Application Note: Keeping your eye on the sky: A guide to antenna modelling in GNSS testing
Like a radio or a television, the performance of any GNSS-enabled device is limited by the quality of its antenna. But developers often overlook the need to realistically model antennas during testing—and in doing so, run the risk of drawing false conclusions about their receiver’s performance.
Download the Application Note to discover:
- The problem with relying on ‘ideal’ antenna modelling
- How to accurately model an antenna in a Spirent simulator
Simply enter a few details opposite to receive the Application Note—and happy reading!
About Spirent
Spirent has been the global leader in GNSS testing for 25 years. Spirent delivers navigation and positioning test equipment and services to governmental agencies, major manufacturers, integrators, test facilities and space agencies worldwide.
Application note DAN019-TM
Keeping your eye on the sky
The importance of antenna modelling in GNSS testing
Keeping your eye on the sky: The importance of antenna modelling in GNSS testing
Spirent Application Note DAN019-TM Page 2
Spirent Communications PLC
Paignton, Devon, TQ4 7QR, England
Web: http://www.spirent.com/positioning
Tel: +44 1803 546300
Fax: +44 1803 546301
Copyright
© 2009 Spirent. All Rights Reserved.
All of the company names and/or brand names and/or product names referred to in this document, in
particular, the name “Spirent” and its logo device, are either registered trademarks or trademarks of Spirent
plc and its subsidiaries, pending registration in accordance with relevant national laws. All other registered
trademarks or trademarks are the property of their respective owners.
The information contained in this document is subject to change without notice and does not represent a
commitment on the part of Spirent. The information in this document is believed to be accurate and
reliable; however, Spirent assumes no responsibility or liability for any errors or inaccuracies that may
appear in the document.
Multi-GNSS: Benefits, Challenges and test considerations for Technology Developers
Scope
This Application Note explains why modelling the performance of your GNSS system’s antenna
is important, and how to accomplish it with your Spirent GNSS simulator.
Introduction
Any radio system is only as good as its antenna. You can have the most expensive, highest-
definition television money can buy, but if you pair it with a poor antenna you will be
throwing away the benefit of all that costly performance. This philosophy is no less true for
GNSS systems. Capturing as much as possible of the wanted signals from space and rejecting
everything else is the goal for any aspiring GNSS antenna, and a tough goal it is to attain.
GNSS signals are extremely weak when they reach the Earth’s surface and collecting them and
passing them unperturbed into the receiver requires good design.
The perfect antenna of course does not exist. As with all design there is trade-off of one
attribute against another, which ultimately has an effect on what would otherwise be the
ideal performance of your system.
Now turn the issue on its head and consider performance testing. The goal of any test is to
ensure your device is subjected to stimuli which appropriately synthesised reality. For
example, with a GNSS simulator test you would want to ensure you are modelling the effects
of the atmosphere. Of course, in some situations you can leave out multipath and
obscuration, as they do not exist in every scenario in reality, but in all terrestrial applications
the atmosphere is always there. So too is your antenna, yet it is surprising how many people
ignore the antenna and conduct their tests with complex motion, multipath, obscuration and
all but with the default, ideal antenna model selected. Doing this can lead you to false
conclusions about the performance of your receiver in certain applications. Even the most
basic GNSS simulators allow you the ability to model your antenna, at the very least allowing
you to use a realistic gain pattern.
This Application Note explains how to model an antenna in the simulator and shows how much
of a difference there is in a receiver’s performance when the antenna (the real one of which
is normally missing from the test) is correctly modelled.
Keeping your eye on the sky: The importance of antenna modelling in GNSS testing
Spirent Application Note DAN019-TM Page 4
Antenna basics
It is worth reminding ourselves of some basic antenna understanding.
The ideal antenna does not exist. The theoretical ideal antenna is called an isotrope. An
isotropic antenna would have equal gain in all directions. This is not possible in practice
because you need to physically connect a signal feed at some point on the antenna, and it
would not be possible for the antenna to radiate from this point. The 3-dimensional gain
pattern for an isotropic antenna is a perfect sphere, as shown in Figure 1
FIGURE 1 ISOTROPIC ANTENNA GAIN PATTERN
The main purpose of the isotropic antenna is to provide a reference against which other
practical antennas are compared. For example, you may see a real antenna with gain stated
in dBi, which means dB relative to an isotropic antenna.
Antenna gain G is a function of the radiation efficiency (due to mismatch, conduction and
dielectric losses) and the directivity D as given by the formula:
G = .D
Expressed in dB:
G(dB) = 10 log
10
(G)
Relative gain (such as dBi) is defined as the power gain ratio in a specific direction of the
antenna (normally the direction of maximum gain, or the ‘bore-sight’), to the power gain
ratio of a reference antenna in the same direction.
