Project Report

ATC-238

GPS Antenna Multipath Rejection Performance

A.-M. Dinius

7 August 1995

Lincoln Laboratory
MASSACHUSETTS INSTITUTE OF TECHNOLOGY
LEXINGTON, MASSACHUSETTS

Prepared for the Federal Aviation Administration,

Washington, D.C. 20591

This document is available to the public through

the National Technical Information Service,
Springfield, VA 22161
This document is disseminated under the sponsorship of the Department
of Transportation in the interest of information exchange. The United
States Government assumes no liability for its contents or use thereof.
 TECHNICAL REPORT STANDARD TITLE PAGE
1. Report No. 2. Government Accession No. 3. Recipient's Catalog No.
ATC-238

4. Title and Subtitle 5. Report Date

7 AUllUst 1995
GPS Antenna Multipath Rejection Performance 6. Performing Organization Code

7. Author(s) 8. Performing Organization Report No.

Anne-Marie Dinius ATC-238

9. Performing Organization Name and Address 10. Work Unit No. (TRAIS)
Lincoln Laboratory, MIT
244 Wood Street 11. Contract or Grant No.
Lexington, MA 02173-9108 DTFAOI-93-Z-02012
12. Sponsoring Agency Name and Address 13. Type of Report and Period Covered
Department of Transportation
Project Report
Federal Aviation Administration
Systems Research and Development Service 14. Sponsoring Agency Code
Washington, DC 20591
15. Supplementary Notes

This report is based on studies performed at Lincoln Laboratory, a center for research operated by Massachusetts Institute of
Technology under Air Force Contract FI9628-95-C-0002.

16. Abstract

A GPS antenna multipath rejection perfonnance evaluation was conducted at Lincoln Labora-
tory. Ground reference station antennas and aviation patches were tested for their ability to reject a
multipath signal. Different types of ground plane structures were used such as choke rings, ground
planes, and mock sections of fuselage. Frequencies transmitted were Ll (1575 MHz), L2 (1227 MHz),
and the median GLONASS frequency (1609 MHz ). Receive amplitude and phase were measured on
each antenna. Subsequently, these data were converted to absolute gain for a right-hand and left-
hand circularly polarized signal as a function of satellite elevation angle.

Two types of multipath signals were considered: ground bounce multipath and building or
structure bounce multipath. Ground bounce multipath typically occurs at low satellite elevation
angles while structure bounce multipath can occur at any satellite elevation angle. Separate analysis
methods were used to assess an antenna's ability to reject either type of multipath.

This report describes the data collection methods, data reduction and analysis, and the results.

17. Key Words 18. Distribution Statement

GPS GPS antenna
This document is available to the public through the
multipath choke ring
National Technical Information Service,
Springfield, VA 22161.

19. Security Classif. (of this report) 20. Security Classif. (of this page) 21. No. of Pages 22. Price

Unclassified Unclassified 70

FORM DOT F 1700.7 (8-72) Reproduction of completed page authorized

 EXECUTIVE SUMMARY

BACKGROUND

This report documents a performance evaluation of GPS antennas typically used in

real-time aircraft navigation applications. The study was conducted in order to determine the
ability of an antenna to reject multipath signals. Since our GPS application was aircraft
navigation, the antennas we tested were either aircraft antennas or ground station antennas
typical of what would be used in a GPS differential navigation system.

STUDYMEmOD
Data were collected by mounting the test antenna on a pedestal in an anechoic chamber.
The test antenna was then illuminated with a linearly polarized source whose frequency was
1575 MHz (Ll). The amplitude and phase of the test antenna while it was being illuminated
were measured. Variations in source elevation angle with respect to the test antenna and test
antenna azimuth angle with respect to the source were introduced in order to simulate the
variability of GPS satellite elevation and azimuth angles. Receive amplitude and phase of the
test antenna were then converted to absolute gain, for both a right-hand and a left-hand circularly
polarized signaL Absolute gain for both polarizations was plotted as a function of elevation
angle.
Three of the antennas tested were capable of receiving at frequencies other than Ll
(1575 MHz). Accordingly, they were tested at L2 (1227 MHz) and the median GLONASS
frequency (1609 MHz).
One of the purposes of this study was to determine if mounting antennas on different
ground plane structures improved their ability to reject multipath signals. Again, a typical GPS
differential navigation scenario was created by mounting ground-based antennas on choke rings
or ground planes and aircraft antennas on a mock section of a 727 fuselage. Some antennas were
tested on more than one ground plane structure.
Two types of multipath signals were considered: ground bounce multipath and building
or structure bounce multipath. Ground bounce multipath typically occurs at low satellite
elevation angles. Structure bounce multipath can occur at any satellite elevation angle. We
measured rejection of ground bounce multipath by determining the antenna's cross polarization
ratio at a five-degree satellite elevation angle. The larger the cross polarization ratio, the better
the antenna would be at rejecting ground bounce multipath. Conversely, minimum cross
polarization ratios at all satellite elevation angles were considered when determining an
antenna's ability to reject a structure bounce multipath signal. In our analysis, it was assumed
that GPS signals would, upon reflection from a perfectly conducting surface, switch polarizations
from right-hand to left-hand circular polarization.

RESULTS
The choice of antenna ground plane structure was important for improving an antenna's
ability to reject ground bounce and structure bounce multipath signals. When antennas were
mounted in choke rings, there were better at rejecting both types of multipath than when they

III
were mounted on ground planes. Variations in antenna type (Le., crossed dipoles, patch,
quadrahelix) also affected multipath rejection ability.
The broad-band antennas performed roughly as well at L2 (1227 MHz) or the GLONASS
frequency (1609 MHz) as they had at Ll (1575 MHz). Antennas tested on both a choke ring and
ground plane were better at rejecting ground bounce and structure bounce multipath when they
were mounted on the choke ring.

RECOMMENDAnONS

Since multipath is typically the largest source of error in a differential GPS system, great
care should be taken when choosing the type of antenna as well as the antenna location for the
ground-based station.
Future studies should focus on characterizing the environment surrounding specific
locations to allow the user a means of minimizing the effects of multipath. With a site specific
multipath profile, the user could determine whether most of the multipath signals would be
ground bounce or structure bounce. Accordingly, one could choose an antenna whose multipath
rejection performance matched the type of multipath predominant at that site.

