Interference Detection and Analysis in the Field of GNSS Verification

Steffen Thölert, Stefan Erker, Johann Furthner, Michael Meurer

German Aerospace Center (DLR)

Institute of Communications and Navigation BIOGRAPHIES Steffen Thölert received his diploma degree in Electrical Engineering with fields of expertise in high-frequency engineering and communications at the University of Magdeburg in 2002. The next four years he worked on the development of passive radar systems at the Microwaves and Radar Institute at the German Aerospace Centre (DLR). In 2006 he changed to the Department of Navigation at German Aerospace Centre (DLR), Institute of Communications and Navigation. Now he is working within the topics of calibration, civil security and automation of technical processes. Stefan Erker received his diploma degree in Communication Technology at the Technical University of Kaiserslautern, Germany in 2007. In the same year he joined the Institute of Communications and Navigation of the German Aerospace Center (DLR) at Oberpfaffenhofen. He mainly works on the topics of GNSS verification and corresponding signal analysis Johann Furthner received his diploma degree in Physics with fields of expertise of laser physics at the University of Regensburg in 1990. In 1994 he finalized his Ph.D. work in laser physics at the University of Regensburg. Since 1995 he is scientific staff at the Institute of High Frequency, since 2000 of the Institute of Communication and Navigation, both at German Aerospace Centre (DLR). In 2008 he stayed half year at ESA/ESTEC in the Galileo Project Team as Navigation Performance Engineer. Johann Furthner is working since 1995 on the development of navigation systems in a number of areas (systems simulation, timing aspects, GNSS analysis, signal verification, calibration processes). Michael Meurer received the diploma in Electrical Engineering and the Ph.D. degree from the University of Kaiserslautern, Germany. After graduation, he joined the Research Group for Radio Communications at the Technical University of Kaiserslautern, Germany, as a senior key researcher, where he was involved in various international and national projects in the field of communications and navigation both as project coordinator and as technical contributor. Since 2005 he has been an Associate Professor (PD) at the same university. Additionally, Dr. Meurer joined the German

Aerospace Centre (DLR), Institute of Communications and Navigation in 2006. Since June 2008 he is the director of the Department of Navigation. ABSTRACT A constant growing number of applications based on satellite navigation systems especially safety critical ones like airplane landing require highest precision and robustness of the positioning solution also available under rough conditions. In order to provide sufficient quality of service the positioning performance of the satellite navigation system is a main key factor. Accurate and comprehensive interference measurements of unwanted and possible perturbing emissions and a clear separation of the different sources are necessary since they can degrade the performance and integrity of safety critical applications. A main challenge in these kinds of measurements is the determination of the interference source. These could be external terrestrial interference like ground based navaids as the Distance Measuring Equipment (DME) or Tactical Air Navigation (TACAN), TV-Stations and radio operators which affect the signal quality due to their high output power levels and continuous operation. Another source of potential interference is caused by other satellites which transmit their signals in adjacent or even the same frequency bands. Beside the determination of the interference source also the information of the specific form and occupied bandwidth of this measured interference is essential. Starting in 2005 the DLR Institute of Communications and Navigation established an independent monitoring station for the analysis of GNSS signals. A 30 meter antenna located at Weilheim, Germany is the core element of this facility. An absolute calibrated setup with this 30m antenna allows very precise measurements on a single navigation satellite. Using this facility it is also possible to measure and analyse potential interference very detailed. This paper gives an overview of the used setup and will present the results of an extensive interference measurement campaign in the L-Band, the most interesting frequency band for satellite navigation at the moment.

