Interference Effects on Galileo E5b and E6 Frequen- cies ... - IFEN

<strong>Interference</strong> <strong>Effects</strong> on Galileo E5b and E6 Frequencies
observed in the Galileo Test and Development
Environment GATE
BIOGRAPHY
Guenter Heinrichs, Erwin Loehnert, Bjoern Ott and Elmar Wittmann, IFEN GmbH
Dr. Günter Heinrichs received a Dipl.-Ing. degree in
Communications Engineering from the University of Applied
Science Aachen in 1988, a Dipl.-Ing. degree in Data
Processing Engineering and a Dr.-Ing. degree in Electrical
Engineering from the University Paderborn in 1991
and 1995, respectively. In 1996 he joined the Satellite
Navigation department of MAN Technologie AG in
Augsburg, Germany, where he was responsible for system
architectures and design, and digital signal and data
processing of satellite navigation receiver systems. From
1999 to April 2002 he has served as the head and R&D
manager of MAN Technologie’s satellite navigation department.
In May 2002 he joined IFEN GmbH, Poing,
Germany, where he is currently the head of the customer
applications department and business development unit.
Erwin Löhnert received a diploma in Aerospace Engineering
in 1993 from the Munich University of Technology.
In 1994 he joined the Institute of Navigation and
Geodesy of the University of the Bundeswehr Munich as
a Research Associate, working mainly for aerogravimetry
and GPS/INS integration. In 2000 he joined IFEN GmbH
as a Project Manager for integrity determination. Since
2001 he is Head of the Mobile Solutions department,
managing several projects and currently being the technical
manager of the GATE project.
Elmar Wittmann received a Dipl.-Ing. degree in Geodesy
from the Munich University of Technology in 2000
and then joined the IFEN GmbH, where he is now working
as a systems engineer in the field of GPS/Galileo satellite
navigation and software development for mobile
applications.
Björn Ott received a degree in aerospace engineering
(Dipl.-Ing.) from the Technical University of Berlin in
1994. In 1996 he entered the field of satellite navigation
as a research associate at the Institute of Geodesy and
Navigation of the University FAF Munich. He focused on
Pseudolite signal characteristics and propagation effects.
After positions in GNSS engineering with Thales ATM,
Telematica and TeleConsult Austria, he joined IFEN
GmbH in 2006 to lead the development of a Chinese outdoor
Galileo test environment.
INTRODUCTION
GATE is the only Galileo test and development environment
worldwide where already today navigation is
possible with realistic Galileo Signals on all frequencies
in an outdoor area. GATE is a local ground-based “miniature
navigation system” which will provide real Galileo
signals already several years before the full operability of
the Galileo system in space. Thus, GATE is an important
intermediate step for Galileo on its way from the laboratory
to the orbit in terms of realistic (RF) signal transmission.
GATE has successfully passed the final system performance
acceptance testing and pre-operational optimization
phase in early July 2008 and has entered the commercial
operation phase starting on August 1, 2008.
GATE operates according to the same physical principles
as Galileo and GPS to allow users to determine their
position, velocity and time (PVT). The receiver can determine
its PVT by calculating the distances to the virtual
satellites, emulated by a transmitter station, and finding
the intersection point. Through its infrastructure, GATE
is able to radiate the original navigation signals from
Galileo satellites, to simulate natural influences like ionosphere
or troposphere delays, to change characteristic
parameters of signals and to adapt the signal strength as
required.
During the GATE system performance tests, that have
been performed in the test area in Berchtesgaden/Germany
to establish the user positioning performance,
interference effects on the Galileo E5b and E6 frequency
bands have been detected that caused a performance
degradation. After extensive spectrum evaluation
and analysis, it was found that the interference effects are
caused by a military aerial surveillance radar, operating in
the Galileo E5b and E6 frequency bands.
The paper will give an overview on the interference
measurements, the spectrum evaluation and analysis performed
so far. It describes the test activities and shows
the test results dedicated to the user positioning performance
degradation measurements due to the detected interference
effects on the Galileo E5b and E6 frequency
bands. Finally, an evaluation of the influence of the detected
interference effects on the GATE receiver measurements
is presented.