IV
 ACKNOWLEDGMENTS
I would like to thank the following people in the Air Traffic Surveillance Group for their
help on this project: Raymond LaFrey for supervision; Edward Bayliss, for creating the
parameters of this study and guiding me through it from start to finish; Richard Ferranti, for
technical and editorial advice; George Knittel, for help in defining and interpreting this study;
Michael Burrows, for technical clarification; Ralph Halvorsen, for building the hardware;
Timothy Borge, for collecting the data; and Katharine Krozel, for editing, artwork, and creative
suggestions.

v
 TABLE OF CONTENTS

Executive Summary 111

Acknowledgments v
List of Illustrations ix
List of Tables xi
1. INTRODUCTION 1

2. MULTIPATH 3

3. THE DATA COLLECTION 5

3.1 Hardware 5
3.2 Antennas 9
3.3 SetUp 13

4. RESULTS 19
4.1 Absolute Gain 19
4.2 GPS Antenna Multipath Rejection Performance 19
5. GROUND PLANE VERSUS CHOKE RING 27

6. OTHER FREQUENCIES 31

7. CONCLUSIONS 37

APPENDIX 39

REFERENCES 57

Vll
 LIST OF ILLUSTRAnONS
Figure
No. Page
1 Multipath Bounce Scenario. 3
2 Ground Plane Structures Used: Mock Fuselage, Ground Plane, and
Choke Ring. 5
3 Litton Antenna Mounted on the Mock Fuselage. 9
4 Antennas Tested. 11
5 Anechoic Chamber Used to Test Antennas. 13
6 Sketch of Test Setup. 15
7 Azimuth and Elevation Variation. 17
8 Azimuth Multipath Rejection Variation for NovAtel Antenna on a
Choke Ring, 5-degree Elevation Angle 21
9 Smallest Minimum Cross Polarization Ratio Determined From
Absolute Gain, NovAtel (corner fed patch) Antenna on a Choke Ring. 23
10 Largest Minimum Cross Polarization Ratio Determined From Absolute
Gain, NovAtel (corner fed patch) Antenna on a Choke Ring. 23
11 Ground Plane Structure Comparison for NovAtel (corner fed patch)
Antenna, Ground Plane vs. Choke Ring. 27
12 Ground Plane Structure Comparison for Trimble (patch) Antenna,
Ground Plane vs. Choke Ring. 27
13 Absolute Gain for 3S (quadrahelix) Antenna on a Ground Plane at
1609 MHz; Minimum Cross Pol. Ratio =8.4 dB at 89-degree
Elevation. 31
14 Absolute Gain for Dome & Margolin (crossed dipoles) Antenna on a
Ground Plane at 1227 MHz; Minimum Cross Pol. Ratio = 1.9 dB at
6-degree Elevation. 33
15 Absolute Gain for Turbo Rogue (crossed dipoles) Antenna on a Choke
Ring at 1227 MHz; Minimum Cross Pol. Ratio = 12.6 dB at 5-degree
Elevation. 35
A-I Absolute Gain for 3S (quadrahelix) Antenna on a Ground Plane;
Minimum Cross Pol. Ratio = 10.4 dB at 14-degree Elevation. 39
A-2 Absolute Gain for Dome & Margolin (crossed dipoles) Antenna on a
Ground Plane; Minimum Cross Pol. Ratio =3.2 dB at 5-degree
Elevation. 41

ix
 LIST OF ILLUSTRATIONS (continued)

Figure
No. Page
A-3 Absolute Gain for Litton (patch) Antenna; Minimum Cross Pol.
=
Ratio -0.4 dB at 5-degree Elevation. 43
A-4 Absolute Gain for NovAtel (comer fed patch) Antenna on Choke
=
Ring; Minimum Cross Pol. Ratio 11.1 dB at 6-degree Elevation 45
A-5 Absolute Gain for NovAtel (comer fed patch) Antenna on Ground
=
Plane; Minimum Cross Pol. Ratio 1.4 dB at 14-degree Elevation. 47
A-6 Absolute Gain for Trimble (patch) Antenna on Choke Ring; Minimum
=
Cross Pol. Ratio 4.9 dB at 5-degree Elevation. 49
A-7 Absolute Gain for Trimble (patch) Antenna on Ground Plane;
=
Minimum Cross Pol. Ratio -0.2 dB at 5-degree Elevation. 51
A-8 Absolute Gain for Trimble (patch) Antenna on Fuselage; Minimum
=
Cross Pol. Ratio 0.2 dB at 5-degree Elevation 53
A-9 Absolute Gain for Turbo Rogue (crossed dipoles) Antenna; Minimum
=
Cross Pol. Ratio 6.9 dB at 5-degree Elevation. 55

x
 LIST OF TABLES
Table
No. Page
1 Antennas and Ground Plane Structures 7
2 Antenna Specifications 11
3 Peak Azimuth Variation of Ground Bounce Multipath Signal
Rejection 20
4 Smallest Minimum Cross Polarization Ratio 25
5 Largest Minimum Cross Polarization Ratio 25

Xl
 1. INTRODUCTION
The Global Positioning System (GPS) is a satellite-based navigation system that allows
users to determine their location with meter-level accuracy. In the recent past, the aviation
community has become increasingly interested in using GPS for aircraft navigation and, in
combination with a data link, for surveillance. Though the inherent accuracy of the GPS signal
is sufficient for many phases of flight, terminal approach and surface surveillance require that the
satellite signals be differentially corrected by ground stations providing precision references.
GPS signals are transmitted in L-band at 1227 MHz and 1575 MHz. Since satellites can
appear anywhere from the horizon to the zenith, satellite signal reception by both aircraft and
differential ground stations can be corrupted by multipath interference. Multipath occurs when
the GPS signal is reflected from the ground or a structure large compared to its wavelength (such
as a building, another aircraft, or parts of the aircraft itselt). Depending on the signal path
difference between the direct and bounced signal, the multipath constructively or destructively
adds with the direct signal. The result of this addition is either signal fading or a signal so
corrupted that it cannot be used by the GPS receiver.
Three factors influence how much multipath interference will affect a particular GPS
receiver: the antenna's ability to discriminate between a direct and a reflected signal, its location,
(the surrounding ground environment or an aircraft's position), and what, if any, signal
processing techniques the receiver uses to mitigate multipath interference.
The Air Traffic Surveillance Group conducted a study to assess the first of these factors.
As is shown in this report, multipath signals may be mitigated on the basis of polarization: GPS
signals are right-hand circularly polarized while multipath is often left-hand circularly polarized.
Thus, the study's focus was to measure key polarization parameters of six commercial GPS
antennas and to assess the effects of mounting these antennas on different ground plane
structures (choke ring, fuselage section, or flat ground plane). Data were analyzed to determine
the multipath susceptibility of these antennas at elevation angles likely to produce either ground
bounce or structure bounce multipath.
This report describes the data collection methods, the data reduction and interpretation
techniques, and the results of these antenna measurements.