INTRODUCTION High quality assessment of signals in space of navigation satellites means that the signal from the satellite is measured very accurate and is free from external influences and disturbances to determine exactly possible differences in comparison to its specifications. For that purpose it is necessary on one hand to calibrate the system very precise. And on the other hand to measure and analysis all possible interferer which can have an impact to the signal monitoring facility and consequently the raw measurement data. Interference in general can occur in the measurement system itself or coupled in from outside of the antenna. The interferer can be distinguished by several types of characteristics. One of the main interesting criteria within the interaction of unwanted signals and signals they should transmit by the satellite is the bandwidth. We can distinguish into narrow-band and wide-band interferences. A popular example of narrowband signal sources are AM and FM radio stations. Also to the group of this interferer belongs tone interference. They can be considered as narrowband interference with the smallest bandwidth. It concentrates all of its power at one frequency and can be modelled as a sine wave. Wideband sources are e.g. radars, TV stations and communication systems. Further signal properties of interest of the interference are signal power, modulation, transmit duration and repetition time (if signal is pulsed). The location of the interferer is also of great interest. With the known of the position there can be tried to avoid measurements in this special direction to reduce interference signals in the raw data. In general it can be separated in ground-based and air-based interferences. Why it is so important to analyse and determine all these interferences in that detail? The separation of the different interference sources is the main challenge to detect all unwanted signals transmitted by the satellite under test, which is a strong criteria to declare the satellite as usable. Furthermore interferences can overlay and disturb parts of the satellite signal so that for instance an analysis of the spectral purity or the modulation quality is not possible. With the exact known of the interference properties the special measurements can skipped out of the signal quality analysis or maybe the interference can be filtered out. Interferences can not only overlay the signals from the satellite of interest. Strong external signals can causes saturation effects in the amplifiers of the measurement system or cause an over range of the vector signal analyzer. With the precise known of interferences negative effects onto the measurements can be avoid by filters or pre-attenuation before the LNA’s or the signal analyzer. Note that the use of additionally attenuation raises the noise level and consequently decrease the C/N0.

MEASUREMENT FACILITY AND SETUP Starting in September 2005 the Institute of Communications and Navigation of the German Aerospace Center (DLR) established an independent monitoring station for the analysis of GNSS signals. The core element of this facility is a 30 meter deep space antenna located at DLR ground station near Weilheim, Germany and was originally built up in the early seventies for the first US/German interplanetary satellite mission HELIOS-A/B. Launched in 1974 and 1976 the two probes approached the sun closer than any space probe before. Later the antenna was used to support other scientific space missions and various experiments. For the new challenge the antenna has been adapted to the requirements in the navigation field. A newly developed broadband circular polarized feed and a new receiving chain including an online calibration system were installed at the antenna during the preparation for the GIOVE-B IOT campaign in spring 2008. In this time also intensive work on the system calibration was performed using well known signals from radio stars and EGNOS satellites for antenna gain determination and suitable calibration methods for the receiving system. This calibration leads to a remaining cumulated absolute measurement uncertainty significantly less than 1dB. The use of this antenna which is characterized by its high gain and small beam width and the absolute calibration of the whole measurement setup including also the 30m antenna allow very precise and absolute measurements on a single navigation satellite.

Figure 1: The 30m High Gain Antenna at DLR ground station Weilheim

The antenna is based on a shaped Cassegrain system with elevation over azimuth mount. It is characterized by a gain value of over 50dB in the L-band and a beam width around 0.5°. The absolute position accuracy of this antenna is 0.001° in each direction. The signals are directed from the parabolic main reflector to a hyperbolic sub-reflector with 4 meters in diameter. This sub reflector sends the signals via a waveguide and a second sub-reflector in a measurement cabin. One big benefit of this construction is the direct access to the installed feed in the cabin and the possibility to place the complete

measurement equipment next to the feed and avoid the use of long connection cables.

The measurement system which is adapted to the 30 meter antenna include two Low Noise Amplifiers (LNA) with a total gain of almost 70dB. Several directional couplers are used for the injection of pilot signals. Thus it is possible to calibrate the receiving system during operation near real-time. The setup also has several signal outputs for external equipment like bit grabbers or navigation receivers. Another main item of the receiving system are the band pass filters dedicated to the individual Galileo navigation bands and an all-pass to measure out-of-band emissions. The signals are recorded using a vector signal analyzer with at least 80 MHz bandwidth. For signal recording of wideband signals the analyzer is used for down conversion of the signal. A digital oscilloscope connected to the analyzer is able to record with 300 MHz signal bandwidth.