GATE INFRASTRUCTURE & TEST AREA
GATE System Architecture
The GATE system is partitioned into four segments -
Transmit Segment (GATS), Mission Segment (GAMS),
Control Segment (GCS), and the Support Segment. Figure
1 presents an overview of the GATE system architecture.
GTF
Timing
Facility
GMCF
Control Segment
GPS
GMS
GPF
GAMS
SIS
Mission Support
Facility
Steering
Ground (Field) Segment
Support Segment
Galileo IOV
6 Transmit
Stations
GATS
User Trajectory
GATE
User Receiver
Figure 1: GATE Infrastructure Overview
User
GSTB
GIOVE
The ground-based transmitters, which are part of the
GATE Transmit Segment (GATS), emit signals on all
Galileo frequencies. They are flexible in signal generation
and adaptive to changes in signal structure. As
GATE is a real-time system, it is necessary to feed the
navigation message in real-time to the transmitters. They
are also equipped with stable atomic clocks.
The GATE Mission Segment (GAMS) monitors the
navigation signals by using two GATE Monitoring Stations
(GMS), performs the time synchronisation of all
system clocks and generates navigation messages and
steering commands to be sent to the six transmitters. The
tasks denoted above are mainly performed by the two
GAMS core elements, the GATE Processing Facility
(GPF) and the GATE Monitor Receivers (GMRx), both
developed by IFEN GmbH.
The GATE Control Segment (GCS) includes all the
functionality and facilities that are required for the mission
control and operation. The main tasks are to monitor
and control the entire GATE system, to host and operate
the control centre, to provide accurate and stable GATE
system time aligned with TAI, and to archive GATE mission
data.
The main tasks of the GATE Support Segment (GSS)
finally comprise the appropriate preparation, i.e. simulation
and planning, of the GATE experiments with dedicated
software tools, as well as the provision of the
GATE User Terminals equipped with a combined Galileo/GPS
receiver.
GATE Test Area Berchtesgaden / Germany
The GATE test area is located in the region of Berchtesgaden
in the very south-eastern part of Germany / Bavaria.
The service area is depicted in the maps shown in
Figure 2 a and b. The GATE test area, which is roughly
limited by the imaginary connection lines between the
signal transmitters, has a size of about 65 km², while the
GATE core test area, as marked in Figure 2 b, is about 25
km². The two monitoring stations are located at an exposed
central site GATE test area. As can be seen on figure
2c, Berchtesgaden is surrounded by high mountains,
that are rising up to over 2000m. The installation of the
GATE transmitters on well exposed positions allows for
emission of the GATE signals with average elevation
angles between 10 to 15 degrees from a user’s point of
view, when located within the GATE test area.
Figure 2 a, b, c: GATE Service Area Berchtesgaden / Germany (core test area marked in red, transmitters
marked as red dots) and view into the GATE test area

INTERFERENCE MEASUREMENT SET-UP IN
THE GATE TEST AREA
In the framework of the GATE system AIV (Assembly,
Integration and Verification) activities extensive system
tests have been performed for all three operational GATE
modes, i.e. the Base Mode (BM), the Extended Base
Mode (EBM) and the Virtual Satellite Mode (VSM). In
BM the actual GTS positions are transmitted in the navigation
message at constant power levels, this means the
transmit platforms act like pseudolites. In EBM the Base
Mode is extended by a dynamic adjustment of the signal
power levels according to the current user position, which
is fed back into the system via data link. In VSM a virtual
dynamic Galileo satellite constellation is emulated. The
Doppler shift and power levels of the transmitters are
steered to be identical with signals coming from orbit.
For the evaluation of the system performance within the
GATE core area a test vehicle equipped with the GATE
user receiver and terminal was used. With this vehicle
static system tests as well as dynamic tests can be performed.