1
 2. MULTIPATH
GPS receivers use multilateration techniques to determine their location based on the
time of arrival of the GPS satellite signal. Four satellites are needed to determine a three
dimensional position and correct the receiver's clock bias. Since the exact location of each
satellite is known, solving for the precise time it takes a signal to propagate to the GPS antenna
uniquely determines its location.
Multipath occurs when a direct signal is reflected off of a structure that is large compared
to the GPS signal's wavelength, resulting in an inaccurate measure of the propagation time.
Figure I depicts three satellite-to-ground paths: a direct line-of-site path, structure bounce path
to the GPS antenna, and a ground bounce path to the GPS antenna. The angle 9 in Figure I
represents the satellite elevation angle.
Snell's law of reflection states that when an electromagnetic wave undergoes a reflection,
the angle of incidence is equal to the angle of reflection. Given Snell's law, and the geometry
pictured in Figure I, it is clear that, with respect to the GPS antenna, a satellite can appear to be
transmitting from a negative elevation angle (-9). With this in mind, this data collection was
designed to simulate satellite elevation angles varying from -30 degrees to 90 degrees, where 90
degrees is a satellite at the zenith.

GROUND BOUNCE
MULTIPATH SIGNAL

6• SATELUTE ELEVATION ANGLE

K0421••2

Figure 1. Multipath Bounce Scenario.

When a circularly polarized signal is reflected off of a perfectly conducting reflector, it

reverses its polarization sense. Polarization sense continues to be reversed for each subsequent
bounce. Therefore, a right-hand circularly polarized signal that has bounced an odd number of
times will become a left-hand circularly polarized signal. Similarly, a right-hand circularly

3
polarized signal that has bounced an even number of times will remain a right-hand circularly
polarized signal [}].
When a signal bounce off of a surface, amplitude is typically attenuated and phase is
shifted. The amount of attenuation and phase shift are determined by the reflection coefficient
(r) of the reflecting surface. For a perfect reflector, the amplitude remains constant after
reflection. That is:

In=}

However, for all other surfaces, amplitude is attenuated. In this data collection, it was assumed
that all multipath signals were reflected off of a perfect reflector. Additionally, differences in
phase between the direct signal and a rnultipath signal were not computed except for possible
circular polarization reversals. The results therefore represent a worst-case estimate of the
antenna multipath rejection capabilities.
Circular polarization is the combination of horizontal and vertical polarizations of equal
amplitudes and orthogonal phases. Because these two linear polarizations have different
reflection coefficients, horizontal and vertical components of a circularly polarized signal
undergo different reflections. The result is that circularly polarized signals, upon reflection, off
any surface except a perfect conductor, become elliptically polarized signals. The extent of the
ellipticity is primarily determined by the reflection coefficient of the vertical component [2]. For
simplicity, it was assumed in this study that all reflected multipath signals were circularly
polarized. Again, these results represent worst-case examples of the antenna's multipath
rejection capabilities.

4
 3. THE DATA COLLECTION

Data were collected with the antennas mounted in one of three different ground plane
structures. In all, six antennas were tested (four ground-based and two aviation). Antennas that
are capable of performing outside of 1575 MHz (Ll) were tested at either 1227 MHz (L2) or a
GLONASSI frequency depending on the specifications of the antenna. Data were collected in an
anechoic chamber simulating a GPS antenna receiving a satellite signal.

3.1 HARDWARE

One of the goals of this study was to determine the role of an antenna's immediate
environment in suppressing multipath signals. Three types of ground plane structure were
considered: a ground plane, a choke ring, and a mock section of fuselage. Figure 2 shows these
three types of ground plane structures. Table 1 lists the diameters of the ground plane structures
used in this study. An XX indicates that data were collected with the antenna mounted on that
structure. The two antennas mounted on the mock fuselage were the aviation antennas.

Figure 2. Ground Plane Structures Used: Mock Fuselage, Ground Plane, and Choke Ring.

GLONASS is the Russian equivalent of GPS, operating at approximately 1609 MHz.

5
 Table 1. Antennas and Ground Plane Structures

Antenna Ground Plane Choke Ring Mock Fuselage

38 (Quadrahelix) XX 15"
Dome and Maroolin (dipoles) XX 15"
Litton (patch) XX
NovAtel (comer fed patch) XX 18" XX 18"
Trimble Aviation (patch) XX 24" XX 24" XX
Turbo Rogue (dipoles) XX 15"

3.1.1 Ground Plane

Until recently, ground planes were the only hardware mounting option commercially
available for ground-based GPS antennas. Manufacturers used different diameters and metals for
their ground planes. For the sake of consistency, all ground-based antennas tested were
measured on an aluminum ground plane. Two of these antennas were then measured in a choke
ring. As can be seen from Table 1, the diameter of the ground plane was equal to that of the
choke ring when measurements were made on both.

3.1.2 Choke Ring

Choke rings are designed to reject ground bounce multipath signals. A choke ring, like a
corrugated horn, changes the boundary conditions of the surface the antenna is mounted on. On
a flat surface, the boundary conditions for the E field are different from those for the H field.
This makes the antenna sensitive to the predominantly cross polarized multipath signal. Adding
quarter-wavelength grooves to that flat surface creates the same boundary conditions for both
fields, thereby reducing the antenna's sensitivity to multipath signals [3]. Because of
aerodynamic considerations, it is nearly impossible to affix a choke ring onto the fuselage of an
aircraft. However, in a differential GPS navigation system, the reference ground station's
antenna can be mounted on either a ground plane or a choke ring.