Figure 2: Measurement Setup Overview Figure 2 shows the complete measurement setup including parts used for system calibration and a caesium clock used as frequency standard for the measurement equipment. More detailed information about the 30m antenna and the used measurement setup can be found in [1]. SYSTEM CALIBRATION As the objective of the measurement setup in Weilheim is the “Signal in Space” (SIS) performance characterisation of satellites like GIOVE-B, all relevant distortion contributions of the measurement setup need to be characterized accurately and finally removed within the parameter evaluation. To achieve a combined absolute measurement uncertainty significantly less than 1.0 dB it is essential to calibrate every part of the used system very precisely. This includes beside all RF components of the receiving system also the high gain antenna itself. For the characterisation of the high gain antenna two values are assessed. The first one is the antenna pointing accuracy. This error contributes with a reduced maximum power value caused by the miss pointing of the antenna pattern. So the pointing offset is measured for azimuth and elevation using the geostationary satellite Artemis.

The measurements show a systematic elevation offset of 0.02° and an azimuth offset of 0.03°. This offset is corrected in the antenna control. The second value is the antenna gain. For accurate measurement of the antenna gain natural sources like radio stars or artificial sources like geostationary satellites are well suited. For a characterization over the complete used frequency range the radio star Cassiopeia A is used. Cassiopeia A is one of the strongest wideband radio emitters on the northern hemisphere. The star is circumpolar and therefore usable for calibrations at every time of the year. With the help of the well-known flux density of the celestial radio source Cass A the G/T can be determined, which is the relation between the gain of the antenna and the noise temperature of the receiving system. After a precise determination of this noise temperature the antenna gain can be calculated. More information about the antenna calibrations can be found at [1, 2]. For the receiving system calibration several techniques are used. The State-of-the-Art method is the use of a network analyzer. This analyzer can be calibrated remotely and connected to the receiving system by a group of dedicated switches. These precise measurements of gain and phase are performed periodical. A frequency and power stabilized signal generator is used in combination with two power meters for an online gain determination during measurements. This method is used for detection of gain variations of the low noise amplifiers and passive elements of the receiving system. It is also possible to detect if one of the amplifiers is saturated and working outside specified limits. INTERFERENCE PROPERTIES For the analysis of interferences mostly the same parameters of the signal are of interest as for the SIS verification process:

- Power - Band/Frequency location - Bandwidth - Modulation - Transmit duration (CW or Pulse) - Repetition time - Spatial location,…

The location of the interferer could be on ground or in the air, stationary/fixed or moving, for example a TV-Station is ground based and stationary and a satellite is air based and moving (LEO/MEO) or more or less fixed (GEO). Furthermore it is of interest if the interference affects directly the GNSS band or the out of band region and where is the origin of the interference; the satellite itself or external sources. Out-Of-Band Emissions Out-of-Band Emissions describe any type of unwanted and unspecified emission which is caused by the observed satellite outside the specified frequency band limits. These emissions could lead to Inter-System-Interference.

Isolator

out

in

30 dB

IF OUT

Oscilloscope

Splitter III

out

External Signalor

50 Load

Power Meter I

Power Meter II

out

out

BaySEFout

Signalgenerator

30 dB

Networkanalyzer

eCalKit

LHCP

in

Splitter I

Coupler

Weilheim – Measurement Setup

Coupler

Cs-Clock

10M

Hz

Dis

trib

utio

nA

mpl

ifier

10 dB

LNA

LNA

Spectrumanalyzer

out

inSplitter II

Isolator

Feed

Cou

ple

r

Isolator Isolator

RHCP

In-Band Emissions In-Band Emissions names all unwanted and unspecified radiation inside the systems own specified frequency bands. They can cause Intra-System Interference as well as Inter-System Interference.