In order to provide an independent and very accurate
position reference a GPS RTK solution was applied
using high-precision differential corrections. The
accuracy of the RTK positions was typically in the range
of a few centimetres in static scenarios and several decimetres
within dynamic tests. In this way it is made sure
that on the one hand the steering of the signal transmitters
in the Extended Base Mode and Virtual Satellite Mode is
adjusted properly. On the other hand the precise RTK
position serves as a reliable reference for the performance
evaluation with respect to the absolute GATE accuracy
achieved.
The GATE user receiver as well as a dual-frequency
GPS RTK receiver for the position reference were installed
in the test vehicle (see Figure 3). On the roof of
the van two separate antennas for the receivers were
mounted close to each other.
Figure 3: GATE test vehicle and test setup
Various tests with duration of at least eight hours have
been performed for the three different GATE modes in
order to verify both, system stability and performance
during the maximum daily operating time specified.
The results obtained from these tests gave proof of the
compliance of the GATE system with the performance
specifications, at least for the E1 band. The GATE specification
for the horizontal real-time navigation requests a
positioning accuracy with GATE signals better than 10 m
(2 sigma) with an adequate HDOP (Horizontal Dillution
of Precision) value. The tests showed that this accuracy
can also be achieved with the E5a and E5b signals, however,
often some degradations of the positioning performance
were observed for these frequencies. Some sample
results of the static GATE positioning tests are presented
in figures 10, 11 and 12. After various analyzes of the E5
test results there was clear evidence to suggest that those
degradations may be caused by signal interference within
the E5 band.
Thus, intensive investigations were made with respect to
potential GATE/Galileo interference sources in the area
of Berchtesgaden and surroundings. In the scope of these
analyses several measurement campaigns were performed
in order to detect potential interference sources as well as
to evaluate the impact of interference sources already
identified. Measurements have been performed by IFEN
GmbH and by the Bundesnetzagentur BNetzA (German
Federal Network Agency), which is responsible for securing
efficient and interference-free use of frequencies in
Germany.
Figure 4 a, b: The BNetzA measuring vehicle during
interference measurements in the GATE area
At the GATE Monitoring Station (GMS) a spectrum
analyzer and digital oscilloscope were connected to one
GMS antenna and the relevant frequency bands were
scanned. For the mobile interference measurements in the
area of Berchtesgaden and surroundings, the test car was
additionally equipped with a directional antenna which
was connected to the measurement devices.
In addition to this measurement campaign the German
Federal Network Agency (BNetzA) performed several
measurements with its specially equipped measurement
vehicle in the area of Berchtesgaden. Figure 4 shows this
vehicle equipped with a telescopic antenna mast during

static measurements close to the GMS location (4a) and
at the GATE central point (4b).
INTERFERENCE MEASUREMENT AND SPEC-
TRUM EVALUATION RESULTS IN GATE
Analysis of the spectrum analyzer measurements for the
Galileo frequencies showed that interference signals are
present just outside of the ITU-reserved E5 band and
right within the E6 band. The interference signals exhibited
quite similar characteristics on two frequencies of
interest:
E5 band:
- center frequency: 1217 MHz
- “always on”
E6 band:
- center frequency: 1290 MHz
- “sporadic presence”
Common
- measured bandwidth: > 5 MHz
- pulse spectral power: ca. –75 dBm
- complex pulse sequence and shaping
- average PDC: ~ 5%
The 1217 MHz signal is constantly present, while the
1290 MHz signal is only present at certain times of the
day. After clear reception for a couple of hours, the signal
ceases for another couple of hours, and then comes back.
There are indications the signal is switched to other frequencies
meanwhile. No obvious schedule for this characteristic
could be established.
Snapshots of the interference signals observed at the
spectrum analyzer are depicted in Figure 5 below (E5 in
Fig. 5a and E6 in Fig. 5b).
Figure 5 a, b: <strong>Interference</strong> signals observed at GATE
monitoring station
The interference signals received at the GMS have then
been analyzed in the time domain. Figure 6 shows the E5
recurring pulse sequence. The pulse sequence is rather
complex with pulses of variable duration from 50 to 200
microseconds.