3.1.3 Fuselage Section

Little can be done to minimize the effects of multipath signals being received by an
antenna mounted on an aircraft. The chief source of multipath error for an airborne aircraft
comes from edge diffraction and signal propagation along the aircraft fuselage [4]. During
aircraft approach and landing, however, the dominant multipath is from ground and building
reflections. In order to better understand and quantify both types of multipath sources,
measurements were taken on two aviation antennas mounted on a mock 727 fuselage. The mock
fuselage was built to ARINC standard 743A2 at Lincoln Laboratory. Figure 3 shows the Litton
aircraft antenna mounted on the mock fuselage.

2 4' X 7' with a radius of curvature of 96 inches; ARINC Characteristic 743A, p. 34.

7
 Figure 3. Litton Antenna Mounted on the Mock Fuselage.

3.2 ANTENNAS

In a differential aircraft navigation application, two different types of antennas would be

required: a reference ground station antenna and an aircraft antenna. The difference between the
two is that the ground station is not a dynamic platform nor is its antenna limited by the size
constraints of an aircraft antenna. Most commercially available aircraft antennas are patches.
Accordingly, two patch aircraft antennas were tested. Ground station antennas used in this study
included a quadrahelix, two crossed dipoles, and a corner fed patch. The antennas tested were a
representative sample of the types of ground station antennas commercially available. Table 2
lists the pertinent data on each antenna tested in this experiment while Figure 4 shows the
antennas.

9
 Table 2. Antenna Specifications

Model Number Type of Amplifier Gain

Antenna or PIN Antenna indB Noise Figure
3S 3Snav027014 Quadrahelix 23 1.5
Dome and C146101 Crossed dipoles passive N/A
Margolin antenna
Litton 510116-1 Patch passive N/A
antenna
NovAtel GPS-501 Corner Fed 24 3.0
Patch
Trimble Patch 1624811 Patch 44 2.0<
Turbo Rogue 7490400-4 Crossed dipoles 54 3.0-3.5

Figure 4. Antennas Tested.

11
3.3 SETUP

All data for this study were collected in an anechoic chamber at the Laboratory's Antenna
Test Range (ATR). Figures 5 shows the anechoic chamber used in this study.

Figure 5. Anechoic Chamber Used to Test Antennas.

Figure 6 is a diagram of the test setup. A linearly polarized source, at frequencies of

1227.60 MHz (L2), 1575.42 MHz (L1), and 1609.00 MHz (GLONASS), was used to illuminate
the test antennas. Since a circularly polarized source antenna was not available, we transmitted
orthogonal linear polarization and mathematically converted these to right-hand and left-hand
circular polarizations. Test antenna amplitude and phase were measured and calibrated by
replacing the test antenna with a standard gain horn.

13
 SOURCE
ANTENN

MEASURE MEASURE
SOURCE AMPLrrUDE TEST AN"rENNA AMPLrrUDE
POLARIZATION PURITY TEST ANTENNA PHASE

Figure 6. Sketch ofTest Setup.

3.3.1 Elevation and Azimuth

GPS satellite elevation angles vary from 0 degrees to 90 degrees (where 90 degrees represents a
satellite at the zenith); while azimuth angles vary from 0 degrees to 360 degrees. In this test, we
emulated these variations in azimuth and elevation angles by varying the antenna's orientation
with respect to the source. In this way, we captured the real-life scenario of the GPS signal
moving with respect to the antenna. To take into account aircraft banking angles and multipath
received from the ground at negative elevation angles, the test antenna elevation angles were
varied from -30 degrees to 90 degrees. Data were collected at I-degree increments across this
range of elevation angles while holding the azimuth angle fixed. Test antenna azimuth angle was
varied from 0 degrees to 360 degrees in 45-degree increments because we assumed this would
give us an accurate representation of antenna variation due to azimuth angle while not requiring
us to collect prohibitively large amounts of data. Figure 7 illustrates the range of azimuth and
elevation angles.

15
AZIMUTH AND ELEVATION ROTATION FOR
GPS ANTENNA TEST
y "<.-
, , .....
~ "... -

~(~~~
-- ,

AZIMUTH

ELEVATION

240970.1 X

Figure? Azimuth and Elevation Variation,

17
 4. RESULTS
To evaluate the antenna's performance in the presence of cross-polarized multipath
interference, right-hand circular polarization and left-hand circular polarization absolute gains
versus elevation angle were plotted for all azimuths. These results are discussed in the following
sections.

4.1 ABSOLUTE GAIN

The two orthogonal linear polarizations were converted to right-hand and left-hand
circular polarization using Equation 4.1

(4.1)

where G = Gain
+ =Right-Hand Circular Polarization (RHCP)
- =Left-Hand Circular Polarization (LHCP)
h = Horizontal Polarization
v = Vertical Polarization
b = (Horizontal phase -Vertical phase)
Right-hand circular polarization and left-hand circular polarization signal levels were
converted to absolute gain using the calibration data from the standard gain hom. The Appendix
contains one absolute gain plot for each antenna tested. Though GPS satellites transmit right-
hand circular polarization signals, left-hand circular polarization absolute gains were calculated
as a means of assessing the antenna's multipath rejection capability.
4.2 GPS ANTENNA MULTIPATH REJECTION PERFORMANCE
Two types of multipath can corrupt a GPS measurement: ground bounce multipath and
multipath signals that are reflected off buildings or similar structures (structure bounce
multipath). Typically, ground bounce multipath occurs at relatively low satellite elevation
angles. In this study we assumed that ground bounce multipath would occur at satellite elevation
angles less than 30 degrees. Alternatively, building bounce multipath can occur at any satellite
elevation angle, so we considered all positive satellite elevation angles. Given the two different
characteristics of these types of multipath, separate methods of quantifying an antenna's
performance in either circumstance were used.