If we consider a typical antenna for a GNSS receiver as shown in Figure 2, we see that it has a
hemispherical antenna gain pattern in the azimuth plane. This is because for all terrestrial
GNSS applications the satellite signals are in practice always arriving from angles greater than
Multi-GNSS: Benefits, Challenges and test considerations for Technology Developers
zero, and we therefore want out antenna to have directionality.
FIGURE 2 TYPICAL GNSS ANTENNA GAIN PATTERN
GNSS simulator and test set-up
Before we look into antenna modelling it is worth reminding ourselves of what a GNSS
simulator does and the typical test set-up. A GNSS simulator reproduces the environment of a
GNSS receiver on a dynamic platform by modelling the vehicle and satellite motion, signal
characteristics, atmospheric and other effects, such that the receiver will actually navigate
according to the parameters of the test scenario. The test signal is a combined, calibrated
and presented to the RF output port connector, of the simulator, the face of which is deemed
to be the physical/electrical boundary of your system’s antenna. In other words,
theoretically, when you look into the simulator output, you are seeing the sky – just as your
GNSS receiver’s antenna would in real life. The simulator output is of course a coaxial
connector, and you connect via a coaxial cable to your device of system to be tested. As can
be seen, the real antenna is omitted. There is no way to physically test it in this kind of set-
up. A large anechoic chamber with a spatially reproduced satellite constellation would be
required, and for standard testing that is prohibitively expensive.
Figure 3 shows the basic connection between a simulator and device under test.
Keeping your eye on the sky: The importance of antenna modelling in GNSS testing
Spirent Application Note DAN019-TM Page 6
Figure 3
Modelling an antenna in the simulator
With Spirent’s simulator software you can easily model the performance of your GNSS
antenna.
The Antenna Pattern Editor allows you to define the signal power level attenuation (relative
gain) and phase delay for elevations of +90 to -90 for the full 360 degrees of azimuth around
the antenna. The resolution can be as fine as 1 degree x 1 degree. The Antenna Pattern
Editor allows you to manually define the characteristics of the antenna, but is also able to
read in antenna manufacturer’s performance data if it is in CSV format.
Multi-GNSS: Benefits, Challenges and test considerations for Technology Developers
Power Level Pattern
Figure 4 shows the Antenna Level Pattern Editor. You can see the AZ/EL segments and their
individual values of attenuation in dB. To the left side is an overview of the entire pattern, a
3-D view and the data entry box used to set the attenuation value.
FIGURE 4 ANTENNA LEVEL PATTERN EDITOR WINDOW
Antenna gain is actually modelled as normalised attenuation. The segments representing the
antenna’s maximum ‘gain’ is always entered as 0dB, all other values being attenuation
relative to that. Any real gain in the antenna must be applied using the simulator’s power
control range and / or a Low Noise Amplifier inserted before the input to the Receiver under
test.
Constructing a realistic antenna level pattern
Now we have established the basic principles of the antenna Pattern editor, and how the
Simulator Software controls the position and orientation of the antenna with respect to the
vehicle on which it is mounted, we can look at designing realistic antenna patterns based on
manufacturer’s data. Figure 5 shows a snapshot of the manufacturer’s gain data for a typical
Keeping your eye on the sky: The importance of antenna modelling in GNSS testing
Spirent Application Note DAN019-TM Page 8
GNSS antenna in XML format.
<?xml version="1.0" encoding="ISO-8859-1"?>
<antenna_pattern>
<az_res> 1.00000000e+000 </az_res>
<elev_res> 1.00000000e+000 </elev_res>
<high_rotation_rate> yes </high_rotation_rate>
<aperture_angle> 6.2831853071795862e+000 </aperture_angle>
<data>
-179.5 -178.5 -177.5 -176.5 -175.5 -174.5 -173.5 -172.5 -171.5 -170.5 -169.5 -168.5 -167.5
89.5000000000000
87.5000000000000
6
85. 000000000000
84.5 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1
83.5 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1
82.5 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1
81.5 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1
80.5 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1
79.5 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2
78.5 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2
77.5 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2
76.5 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2
75.5 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2
74.5 0.4 0.4 0.4 0.4 0.4 0.4 0.4 0.4 0.4 0.4 0.4 0.4
73.5 0.4 0.4 0.4 0.4 0.4 0.4 0.4 0.4 0.4 0.4 0.4 0.4
72.5 0.4 0.4 0.4 0.4 0.4 0.4 0.4 0.4 0.4 0.4 0.4 0.4
71.5 0.4 0.4 0.4 0.4 0.4 0.4 0.4 0.4 0.4 0.4 0.4 0.4
FIGURE 5 SNAPSHOT OF ANTENNA PATTERN DATA XML FILE
The Antenna Pattern Editor imports this data allowing you to view and/or edit it as shown in
Figure 6. Alternatively, you can manually enter the data into the editor, or edit the data in a
spreadsheet. In the ‘View’ menu you can select ‘3D view’ which allows you to see a 3-
dimensional representation of the antenna pattern in terms of elevation and azimuth. Using
the control sliders you can rotate the pattern to view it from different angles. Figure 7 shows
the pattern viewed from three different elevations. As you can see, this pattern simulates a
ground plane below zero degrees elevation. In the editor the cells representing negative
elevation angles are all set to 46dB attenuation (the maximum value)
Multi-GNSS: Benefits, Challenges and test considerations for Technology Developers
FIGURE 6 REAL ANTENNA DATA IN THE EDITOR WINDOW
FIGURE 7 3D VIEW TOOL
Once you are happy that the antenna pattern is correct, simply save the file with an
appropriate name. The extension ‘ant_pat’ is automatically added.
