inward-propagating waves.
The figure above shows two multiport networks with different numbers of
outward- and inward propagating waves. The two networks above are
described by their scattering matrices. The first network has M and N ports;
the second has N and O ports. For both networks the following equations
yield
b1M 1 = S11M M a1M 1 + S12M N b2N1
(C1)
a2N 1 = S21N M a1M 1 + S22N N b2N1
(C-2)
b2N 1 = S11N N a2N 1 + S12N O a3O1
(C-3)
O N N 1
OO O1
b3O1 = S21
a2 + S22
a3
(C-4)
By solving the equations for b1Mx1 and b3Ox1 as a function of a1Mx1 and a3Ox1 , the
following equation is obtained
73
(
b1M 1 S11M M
O1 = ( O M
b S
3 21
( M O
S 12 a1M 1
( OO O1
S 22 a 3
(C-5)
where
(
1
S11M M = S11M M + S12M N ( I S11N N S22N N ) S11N N S21N M
(C-6)
( M O
M N
N N N N 1
S12 = S12 ( I S11 S22 ) S12N O
(C-7)
( O M
1
O N
S 21 = S 21
( I S 22N N S11N N ) S 21N M
(C-8)
( OO
OO
O N
N N N N 1
S22 = S22 + S21 ( I S22 S11 ) S 22N N S12N O
(C-9)
The resultant network which is equivalent to the two networks shown in

figure C-1 is depicted in the figure below
Figure C-2 Network resulting from merging the two networks in Fig.C-1.
74
Figure C-3 Multiport network terminated with source and drain.

The above figure shows a network with source and drain termination. The
wave vectors a1, a2, b1 and b2 are described as
a 1M 1 = b 0M, s 1 + rsM M b1M 1
(C-10)
b1M 1 = S11M M a1M 1 + S12M N a2N 1
(C-11)
a2N 1 = rDN N b2N 1
(C-12)
b2N 1 = S21N M a1M 1 + S22N N a2N1
(C-13)
After some algebraic transformations, the equations can be solved for a1, a2,
b1 and b2 as a function of b0,s
a1 = I rs S11 rs S12 ( I rD S22 ) rD S21

1
a2 = I rD S22 rD S21 ( I rs S11 ) rs S12
b1 = I S11rs S12 rD ( I S 22 rD ) S 21rs
b0, s
(C-14)
rD S21 ( I rs S11 ) b0,s

1
) (S
b2 = I S22 rD S21rs ( I S11rs ) S12 rD

1
) (S
(C-16)
(C-17)
+ S12 rD ( I S 22 rD ) S 21 b0, s
1
11
(C-15)
+ S21rs ( I S11rs ) S11 b0,s

1
21
75
References
References are given according to the schemes given below for some
applications: papers, publications, books and handbooks/codes.
[1] Eyad Arabi, Sadiq Touri, Behavioural modelling of RF front end
devices using Simulink, Master thesis, Chalmers University, signal and
systems group, 2008.
[2] Antenna theory - Analysis and Design (Constantine A.Ballanis) (2nd
Ed) [John Willey]
[3] A. Derneryd and G. Kristensson, Signal correlation including antenna
coupling, IEE ELECTRONICS LETTERS, 2004
[4] Andreas Nilsson, Peter Bodlund, Compensation of mutual coupling
between two antennas, Master thesis, Chalmers University.
[5] http://en.wikipedia.org/wiki/Transmission_line
[6] D.S. Shiu, G.J. Foschini, M.J. Gan and J.M. Kahn, Fading correlation
and its effect on the capacity of multielement antenna systems, IEEE
Trans. Commun, vol.48, pp.502-513, Mar. 2000.
[7] A.F. Molisch and F. Tufvesson, MIMO capacity and measurements,
in Smart Antennas-State of the Art, T. Kaiser, Ed. New York: Hindawi, 2005,
ch 24, pp.467-490
[8] R.G. Vaughan and N.L. Scott, Closely spaced monopoles for mobile
communications, Radio Sci, vol.28, pp.1259-1266, Nov.-Dec.1993
[9] R. Vaughan and J.B. Andersen, Channels, Propagation and Antennas for
Mobile Communications. London, U.K, 2003, Inst. Elect. Eng.
[10] P.S. Kildal and K. Rosengren, Electromagnetic analysis of effective
and apparent diversity gain of two parallel dipoles, IEEE Antennas
Wireless Propag. Lett, vol.2, no. 1, pp.9-13, 2003.
[11] D. M. Pozar, Microwave Engineering. New York:Wiley, 1998, ch 4.
[12] S. M. S. Ow, Impact of mutual coupling on compact MIMO systems.
Lund, Sweden, Dept. Electroscience, Lund University, 2005, M.Sc.thesis.
76
[13] K. Rosengren, J. Carlsson, and P.S. Kildal, Maximizing the effective