4.2.1 Ground Bounce Multipath

Ground bounce multipath occurs at low satellite elevation angles. Because the magnitude
of the ground's reflection coefficient increases as the satellite angle decreases, it is assumed that
the worst multipath interference occurs at the lowest satellite elevation angles. For this study, a
satellite elevation angle of 5 degrees was assumed to be a worst-case multipath. In practice,
satellites below 5 degrees of elevation are not typically used for aircraft navigation because of
the high likelihood they are corrupted by multipath. Therefore, in this study, we calculated the

19
ratio between the level of a direct signal at 5 degrees of elevation and a ground bounce signal.
As was shown in Figure 1, the ground bounce signal is left-hand circularly polarized and has a
negative elevation angle. Accordingly, the numbers listed under the smallest ratio column in
Table 3 are the worst-case ratio of the direct right-hand circularly polarized signal at 5 degrees of
elevation to the ground bounce left-hand circularly polarized signal at negative 5 degrees of
elevation.
Changing the azimuth angle of the test antenna affected the antenna's performance.
Table 3 also lists the largest or best ratio of direct to ground bounce signal for each antenna
tested. Azimuth angles are not included in Table 3 because their orientations were arbitrarily
assigned at the start of the data collection. Again, Table 3 shows the data used to assess an
antenna's ability to reject a ground bounce multipath signal. Structure bounce multipath will be
discussed in Section 4.2.2.

Table 3. Peak Azimuth Variation of Ground Bounce Multipath Signal Rejection

Smallest Ratio Largest Ratio

Antenna Ground Plane Direct/Ground Direct/Ground
Structure Bounce (dB) Bounce (dB)
38 (quadrahelix) ground plane 13.9 23.1
Dome and Margolin(dipoles) ground plane 2.8 5.3
Utton (patch) mock fuselage 3.6 10.8
NovAtel (corned fed patch) choke ring 10.3 23.0
NovAtel (comer fed patch) ground plane 3.9 10.8
Trimble Aviation (patch) choke ring 9.4 20.3
Trimble Aviation (patch) ground plane 4.1 14.3
Trimble Aviation (patch) mock fuselage 2.8 8.0
Turbo Rogue (dipoles) choke ring 9.8 15.3

A comparison of smallest to largest minimum ratio of the direct signal to a ground

bounce multipath signal shown in Table 3 indicates that satellite azimuth angle plays a
significant role in an antenna's ground bounce multipath rejection performance. For example,
there was a 13 dB difference in the smallest and largest ratios for the NovAtel (comer fed patch)
on a choke ring. Figure 8 shows more clearly the variation in direct to ground bounce ratios due
to azimuth angle changes. In Figure 8 the direct to ground bounce ratio is plotted for all azimuth
angles tested for the NovAtel antenna on a choke ring. As can be seen in Figure 8, variations in
azimuth angle had a significant effect on this antenna's ability to reject a ground bounce
multipath signal. As previously mentioned, direct signal to ground bounce multipath signal
ratios were calculated at a 5-degree satellite elevation angle.

Azimuth angle was not the only factor which affected the antenna's ability to reject
multipath signals. Ground plane structures can enhance an antenna's performance. For example,
the Trimble aviation patch mounted on a choke ring measured 6.6 dB higher than when mounted
on a fuselage section and 5.3 dB higher than when mounted on a ground plane. Section 5
discusses ground plane structures in more depth.

20
 24
22
D:I
"CI
20
.5
C
c 18
:e 16
'-a:
Q)
Q)
14
.c
11 12
§=- 10
~
Q)
(oJ 8
C
=
c
D:I
6
"CI 4
c
=
e 2
t!'
0
0 45 90 135 180 225 270 315 360

Azimuth Angle in Degrees

Figure 8. Azimuth Multipath Rejection Variation for NovAtel Antenna on a Choke Ring,
5-degree Elevation Angle.

4.2.2 Structure Bounce Multipath and Cross Polarization Ratio

Our means of evaluating an antenna's structure bounce multipath performance is to
determine its minimum cross polarization ratio over the entire gain pattern. Cross polarization
ratio is the ratio between right-hand circular polarization gain and left-hand circular polarization
gain. All satellite elevation angles, above 5 degrees, are considered in this analysis because
structure bounce multipath can occur at any satellite elevation angle. The larger the cross
polarization ratio, the better the antenna will be at rejecting a multipath signal that has bounced
off of a structure.
Nine satellite azimuth angles were tested; minimum cross polarization ratios were
determined for each azimuth angle. Since these antenna minima varied, figures and tables are
presented to provide the reader with a sense of the range of variation due to azimuth angle.
Figure 9 shows the absolute gain for the NovAtel antenna on a choke ring at a relative azimuth
angle of 135 degrees. The minimum cross polarization ratio for this azimuth angle was smaller
than any other azimuth angle tested. Thus, Figure 9 is referred to as the smallest minimum cross
polarization ratio. Similarly, Figure 10 shows the largest minimum cross polarization ratio for
the NovAtel antenna on a choke ring, occurring at a relative azimuth angle of 45 degrees.

21
Table 4 and Table 5 list the smallest and largest minimum cross polarization ratios, respectively,
for all antennas tested. Also listed are the elevation angles where the minimum cross-
polarization ratio occurred. In this computation, only elevation angles greater than 5 degrees
were used because generally satellites below 5 degrees of elevation are not used for navigation.

22
 10

-10

-20 .....

II
I
-30 . . . . ". t ...

-RHCPGain
-40·
-LHCPGain

-50

-60 f---,.---,--,-----'--,..----,-----.-----r------,---,---,-----,----,
0
-30 0
30
Elevation Angle

Figure 9. Smallest Minimum Cross Polarization Ratio Determined From Absolute Gain,
NovAtel (corner fed patch) Antenna on a Choke Ring: Azimuth Angle = 135 degrees.

10

-10

-20 ...

-30 - ...

-RHCPGain
-40 ..
-LHCPGaln

-50 .

-60 f - - - , - - - , - - - , - - - - , - - L , - - - , - - - , - - , - - , - - , - - - , - - - - - , r - - - - ,
-30 0 30 0

Elevation Angle

Figure 10. Largest Minil'num Cross Polarization Ratio Determined From Absolute Gain,
NovAtel (comer fed patch) Antenna on a Choke Ring: Azimuth Angle = 45 degrees.