Intra-System Interference Intra-System Interference can be caused by interference which is radiated by a single navigation satellite inside the own used signal bands and could therefore result in performance degradation of the satellite or other satellites of the navigation system which of course also operate in the same frequency band (In-Band Emissions). Inter-System Interference This type describes the impact of interference on other (GNSS or L Band) satellite systems. This can be caused either by Out-of-Band Emissions - if the interfered systems frequency band is adjacent to the system which is generating the interference - or also by In-Band Emissions – if the interfered system uses the same or a part of the interfering satellite frequency band (e.g. GPS / Galileo L1) External Interference This type of interference characterises nearly any other interference that is not radiated by the satellite itself. There are various sources. The strongest are mainly ground based interferer such as navaids like TARCAN and DME, TV and Mobile Stations or Radars. But also emissions of stars like the sun or a strong radio star like Cassiopeia can cause observable influence on the measurements due to the use of a high gain antenna. INTERFERENCE MAESUREMENTS Spectral interference scans In a first step all potential and present interference sources have to be determined. For that purpose the spectrum from 1020 MHz up to 1820 MHz is measured 360° around the signal monitoring facility in steps of 0.5°. Such interference scans consist of a set of 10 measurements with a bandwidth of 80 MHz each and a resolution of 10 KHz. Each data acquisition collects the signals for about 10 seconds. So we can ensure that most of all interferences in the observed bandwidth are captured within the record. Figure 3 shows an example of an interference scan with 800 MHz bandwidth. In the plot the Galileo bands are marked and it could be seen that in L1 and E6 no significantly interference was recorded. But in E5 we can see a non GNSS signal. These signals (*1) in and below the E5 band are mostly radiated by radars (airport radars, DME/TACAN). The frequencies around the L1 band (*2,*3) are known to be used from global satellite communication systems like Inmarsat, Thuraya or Iridium. The signals (*4) over 1750 MHz are transmitted by mobile phone communication systems.

*1*2 *3 *4

E5a+E5b E6 L1

Figure 3: Spectral interference scan 1020 MHz – 1820 MHz To get an overview about possible interference not only from one direction/snapshot the whole azimuth range of 360° was measured. Figure 4 shows the result of a scan around the 30m antenna in Weilheim for 10° elevation and the frequency range from 1020 MHz to 1820 MHz. 10° is in general the lowest elevation which is used for SIS verification in order to guarantee high quality analyses. Thus these measurements can be seen as a worst case (concerning ground based interference). What we can see from this plot are the frequently used channels of mobile communication between 1760 MHz and 1820 MHz. The direction of most signals is correlated with the direction to the city of Weilheim, which is located around 8 kilometers south-east direction from the ground station. Furthermore the radars/DMEs and the satellite communication signals could be seen well. In the right part of Figure 4 the frequency ranges around the Galileo bands are blanked out to assess the interference impact for these special bands. The L1 and L6 band is nearly free from external interferences and only the E5 band contains several significant interferers. To separate ground based interferer from air based the measurements are performed for different elevations. Ground based signals should be decreased in signal strength with higher elevations and air based not consequently. Figure 5 shows the 360° scan for different elevations. Obvious can be seen the decrease of the mobile communication signals around 1800 MHz and also the decrease in the interferers between 1020 MHz and 1200 MHz (mostly DMEs). The frequency area just below the L1 band is filled with a lot of signals. The maximum is at an elevation angle of around 30° in southern azimuth direction. The explanation is that these signals mainly are coming from satellite communication systems in geostationary orbits. Most of these satellites that transmitting signals over Europe are in elevation angles between 25°-35° seen from the ground station location.

Sp

ectr

alF

lux

Den

sit

y[d

BW

/m2/

Hz]

NN

L1

E5 E6

1020 MHz 1820 MHz Figure 4: Spectral interference scan (360° azimuth, 10° elevation) around the verification facility in Weilheim

1020 MHz 1820 MHz

N

Sp

ectr

alF

lux

Den

sity

[dB

W/m

2 /H

z]

1020 MHz 1820 MHz

NN

Sp

ectr

alF

lux

Den

sity

[dB

W/m

2 /H

z]

Figure 5: Spectral interference scans at 10°, 20°, 30° and 40° elevation

Besides the external interferences it should also taken into account that the measurement system itself may produce unwanted signals which are not determined by the calibration process. Possible interference sources are CPU clock, front side bus or USB of the used control and data processing hardware. Usually such emissions are very weak and below the noise floor. In general they are shielded by the Faraday Effect of the metallic case. But in cases of manufacture errors at the case or damages stronger signals may couple into the signal path of the receiving system. Figure 6 shows a small interference originated from the USB ports (detected by switch on and off the port). To eliminate the emissions unused USB ports should be switched of or terminated.