Keeping your eye on the sky: The importance of antenna modelling in GNSS testing
Spirent Application Note DAN019-TM Page 10
Phase pattern
We can now look at the antenna’s phase pattern. Modelling antenna phase is important
because –as with gain – this will vary depending on the arrival angle of the satellite signal.
Phase discontinuities cause delays and signal cancellation. Both are undesirable yet
unavoidable effects. One particular characteristic of a GNSS antenna is its electrical phase
centre. The electrical phase centre of an antenna is not the same as the physical centre, and
may in fact be a theoretical point in space above the antenna. For measurements in precision
applications such as surveying you need to know where the electrical phase centre of an
antenna is because it is between this point and the electrical phase centre of the satellite’s
transmitting antenna that the pseudorange is calculated.
The effect on carrier phase due to the variation in the electrical phase centre can be
simulated by applying the correct phase pattern – in just the same way as we did for the gain.
Figure 8 Shows the Phase pattern Editor, which is basically the same as the level pattern
editor
FIGURE 8 ANTENNA PHASE PATTERN EDITOR WINDOW
Multi-GNSS: Benefits, Challenges and test considerations for Technology Developers
You can see the AZ/EL segments and their individual values of phase in degrees. To the left
side is an overview of the entire pattern, a 3-D view and the data entry box used to set the
phase value. Positive values retard the carrier phase (receiver moving away from satellite),
negative values advance the carrier phase (receiver moving towards satellite).
It is also possible (as with the level pattern) to edit the pattern in a CSV file or read-in
manufacturer’s data.
Antenna orientation and location
Now we have modelled the characteristics of the antenna itself, we need to consider the
physical placement of the antenna on the host vehicle, as this can have angular effects due to
the vehicle motion. By default, the antenna position and angular offsets are set to zero. The
convention is for zero degrees antenna azimuth and zero degrees antenna elevation to lie
along the vehicle x- axis, x
b
in Figure 9. This figure also shows a typical hemispherical
antenna pattern with its axes. Placing this antenna on the vehicle shows the default
alignment of the antenna pattern axes and the vehicle axes. An aircraft is shown in this
illustration, but the convention applies equally to the other vehicle models.
x
b
= Nose
x
a
= 0 Azimuth,
0 Elevation
y
b
= Right wing
y
a
= +90 Azimuth
-z
b
= Up
z
a
= +90 Elevation
z
b
= Down
z
a
= -90 Elevation
CofG
Typical
antenna
pattern
0º Azimuth,
0º Elevation
+90 º
Azimuth
+90 º
Elevation
FIGURE 9 ANTENNA ORIENTATION AND LOCATION
Ignoring antenna angular offsets, the following conventions are used for receiver antenna
azimuth and elevation:
• 0.0° Azimuth, 0.0° elevation in the grid of the antenna pattern editor is along the
vehicle’s x-axis. This is normally towards the front of the vehicle, which is NOT
necessarily the direction of travel.
Keeping your eye on the sky: The importance of antenna modelling in GNSS testing
Spirent Application Note DAN019-TM Page 12
• Looking from above, positive azimuth is a clockwise rotation in the vehicle's x-y plane.
• For receiver antenna patterns, positive elevation is towards the negative z-axis.
• For a typical receiver antenna pattern, the direction of maximum antenna gain (the
bore-sight) will be at 90° elevation in the grid of the antenna pattern editor.
By default, the antenna is located at the vehicle Centre of Gravity (C of G) see Figure 9. You
can offset the position of the antenna from the C of G, along each body axis; and offset the
angular position by rotating the antenna pattern from its default. The antenna reference axis
set (x
a
, y
a
, z
a
) will be coincident with the vehicle body frame, by default. The angular offsets
are defined as a sequence of Euler angle rotations of the antenna pattern frame in heading
(
a
), elevation (
a
), and bank (
a
), as for body attitude with respect to local level.