23
 Table 4. Smallest Minimum Cross Polarization Ratio

Ground Plane Minimum Cross Elevation

Antenna Structure Pol Ratio (dB) Angle
38 (quadrahelix) ground plane 10.4 14°
Dome and Margolin (dipoles) ground plane 3.2 5°
Litton (patch) mock fuselage -{).4 5°
NovAtel (comer fed patch) choke ring 11.1 6°
NovAtel (comer fed patch) ground plane 1.4 14°
Trimble Aviation (patch) choke ring 4.9 5°
Trimble Aviation (patch) ground plane -{).2 5°
Trimble Aviation (patch) mock fuselage 0.2 5°
Turbo Rogue (dipoles) choke ring 6.9 5°

Table 5. Largest Minimum Cross Polarization Ratio

Ground Plane Minimum Cross Elevation

Antenna Structure Pol Ratio (dB)
Angle
38 (quadrahelix) ground plane 15.1 84°
Dome and Margolin (dipoles) ground plane 5.5 5°
Litton (patch) mock fuselage 5.0 5°
NovAtel (comer fed patch) choke ring 17.3 13°
NovAtel (comer fed patch) ground plane 10.5 85°
Trimble Aviation (patch) choke ring 12.5 5°
Trimble Aviation (patch) ground plane 10.3 5°
Trimble Aviation (patch) mock fuselage 10.4 69°
Turbo Rogue (dipoles) choke ring 12.1 12°

A comparison of Table 4 and Table 5 shows the choke ring improved both the NovAtel's
(comer fed patch) and the Trimble's (patch) performance. Improvements between 2 and 10 dB
were measured. Section 5 discusses the effects of these ground plane structures on the antenna's
performance in more detail.

25
 5. GROUND PLA E VERSUS CHOKE RI G

A closer look at the ovAtel (corner fed patch) and Trimble (patch) antennas'
performances shows a significant improvement in cro polarization ratio when these antennas
were mounted on choke rings. Figure 11 represents the absolute gain for the ovAtel antenna on
two different ground plane tructures: a choke ring and a ground plane. For these absolute gain
plot . the azimuth angle of the antenna when it was tested was the same. It can be seen that the
cross polarization ratio was greatly improved by mounting this antenna on a choke ring. As
previously discussed, increasing the cross polarization ratio increases the antenna's multipath
rejection capabilities. Figure 12 shows a similar gain plot for the Trimble antenna on two
different ground plane structures: a choke ring and a ground plane. Again, cross polarization
ratio was improved across all elevation angles when the antenna was mounted on a choke ring.

10 - - - - - ..

o - - - - .

Minimum Cross Polarization Ratio

m
'tJ
-30 . - -.- - -.. - -.- - -.. - - -.- - .
RHCP Ground
LHCP Ground
-40 ..
RHCP Choke
LHCP Choke
-50 - - - -- - - -.. - ...

-60 f----.-----.,-------,---,..----,----,------,---.-----.,-------,---,..----,-

-30 0 30 0
Elevation Angle

Figure 11. Ground Plane Stntcture Comparison for NovAtel (comer fed patch) Antenna.
Ground Plane vs. Choke Ring.

27
 10 .-----. -- ... -.. ---.------ .. --------- .------.------.------------------------ .. ----- .. -----. ---

o .--.-- - .

-10 .-- .... -- .. ~_..~

.. ';..- __

·20 . . .... ~

,
. t - .. , - - . - - ..• - - . . . - .• -, .- - •.•. - • • - - ., .. - - . - ~ .... - - . - - ...... - . . - . '•.. - .... . • • • - -... - -
~

ttl
'C
-30 - '" - -.-- -.-- -. .--" -.- - -.-- -. -- - - - -.- -.. - - -.-.
RHCP Ground Plane
LHCP Ground Plane
-40 - , -..' - - -.- -- -. "- -- -.--.. "- -.-- - -..
RHCP Choke
LHCP Choke
-50 -- _ - -.--.... . .- - _ -_ - -.-- -- - - - - .. -- - -.. - -

.60 1 - - - , . . - - - - - , - - - - - - - - - , - - - - - , - - , . - - - - - . - - - - - - , - - - - - , - - - - - - - , , - - - . - - - - . - - - - - ,
.30 0
30 0

Elevation Angle

Figure 12. Ground Plane Structure Comparison for Trimble (patch) Antelllla..
Ground Plane vs. Choke Ring.

29
 6. OTHER FREQUE CIES

Three of the antennas were tested at frequencies other than Ll (1575 MHz). Aside from
changing the frequency, these data were collected in exactly the same manner as described in
Section 3. The 3S (quadrahelix) antenna (a GLONASS antenna) was tested at 1609 MHz, which
is the median frequency in the GLONASS frequency allocation. The Dome and Margolin
(crossed dipoles) and Turbo Rogue (crossed dipoles) antennas were tested at L2 (1227.60 MHz).
Absolute gain computations were made on these data in the same fashion as described in
Section 4 for data collected at Ll (1575.42 MHz). Figures 12 through 14 are the absolute gain
plots where the smallest minimum cross polarization ratio was measured. These plots represent
the worst-case cross polarization ratio for each antenna. Cross polarization ration is used to
determine an antenna's ability to reject a structure bounce multipath signal.

10 .. , -.. __ ._- .__ .- .. -.. - , _.. _.. _ _.--

o ._ --. --- -_ '.:.::.:;'~,; ;.; .; ,; ; ; .----~--.~_-.-:'--~~--

:. ..~-..

-20

-30 .-.. -.. - -_ - _ -- -.. --. - - , _ _.. -..

- RHCPGaln
-40
- LHCPGaln

-50 .--.-- -.--- --.. -.--.- -. -- ..

-60 t---.----,---,----r----r--r--.----,---,----r----r---,
0
-30 0
30
Elevation Angle

Figure 13. Absolute Gain for 35 (quadrahelix) Antenna on a Ground Plane at 1609 MHz:
Minimum Cross Pol. Ratio =8.4 dB at 89-degree Elevation.

31
 10 "'" _ _ _ _ ..

o .. .. ._ ~~_~~ ===._
.. ~~_~---........~

-10· .. _........ ·

-20 _ - _ _ ,.