diversity gain of two paralle dipoles by optimizing the soucrce
impedances, Microwave Opt. Technol. Lett., vol. 48, no. 3,pp. 532-535, Mar.
[14] J. W. Wallace and M. A. Jensen, Mutual coupling in MIMO wireless
systems: A rigorous network theory analysis, IEEE Trans. Wireless
Commun., vol. 3, pp. 1317-1325, Jul. 2004.
[15] R. G. Vaughan and J. B. Andersen, Antenna diversity in mobile
communications, IEEE Trans. Veh. Technol., vol. VT-36, pp. 147172,
Nov.1987.
[16] J. L. Allen and B. L. Diamond, Mutual coupling in array antennas,
Lincoln Laboratory, MIT, Tech. Rep. 424 (ESD-TR-66-443), 1966.
[17 ]R. A. Speciale, Advanced design of phased-array beam-forming
networds, IEEE Antennas Propagat. Mag., vol. 38, pp. 2234, Aug. 1996.
[18 ]H. A. Haus and R. B. Adler, Circuit Theory of Linear Noisy Networks.
New York: Wiley, 1959.
[19] http://en.wikipedia.org/wiki/Spectral_theory
[20] J. W. Wallace and M. A. Jensen, Termination-dependent diversity
performance of coupled antennas: Network theory analysis, IEEE Trans.
Antennas Propag, vol. 52, pp. 98-105, Jan. 2004
[21] Buon Kiong Lau, J. B. Anderson, G. Kristensson and A. F. Molisch,
Impact of Matching Network on Bandwidth of Compact Antenna
Arrays, IEEE Trans. Antennas Propag, vol. 54, pp. 3225-3238, Nov. 2006
[22]
Ali Soltani Tehrani, Nima Seifi, Simulation of
a MIMO
Communication System in Simulimk, Master thesis, Chalmers University,
Signal and System group, 2007.
[23] http://www1.tip.nl/~t010660/matlab/news/algebraic_loops.txt
77

Modelling and Simulation of Matching Networks For Multi-Antenna Communication Systems

Uploaded by

Copyright:

Available Formats

Modelling and Simulation of Matching Networks For Multi-Antenna Communication Systems

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Modelling and Simulation of Matching Networks For Multi-Antenna Communication Systems

Uploaded by

Copyright:

Available Formats

Modelling and Simulation of

Matching Networks for MultiAntenna Communication

CHALMERS, Signal and System, Masters Thesis EX022/2008

CHALMERS, Signal and System, Masters Thesis EX022/2008

Basic principles of the antenna

2 DESCRIBING N-PORT NETWORK BY MEANS OF SCATTERING

Transmission line theory

Describing N-port networks as a 2-port network

COMPLETE CHAIN OF MIMO SYSTEM USING S-PARAMETERS

CHALMERS, Signal and System, Masters Thesis EX022/2008

Simulink and RF Blockset

Introduction to our system model

Representing Tx Antenna by a 4-port network

Representing Matching Network with RF Blockset

Combining Antennas and Matching Network

CONCLUSION AND FUTURE WORK

A-1 RF blocks used

Converting to and from Simulink Signals

CHALMERS, Signal and System, Masters Thesis EX022/2008

Merging Two Consecutive Networks

CHALMERS, Signal and System, Masters Thesis EX022/2008

1.2 Basic principles of the antenna

CHALMERS, Signal and System, Masters Thesis EX022/2008

1.3.2 Radiation Intensity

1.3.5 Antenna efficiency

CHALMERS, Signal and System, Masters Thesis EX022/2008

1.3.6 Input impedance

1.4 Antenna Arrays

CHALMERS, Signal and System, Masters Thesis EX022/2008

1.5 Mutual coupling

Figure 1.1 Receiving mode coupling paths between antennas 1 and 2.

V1 and V2 are the voltages at antenna 1 and 2 respectively

CHALMERS, Signal and System, Masters Thesis EX022/2008

Z11 and Z 22 are the self impedances on antenna 1 and 2 respectively

The main effects of mutual coupling are

CHALMERS, Signal and System, Masters Thesis EX022/2008

2. The mutual coupling between antennas also changes the correlation of

3. The radiation efficiency of the antenna is also affected by the mutual

CHALMERS, Signal and System, Masters Thesis EX022/2008

Describing N-port network by means

In this chapter, we begin with a short statement of the transmission line

2.1 Transmission line theory

Figure 2.1 Electrical transmission line modelled as a two port network.

CHALMERS, Signal and System, Masters Thesis EX022/2008

2.2 Network Characterization

Figure 2.2 Two port network.

A dm ittance ( Y P aram eters )

CHALMERS, Signal and System, Masters Thesis EX022/2008

These parameters can be measured by applying open circuit or short circuit to

Figure 2.3 Physical components of a high frequency system.

CHALMERS, Signal and System, Masters Thesis EX022/2008

2.3 S-parameters representation

CHALMERS, Signal and System, Masters Thesis EX022/2008

It is possible to relate these four travelling waves by some parameter set.

CHALMERS, Signal and System, Masters Thesis EX022/2008

By substituting the expressions for total voltage and total current on a

V1 = F11 ( H )V1+ + F12 ( H )V 2+

CHALMERS, Signal and System, Masters Thesis EX022/2008