Figure 6: Measurement system interference caused by USB ports of the PCs Detection of in-band unwanted signals The Antenna Offset Tracking uses the precise knowledge of the high gain antenna pattern, which was measured during several calibration campaigns for different frequency ranges. With a modified antenna tracking which is adapted to individual pass of the observed satellite a similar effect as described for the low elevation scan is achieved. The antenna tracking performs a set of well defined azimuth and elevation offsets to the original satellite track (see Figure 7 and Figure 8). These offsets result in a power reduction that directly depends on the selected offset value and the corresponding point of this offset at the antenna pattern. All interference signals which are radiated by the satellite vary with this pattern dependent power value. For this measurement we assume that the radiated interference shows a quite stable power level over time. Quite stable means that the fluctuation of the satellite observed interference is significantly smaller as the power variance caused by the antenna offset tracking. Interference which is radiated by another L-Band satellite that is accidentally passing the antenna beam should be only visible for a short time and shows a different behaviour in the offset tracking power variance. Possible ground based interference is strongly dependent on the elevation angle. The power level of this kind of interference is much lower at high elevation values. So this issue can be excluded from the analysis by reviewing for example the corresponding spectrogram plot for the complete measurement.

Figure 7 Example for Azimuth Offset Tracking Values

Figure 8 Example for Elevation Offset Tracking Values The GIOVE CW EIRP measurements which are performed during the In-Orbit-Test of the GIOVE satellites and on re-characterisation campaigns make the detection of In-Band Interference possible. For the L1 Offset Carrier CW two pilot signals located on the BOC(15,2.5) main lobe frequencies are transmitted by the satellite (see Figure 9).

Figure 9 GIOVE-B Offset Carrier CW Measurement 25.02.2009 An Offset-Tracking was performed during this satellite pass. Figure 10 shows the power behaviour of both pilot signals during the Offset-Tracking. It can be seen that the pilots change with the same ratio at the different offset values. Using this technique interfering signals from the satellite can be detected. They should show a similar behaviour with respect to their power level.

Figure 10 Power Variation of GIOVE Offset Carrier Pilots caused by Offset Tracking

Figure 11 Offset Tracking result overview Figure 11 presents the result of one Offset Tracking. Four different data sets with different offset values are compared. To possible In-Band interferer are marked in this overview spectrum. In Figure 12 the Left Pilot Peak is plotted for all offset values. The different power levels caused by the offset tracking are visible.

Figure 12 Offset Tracking Result for Left Pilot Peak

Figure 13 shows an detailed plot for the possible interference #1. In the Figure we see that the power level of this peak shows similar power behavior as the pilot peak at Figure 12. So this peak could be an interference caused by the satellite itself. The peak frequency of this peak is identically to the Galileo L1 center frequency. So this peak seems to be an unsuppressed residual of the L1 carrier. The second possible interference (see Figure 14) peak does not significantly change its power level during the offset tracking, so this interference seems to be not from GIOVE-B.

Figure 13 Possible interference #1 seems to be GIOVE-B L1 carrier residual

Figure 14 Possible interference #2 seems to be external interference

Measurement of interferences between satellites To investigate interferences between satellites a MATLAB tool was developed to calculate such satellite crossing events in advance. That means that during these events not only the observed satellite is located directly in the antenna beam also one or even more satellites are in the main beam or the side lobes of the antenna. For the 7th of September 2009 we identified a track with three satellite crossing events. The path of the observed satellite seen from the ground station is presented in the sky plot shown in Figure 15. In this graph we included a short fraction of the other satellite tracks which were crossing the path of GIOVE-B within the 0.5° beamwidth of the high gain antenna.

Figure 15: Sky plot for GIOVE-B Track 7.9.2009 (blue) including tracks of satellites which crossing GIOVE-B path For a basic determination of interference during a satellite observation over several hours the recorded data in spectral domain are analysed. The spectrum of the observed satellite, in this case GIOVE-B L1, is shown in Figure 16 and the corresponding IQ constellation diagram is shown in Figure 17.