-30 -.. . .. .. . . .. .. .. . - -- -- - . ... .. .. ... . - ..

-RHCPGain
-40 - _ _ -. _ LHCP Gain

·50 -. .... . .. . . ... . ... . . .. . - -.. - -. .. .. . . .. . ..

-60 I---,__--.---__,__--..---r-----r--~-,__--.---__,__-----,---

-30 0 30 0
Elevation Angle

Figure 14. Absolute Gaill for Dome & Margolin (crossed dipoles) Antell II a Oil a Groulld Plalle at
1227 MHz: Millimum Cross Pol. Ratio = 1.9 dB at 6-degree EleWltioll.

33
 10 .

-10 .

-20 ....

m
'C
-30

-RHCPGain
-40 ._ .
-LHCPGain

-50 .

-60 1-----.-----,-------,--.-----,-------.--.-------.----,-------,--.-------,
-30 0 30 0
Elevation Angle

Figure J 5. Absolute Gain/or Turbo Rogue (crossed dipoles) Alltenna all a Choke Ring at 1227 MHz:
Mini/1/u/1/ Cross Pol. Ratio = 12.6 dB at 5-degree Elevation.

35
 7. CONCLUSIONS

The GPS antenna evaluation conducted at Lincoln Laboratory tested a total of six
antennas mounted in various ground plane structures. The antennas were evaluated on the basis
of their ability to reject two different types of multipath: ground bounce and structure bounce.
Antennas were tested in an anechoic chamber. Linearly polarized orthogonal polarizations were
transmitted; receive amplitudes and phases for each antenna were recorded. These data were
converted to right-hand and left-hand circularly polarized signals using well-known
mathematical methods.
Antenna performance evaluations were made based on each antenna's absolute gain
properties. To asses an antenna's ground bounce multipath rejection performance, ratios of
direct to ground bounce signals at a 5-degree elevation angle were calculated. Structure bounce
rejection performance was determined by computing a minimum cross polarization ratio of
elevation angles from 5 degrees to 90 degrees from an absolute gain data. Different ground
plane structures were also used to see if changing the mounting environment for an antenna
improved its overall performance.
Each antenna's performance was assessed using the ratios of direct signal gain to ground
bounce signal gain at a 5-degree elevation angle and the minimum cross polarization ratio for all
satellite elevation angles greater than 5 degrees. The antennas that performed best were the
NovAtel (comer fed patch) on a choke ring and the 35 (quadrahelix) on a ground plane. Both
had direct to ground bounce signal ratios and cross polarization ratios of 10 dB or greater for all
azimuth angles. Therefore, these two antennas would be very good at rejecting either type of
multipath.
The antennas that performed fairly well were the Turbo Rogue (crossed dipoles) on a
choke ring and the Trimble (patch) on a choke ring. These antennas had minimum direct to
ground bounce ratios of nearly 10 dB and minimum cross polarization ratios of approximately
5 dB. All other antennas tested did not perform as well as these.
Ground plane structure did improve an antenna's ability to reject a multipath signal. For
instance, the NovAtel (corner fed patch) antenna increased its smallest minimum cross
polarization ratio by nearly 9 dB when it was mounted on a choke ring instead of a ground plane.
Similarly, the Trimble (patch) antenna gained 5 dB in its smallest cross polarization ratio when
mounted on a choke ring. Though cross polarization is a measure of ability to reject structure
bounce multipath, similar results were attained when evaluating ability to reject ground bounce
multipath. The NovAteI and Trimble antennas improved 6 dB and 5 dB, respectively, when
mounted on a choke ring rather than a ground plane, when assuming a ground bounce multipath
signal occurred at a 5-degree elevation angle.
In addition to varying the ground plane structure, three of the antennas were tested at
frequencies other than 1575 MHz. The Turbo Rogue (crossed dipoles) and the Dome and
Margolin (crossed dipoles) were tested at 1227 MHz. The Turbo Rogue antenna performed very
well at 1227 MHz. Minimum cross polarization ratios exceeded 12 dB for all azimuths. The
Dome and Margolin antenna did not perform as well, providing minimum cross polarization
ratios ranging from 2 to 5 dB. The 35 (quadrahelix) antenna was tested at 1609 MHz, and

37
performed as well at this GLONASS frequency as it did at L1 (1575 MHz). Minimum cross
polarization ratios at 1609 MHz ranged from 9 to 17 dB.
This study was designed to determine which of the GPS antennas provided the best
reduction of multipath signals. Future studies will focus on characterizing environmental
multipath at any given GPS reference location. The latter study, combined with the present GPS
antenna evaluation should allow the user to determine the best antenna type and GPS reference
station location to mitigate the effects of mu1tipath.

38
 APPENDIX

Plots of absolute gain in dBi versus satellite elevation angle are shown in Figures A-I
through A-9. In practice, satellite elevation and azimuth angles vary with respect to the GPS
antenna. Since it was not practical in a testing environment to move the source all over the
anechoic chamber. the antenna was rotated in azimuth and moved at I-degree increments in
elevation. For each sweep through elevation angle, azimuth angle was held fixed. Azimuth
orientation was varied from 0 degrees to 360 degrees in 45-degree increments. Reference marks
were made on each antenna to allow antenna testing personnel to realign the antenna to a
O-degree relative azimuth angle.
Each plot in this appendix represents the absolute gain for the worst-case azimuth
orientation, that is, the orientation with the smallest minimum cross polarization ratio. For a
comparison of worst-case to best-case cross polarization ratios. see Table 3. Since the initial
marking of a O-degree azimuth was arbitrary, azimuth information is not included with these
plots. What is important to note is antenna performance nuctuates depending upon satellite
heading and elevation angle. These plots represent the worst-case scenario in that fluctuation.

10 -- --.-.--.- - - -.--.--.- .. - .

o .. - , "" .:;.;;000- - - ~~~ ~~~

•• • •.•• • .•• -
•.•:•'•
: •••• == .
~~

-30

- RHCPGain
-40 .. ·.. ········· .. ··
- LHCPGain

-50 .