Each of the pulses is itself shaped in a non-repetitious
form. An enlarged representation of a single 200 microsecond
pulse is presented in Figure 7. The pulse gaps are
1.6 ms and 3.5 ms respectively. A 24.7 ms long sequence
of 8 short pulses and 3 long pulses is repeated continuously.
Figure 7 also illustrates the signal measurements in
the receiver front-end.
Figure 6: interference pulse sequence on E5
The signal displayed green was measured before the
AGC (Automatic Gain Control) amplifier and the one
shown in pink after AGC. The latter gives clear evidence
of the AGC going into compression during the pulse.
Essentially what happens is pulse clipping in the analog
domain. This is a desirable effect because the total power
entering the next stage, the A/D converter (ADC), is limited.
An elongation of the pulse due to discharging of
capacitors in the RF system after it has been driven into
saturation (recovery time) is not evident in Figure 7.
Figure 7: E5 interference “long” pulse
In addition to the characteristics described above, the
received interference signal exhibited a superimposed
10 second oscillation. The 10 second interval is assumed
to stem from a rotating transmission antenna, which is
typical for a radar installation. The overall signal characteristics
let us assume that the interference is emitted by a
military long range air surveillance radar. The pulse duration
is longer and pulse complexity higher than that of
civil ATC radar (whose characteristics are rather well
known).
Figure 8 shows an example of a “typical” long range
ATC radar interference signal. It was measured in the
vicinity of a civil ATC radar station (Maitenbeth, GER)
using the same test setup as described above.

Figure 8: Signal pulses of civil ATC radar
Characteristics of this ATC radar simple pulse pair are
- Pulse length 2 µsec
- Pulse gap 2.5 µsec
- Pulse repetition 1.9 to 2.6 msec
- Center frequency 1259 MHz
- Observed spectrum 27.5 MHz wide
The results of the corresponding interference measurements
performed by the BNetzA largely confirmed the
observations described above. The analysis of the
BNetzA campaigns can be summarized as follows:
- E5 band (1155-1228 MHz): An interference spectrum
with centre frequency 1217 MHz was detected
- E6 band (1258 – 1300 MHz): An interference spectrum
with centre frequency 1290 MHz was detected.
- Main result of the analysis according to BNetzA statement:
The interference effects were clearly caused by
radar transmission, presumably from military airspace
surveillance. An inquiry at the German air-traffic control
(DFS) confirmed this assumption.
Figure 9: Kolomansberg Radar site (Austria)
Image source: http://upload.wikimedia.org/wikipedia/commons/b/b2/
Galerie053_768x509_1172521005.jpg
Based on the measurements and the corresponding further
investigations, the Austrian military radar installation
located at the mountain Kolomansberg near Salzburg
could be identified as the most probable interference
source. Figure 9 shows the stations on top of the mountain.
It is obvious from the visible radomes that the site
operates two radar antennas. This would match the observed
interference of two different radar frequencies,
one of those variably present in the E6 band. According
to public sources the station operates radar systems of the
type RAT-31 DL. Its reported pulse power (EIRP) is 84
kW and the typical coverage is about 450 km. The system
is reported to be in operation (in Europe) also in Denmark,
Greece, Poland, Czech Republic, Turkey and Hungary.
The German Bundeswehr recently procured a mobile
version of this radar (RAT-31 DL/M).
A RAT-31 DL radar transmits long pulse bursts which
are compressed in the receiver chain to maximise range
resolution and resistance to electronic countermeasures.
Low- and high-pulse repetition frequency waveforms are
used to provide long-range and medium-range/Moving
Target Indicator (MTI) coverage respectively. Pulse durations
can be up to 800 microseconds. Fixed and adaptive
notch MTI filtering is used.
The technical characteristics of this radar station are
summarized in the table below.