Figure 16: GIOVE-B L1 Spectra

Figure 17: GIOVE-B L1 IQ Constellation The spectrum over the complete measurement time is displayed using a spectrogram (also known as “Waterfall Diagram”) in Figure 18 and as a three dimensional plot in Figure 19. With the help of this representation it is possible to assess the type of interference and its precise duration. This allows also a first separation between other moving satellites that cross the antenna beam during the measurement and external ground based sources.

Figure 18: 2D Spectrogram of GIOVE-B measurement 7.9.2009 During this satellite path of GIOVE-B the three predicted satellites crossed the GIOVE track. But also a third GNSS satellite is seen in the plot (at 1605 MHz).This second weaker interference of an additional GLONASS satellite was received via the side lobes of the high gain antenna.

Figure 19: 3D Spectrogram of GIOVE-B measurement 7.9.2009 Figure 20 shows a detailed spectrum of the interference event for GIOVE-B with the GLONASS satellite (Cosmos-2413). The GLONASS space vehicle was for the duration of 3.5 minutes within the antenna beam.

Figure 20: GIOVE-B L1 signal overlaid with GLONASS satellite

PDFPDF

Figure 21 Distorted IQ Constellation of GIOVE-B L1 signal due to GLONASS Interference Figure 21 shows the IQ constellation diagram, which is complete distorted and no modulation can be seen. The reasons are the partly overlay of the two signals and the two different Doppler shifts of the GIOVE and

GLONASS signal, because it is only possible to correct one of the Doppler shifts.

Figure 22 GIOVE-B L1 signal overlaid with GPS BIIRM L1 signal Figure 22 shows the spectrum of the inter-satellite interference event from GIOVE-B with the GPS NAVSTAR 59 satellite. Caused by the fact that the Galileo and the GPS system using the same center frequency for E1 respectively L1 the signals are complete overlaid in frequency range and both the spectra (Figure 22) and the IQ plot (see Figure 23) and time samples are unusable for verification purposes.

PDFPDF

Figure 23 Distorted IQ Constellation of GIOVE-B L1 signal due to GPS L1 Interference With the developed tool for the calculation of inter-satellite interference events it is also possible to calculate the amount of observation time for a specific satellite that is affected by satellite interference. The tracking of GIOVE-B for example over 10 days is within the beam of 0.5° is 0.1 percent of the time affected by (GPS, GLONASS, COMPASS and Inmarsat). Smaller antennas with a wider beam are more affected by other satellites. CONCLUSION This paper gives a short overview of the DLR GNSS monitoring facility and describes very briefly the used calibration strategies.

In the following different types of interferences are introduced which are important in the field of SIS verification. Based on measurement examples different strategies for interference detection are presented and described in detail. The paper shows that there are several possible/potential interference sources or events, like DMEs or interference between satellites, in the L-band and how to deal with them. The paper dealt mostly with the detection and analysis of interferences in the frequency domain. The next step is to characterize the interesting interference sources also in time domain and analyse their effect on precise and detailed verification measurements and analyses. ACKNOWLEDGMENTS The authors want to thank GSOC for using this great antenna and the colleagues at the DLR location near Weilheim for the operational and maintenance service. The investigations and developments presented in this paper have been partly supported within the scope of the research project 50-NA-0805 in contract of the DLR (German Aerospace Center, Bonn-Oberkassel). The authors gratefully appreciate the support and funding of this project by the BMWi (German Federal Ministry of Economics and Technology). REFERENCES [1] S. Thölert, S. Erker, M. Meurer, ‘GNSS Signal Verification with a High Gain Antenna – Calibration Strategies and High Quality Signal Assessment’, ION-ITM-2009, Anaheim, California [2] J. Banks, N. Roddis, „The Measurement of Moderate Size Reflector Antennas using Astronomical Calibrators“, Antennas and Propagation, IEEE 1975, No.407 [3] S. Erker, S. Thölert, M. Meurer, ‘Concept for Interference Detection and Analysis on Navigation Satellites’, ENC-2009, Naples, Italy