-60 1---,----,------,------.----,--,----,---,------.----,--,-----,

-30 0 30 0
Elevation Angle

Figure A-I. Absolute Gain/or 35 (quadrahelix) Antenna on a Ground Plane:

Minimum Cross Pol. Ratio = 10.4 dB at I4-degree Elevation.

39
 10 - - - - - "- " - - - - - .

-10

-20 _ _ _ _ _ - _ ' .

m
"0
-30 . _ , - - " - - .

- RHCPGain
-40 _ _ LHCP Gain

-50 , -" _ - - ' .

-60 I---..---,----.---r--,--~----,-------.------,--,.------.---__,

-30 0 30 0
Elevation Angle

Figure A-2. Absolute Gain/or Dome & Margolin (crossed dipoles) Antenna 011 a Ground Plane:
Minimum Cross Pol. Ratio = 3.2 dB at 5-degree Elevation.

41
10 .

o .

-10

-30 _. _ - _.

- RHCPGain
-40 - LHCPGain

-50

-60 j----,--,-----.-----r----.----,---rr--,--,----,---,---,

-30 0 30 0
Elevation Angle

Figure A-3. Absolute Gaill for Litton (patch) Antenna:

Minil11ul11 Cross Pol. Ratio = -0.4 dB at 5-degree Elevation.

43
 10 .. - _ _ _._ _._ _..

o ._ _ - _ - _ .

·10 __ _ __ _ __ - - _ _ .

·20 .-- --'" .. -- -. .. ".. -- - - -. .. .. --. ... . .

m
'C
·30 -. -. - . . .. . . . "- _.. .. .. "- - -- .- - .. - - ".. -- .

- RHCPGain
·40 .. __ _ _ _._ _ _- -.- _ __ _. _ LHCP Gain

·50 .. - -.- --.. . .. -" . -.. - --.- _ - -.- -.- - - - - .

.60 f - - - , - - - - . - - - - - . - - - , - - - . - - - . . . , - - - , - - - - - , - - - , - - - , - - - - , - - - - ,
.30 0
30°
Elevation Angle

Figure A-4. Absolute Gain for NovAtel (corner·fed patch) Antenna on Choke Ring:
• Minimum Cross Pol. Ratio = 11.1 dB at 6-degree Elevation.

45
 10 _ __ _ - - .

o . . - .

-10 - .

-20 .. .. -. .. -.... .. .., . .. . - - ... .. . .. .. - _..

CD
'C
-30 .".. -.. '" .. .. .. -- - - - -.- - " .

-RHCPGain
-40 .. - ....
-LHCPGain

-50 -- .. .. . - """" -- -- -. -- .. - - - - .. - .

-60 I-----r--..,..----,--------r-----,-----,--r---r--..,..----,--------r-------,

-30 0 .30 0
Elevation Angle

Figure A-5. Absoillte Gain for NovAtel (comer-fed patch) Antennll 071 Gr01l7ld Plane:
• Minimllm Cross Pol. Ratio = 1.4 dB at 14-degree Elevation.

47
10 - - ..

o _ .

-10 .

-20 .

-30 .. .. . .. .. .

- RHCPGain
-40 .
- LHCPGain

-50

-60 f---~--.----.--.....----r----.-- __r--~--.----.--.....----,

-30 0 30 0
Elevation Angle

Figure A..6. Absolute Gain for Trimble (patch) Antenna on Choke Ring:
Minimum Cross Pol. Ratio = 4.9 dB at 5-degree Elevation.

49
 10 - -....... . __ _ __ ._ - __ _ _ .

o _ _..

• -10 ..
£--.~~~~~~~~~~-.~~~~ .. .. .. _ _ --.. _ .

-20

aJ
'C
-30 -- ....

- RHCPGain
-40 _ -.. .. . . - . .. .. - - -- . - . - -
- LHCPGain

-50 ".. . .. -.. - _ -... . . - -. ... . ... .. _ -.. -- '" .. -

-60 r---.,__--r------r-------r--..---~-__._-__,_--.----.,__--r--___,
-30 0 30 0
Elevation Angle

Figure A-7. Absolute Gain/or Trimble (patch) Antenna on GrOllnd Plane:

Minimum Cross Pol. Ratio = -0.2 dB at 5-degree Elevation.

51
 10 _ " _ " _ _ .

o- - " - ':.: . ".-..,. .

-~, - ".

-20 ... .. . .. - - - " -.- '" .

m
'0
-30 .. - - -.. . - .

- RHCPGain
-40 - - " -.- - -- - ..
- LHCPGain

-50 " - " - " - " " " -. '" " - -.. . .

-60 f---...,---,------,------r-------,-------,---,--...,---,------,------r----,

-30 0 30 0
Elevation Angle

Figure A-8. Absolute Gain for Trimble (patch) Antenna on Fuselage:

Minimum Cross Pol. Ratio = 0.2 dB at 5-degree Elevation.

53
 10 _ _ - .

o .: .:. _.. --_ ~~~~ .

-10 .

-20,·",·"

m
'C
-30 .. .. .. . ... ..,..... ... " "--.. _ -.. - " -.. - - ".. -.. - "..

-RHCPGain
-40 - - _ _ _.. - _ .
-LHCPGain

-50 . ,. - _ - _ - _ - - .

-60 I - - - . - - - - . . , - - - - . - - - - - r - - - - , - - - - - r - - , . - - - . - - - - . . , - - - - . - - - - - r - - - - ,
-30 0 30 0
Elevation Angle

•
Figure A-9. Absolute Gain for Turbo Rogue (crossed dipoles) Antenna:
Minil1lul1l Cross Pol. Ratio = 6.9 dB at 5-degree Elevation.

55
 REFERENCES

[1] Skolnik, Merrill I., "Introduction to Radar Systems," Second Edition, McGraw-Hill Book
Company, New York, 1980, p. 505 .

[2] Ibid, p. 506.

[3] Lo, Y. T.. and S. W. Lee; "Antenna Handbook," Van Norstrand Reinhold Company, New
York; 1988, p. 8-50.

[4] Braasch, Michael S., "On the Characterization of Multipath Errors in Satellite-based
Precision Approach and Landing Systems," Dissertation Presented at Ohio University,
• June 1992, p. 9.

57