RAT31DL technical data (public sources)
Manufacturer SELEX Sistemi Integrati
(former Alenia Marconi Systems
AMS)
Frequency (range) D-band / 1-2 GHz
Peak power: 84 kW / +79.2dBm (-15dBm @
1 km FSPL)
Average power: 2,5 kW / +64dBm
Displayed range: 500 km
Antenna rotation: 10 seconds / 6 seconds
Features: Frequency agile
Pulse compression
Pulse form agile
Pulse length up to 800 µsec
The line-of-sight distance from the radar site to the GMS
antenna is ca. 37 kilometers. From the measurements and
RAT31-DL specification, a path loss of around 184 dB
has been calculated. The free-space path loss of this distance
would be roughly 125 dB. The complete attenuation,
due to diffractive signal dispersion on the hilly
grounds, matches prediction of the ITM Longley-Rice
Model using 1km resolution GLOBE terrain data. It is
obvious that local environmental conditions (humidity of
ground, foliage, atmospheric humidity, snow, etc.) cause
the actual received signal power to vary from day to day,
or even slowly change during the course of a day.
POTENTIAL GATE USER POSITION PERFORM-
ANCE DEGRADATION DUE TO DETECTED IN-
TERFERENCE
In Figures 10 and 11, GATE positioning results of the
official 8-hours static acceptance tests for the E1 band are
depicted.
The signal interference in the E5 and E6 band as identified
above has some negative effects on the reception and

processing of the Galileo/GATE signals in the GATE test
area. The main effects are described in this section.
The effects on a Galileo/GATE receiver, in general, depend
on the received interfering signal features and the
receiver architecture. They are mainly characterized by
the peak power, the pulse repetition rate, and the pulse
width. The receiver architecture determines if and to what
extent the interference effects become operative. Strong
pulses drive amplifiers into compression during the presence
of the pulses and beyond, since active components
require a recovery time.
Figure 10: GATE long-term static position results
E1 band, extended base mode
AGC effects are of prime importance with pulsed interference.
The task of an AGC is to properly set the ADC
input level. This minimizes signal-to-noise losses due to
quantisation. ADC quantisation losses for multi-bit ADC
can increase significantly if the signal input at the ADC
input is not properly balanced. Strong pulses leaking
through AGC may also saturate the ADC. This in consequence
would suppress the pulse and the Galileo signal
and limit the amount of pulse energy that enters into the
correlators. Strong pulses potentially leaving the compressed
AGC amplifier may drive the ADC into compression
and will be digitally blanked.
Figure 11: GATE long-term static position results
E1 band, virtual satellite mode
Pulse duration and sequence characteristics have an impact
in relation to receiver internal loops. The identified
100 µs and 200 µs long pulses overlap one or two E5b
spreading code sequences. One integrate-and-dump takes
4ms, hence 200µs amount to 5%. While it is true that the
pulses “punch holes” into the integration, the tracking
system will just roll over this as it has a time constant in
the order of 1/10 or 1 second. Effectively, these pulses
are averaged. On the other hand, the receiver system will
follow the 10 sec variation in power level.
The overall situation will deteriorate when pulsed interference
combines with static or variable multipath; as has
to be expected with stationary terrestrial transmitters and
never-changing propagation paths. With respect to the
GATE system performance, it has to be considered that a
wrong lock, e.g. to stationary multipath, of the GATE
monitoring receiver has a different overall effect than a
wrong lock of a mobile user terminal receiver. In virtual
satellite mode (VSM), a prevailing pseudorange error of
the monitoring receiver will cause a bias in the frequency-specific
steering of virtual satellites, which will
than become effective for the mobile receiver as pseudorange
bias of the affected signal. This effect is analogous
to DGPS reference receiver biases. In addition, the filtered
clock estimations will deteriorate over time. A
wrong pseudorange of the mobile user receiver, on the
other hand, results only in a momentary incorrect position
solution. In kinematic mode, the receiver will “switch”
between correct and wrong locks. In result, there will be
multiple “clusters” of positioning errors; as typically
shown in figure 12 for an E5 position calculated from
static VSM measurements. The error clusters are caused
and determined by recurring multipath errors of the same
magnitude.