An example application where you would want to specify the physical position offset of an
antenna is on the mast of a ship or yacht. An antenna at the top of mast will see considerably
greater movement in X-Y than one placed at the vessel’s C of G, therefore it is important to
model this in order to maintain the realism of the test scenario. Figure 10 shows the true
motion of the antenna mounted at +50m offset in Z from the C of G on a vessel travelling a
straight path in a ‘Moderate’(Beaufort scale) sea condition. As you can see the antenna is
certainly not moving in a straight line. Designers of such receivers can use this effect to
optimise algorithms to filter out these ‘lever-arm’ effects.
Multi-GNSS: Benefits, Challenges and test considerations for Technology Developers
FIGURE 10 LEVER-ARM EFFECT OF OFFSET ANTENNA MOTION
Relative Obscuration
Another effect which can be applied using the Antenna Pattern is signal blockage due to
obscuration which is fixed relative to the antenna. For example, a GNSS device mounted on
the dashboard of a car has its view of the sky obscured by the physical structure of the car.
This remains fixed relative to the device’s antenna, so can be modelled by setting maximum
attenuation to the appropriate cells of the Level Pattern. Figure 11 shows the 3-D view of a
Level Pattern which includes the effects of obscuration due to the vehicle structure.
Keeping your eye on the sky: The importance of antenna modelling in GNSS testing
Spirent Application Note DAN019-TM Page 14
FIGURE 11 LEVEL PATTERN FOR DASH-MOUNTED ANTENNA
Pattern switching
It is even possible to use different antenna patterns to simulate changes between states of
relative obscuration. For example a spacecraft moving from a static position against the
structure of the launch tower to a stage where it is passing different elements of the launch
tower structure to a final stage where it is in open sky.
Multi-GNSS: Benefits, Challenges and test considerations for Technology Developers
Receiver Test
To highlight how different the results of a receiver test can be when using no antenna pattern
compared to using a realistic antenna pattern a test was carried out using the simulator. The
same receiver was tested twice using an identical test scenario (location, time, date, satellite
orbits, atmosphere etc.) but for the first test no antenna pattern was used and for the second
test a realistic antenna pattern was used. Figure 12 shows the two sets of results. The left-
hand plot is a scatter plot of absolute position for the test with no antenna pattern enabled,
and the right-hand plot is the repeat test but with the antenna pattern enabled. On both
plots the inner circle represents 4.5m distance from the absolute position, and the outer
circle 9.5m
FIGURE 12 NO ANTENNA PATTERN VS. REALISTIC ANTENNA PATTERN
As you can see, there is a significant degradation in the receiver performance when a realistic
antenna pattern is used. This is due to the attenuation and phase shift of signals from certain
satellites due to the characteristics of the antenna. This is just a simple test to demonstrate
the problem, but from it we can conclude that neglecting to model the antenna can lead to a
falsely good result from a receiver under test, which undermines the test integrity, and
ultimately could lead to poor performance in the field, by which time your products are on
the market, you have unhappy customers and it is too late to iterate the design. Many GNSS
simulator systems overlook this and other important, yet fundamental test requirements.
Spirent’s systems give you this functionality, and testing confidence where you need it most.
Keeping your eye on the sky: The importance of antenna modelling in GNSS testing
Spirent Application Note DAN019-TM Page 16
If you found this article of interest
www.spirent.com/positioning
You can find more GNSS related technical articles, white papers and
eBooks at the Spirent Positioning website.
Visit the Spirent GNSS blog
www.spirent.com/Blog/Positioning.aspx
Keep up to date with the news and view within the GNSS community.
Get access to information, tips, and ideas that could help solve some
of your issues by registering for the Spirent Positioning blog
Need more information?
gnss-solutions@spirent.com
If you're curious about how the information in this document can
benefit you and your business, please contact Spirent to discuss your
particular situation and explore opportunities.
Multi-GNSS: Benefits, Challenges and test considerations for Technology Developers
FOR MORE INFORMATION
Please visit our website: http://www.spirent.com/positioning and do not hesitate to contact
your nearest Spirent representative for more detailed information. To find the appropriate
contact details please visit the ‘Contact Us’ page on the website and select your location and
application.
If you found this article of interest
www.spirent.com/testmore
Visit the Spirent GNSS blog
www.spirent.com/Blog/Positioning.aspx
Need more information?
mailto:gnss-solutions@spirent.com
Spirent Communications
Aspen Way,
Paignton, Devon,
TQ4 7QR England
Tel: +44 1803 546325
globalsales@spirent.com
www.spirent.com
Spirent Federal Systems Inc.
22345 La Palma Avenue
Suite 105, Yorba Linda
CA 92887 USA
Tel: 1 714 692 6565
info@spirentfederal.com
www.spirentfederal.com