Figure 12: GATE E5 position result in VSM showing
typical error clusters
Of course there is no option to switch off or otherwise
eliminate interference signals from being, due to numerous
political and jurisdictional reasons; ITU band allocations
being paramaount among them. Accordingly, technical
means to minimize the effect of interference have to
be defined, implemented and verified.
Potential interference mitigation strategies on technical
level are
- Pulse clipping
- Analog or digital pulse blanking
- Advanced techniques
x

Pulse clipping is a simple technique that is effective
with, and inherent to, single-bit ADC receivers. In effect,
it works similar to digital pulse blanking in multi-bit
ADCs, but with much less complexity.
Analog pulse blanking by a discrete pulse detection and
attenuation circuitry has been among the first proposals to
mitigate DME interference to GPS L5, dating back to
around 1998. Later it has been found much more effective
to digitally blank pulses. Digital pulse blanking requires
multi-bit ADCs. Thresholds for excessive power
are defined, and, if exceeded, the instantaneous measurement
is zeroed out or replaced with random values. In
theory, the pulse interference effect is limited to a temporary
loss in correlation energy, but is eliminated from
affecting the critical AGC setting and C/N0 calculation
that may cause a loss-of-lock. The technique requires
proper layout and smart setting of thresholds in order to
work properly. The implementation of a blanking threshold
inherently introduces “weak pulses”, which are below
the threshold. The residual effect of weak pulses needs to
be taken into account.
Advanced techniques comprise detection and elimination
of pulses in the digital domain after (for example)
FFT or wavelet transforms. These techniques are currently
under development with a number of instututions;
their description is beyond the scope of this paper.
The described effects of interference and their mitigation
possibilities have significant impact for teh design of
receivers providing safety-of-life critical navigation and
integrity service. The exact level of performance and the
required settings for performance testing under stress
conditions need to be defined. For civil aviation, Minimum
Operational Performance Specifications (MOPS),
e.g. the well-known RTCA document DO-229D, define
performance and acceptance criteria for such receivers. In
the currently ongoing standardization activities for civil
aviation receivers by European and U.S. bodies, this is
one of the matters under scrutiny by leading experts.
CONCLUSIONS
GATE is a terrestrial test environment for developers of
Galileo (Galileo/GPS) receivers, applications and services.
The test range is situated in the region of Berchtesgaden/Germany.
GATE has reached its full operational
capability (FOC) on August 1, 2008. The terrestrial test
bed is considered to be a necessary intermediate step for
Galileo from laboratory into orbit in terms of realistic RF
signal transmission.
GATE provides the opportunity for receiver, application
and service developers to perform realistic field-tests of
hardware and software for Galileo at an early stage, i.e.
several years before the full operability of Galileo. The
realism of the signal environments includes a well observed
levels and characteristics of typical pulsed interference.
For further information on GATE please refer to
the official project homepage http://www.gatetestbed.com.
During the GATE system testing phase significant interference
effects on the E5 and E6 band were observed.
Intensive investigations revealed a military air surveillance
radar station as the source of the interfering signals.
The presence of this radar interference can lead to a degradation
of receiver positioning performance on the E5
and E6 bands in the GATE area.
It is obvious that the type and power of interference reported
here is not limited to the GATE area. Similar or
worse levels and characterisctics are to expected anywhere
in the world, on ground level or in the air (minding
mobility of the radar systems). Military radars being licensed
by ITU, the threat can not be expected to go away
anytime soon. GNSS receivers using the affected bands
E5b and E6 need to be robust against strong pulsed interference.
Safety-of-Life receivers should be performance
tested under stress of defined interference signal forms
and power levels.
ACKNOWLEDGEMENTS
GATE was developed on behalf of the DLR (German
Aerospace Center, Bonn-Oberkassel) under contract
number FKZ 50 NA 0604 with funding by the BMWi
(German Federal Ministry of Economics and Technology).
This support is greatly acknowledged.