GPS SVN49 L1 Anomaly Analysis based on Measurements with a High Gain Antenna

Transcription

1 GPS SVN49 L1 Anomaly Analysis based on Measurements with a High Gain Antenna S. Thoelert, S. Erker, O. Montenbruck, A. Hauschild, M. Meurer German Aerospace Center (DLR) 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 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 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 Oliver Montenbruck is head of the GNSS Technology and Navigation Group at DLR s German Space Operations Center, where he started to work as a flight dynamics analyst in His current research activities comprise space-borne GNSS receiver technology, autonomous navigation systems, spacecraft formation flying and precise orbit determination. André Hauschild received his diploma degree in mechanical engineering in March 2007 from the Technical University of Braunschweig, Germany, and subsequently became a Ph.D. candidate at DLR s German Space Operations Center. His research focuses on realtime clock-offset estimation of GNSS satellites and the utilization of new GNSS signals. 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 Since June 2008 he is the director of the Department of Navigation. ABSTRACT The constant developments in the field of global satellite navigation systems opened the door for a new type of safety critical applications. These applications that are mostly related to aviation generate stringent requirements on the used navigation system. To meet these challenging requirements which include precision and integrity of the positioning solutions as well as robustness of the service new technologies have to be implemented in developing modernized satellite navigation systems. Only the usage of newly designed signals located on multiple frequencies in combination with advanced receiver architectures and algorithms will allow achieving these demands. In early spring this year GPS IIR-20(M) better known as SVN49 was launched which carries a L5 demo payload. This satellite got famous not only by the transmission of the first GPS L5 signals from a Medium Earth Orbit (MEO) in space shortly after its launch but rather by the detection of an anomaly on the L1 and L2 legacy navigation signal. This anomaly is caused by a L1/L2 signal reflection on the L5 payload via the auxiliary antenna port where the L5 payload is connected. Consequently the satellite transmitted not only the nominal L1, L2 and L5 signals but also an inherent multipath signal on L1 and L2 which has shown to be elevation dependant. The paper gives a brief overview of the SVN49 and its transmitted signals. In the following the paper presents the L1 and L2 anomaly in detail and shows the effect on a GNSS receiver. Then a method is introduced for determining the multi-path error on L1 signal from calibrated measurement data taken with the high gain antenna. In a simulation a multi-path delay, attenuation and phase shift is added to a reference signal of SNV49 to

2 create a distorted constellation diagram that matches the real SVN49 IQ constellation which was recorded using the high gain antenna. So it is possible to assess delay, power and phase of the multi-path. SVN49 Industry presentation the L1 multi-path delay is given with approximately 30ns. 162 In. Cable L5 Filter INTRODUCTION In the course of the GPS modernization the US decided in August 2002 in coordination with the International Telecommunication Union Radio communication Sector (ITU-R) the transmission of a new civil signal on a third frequency known as L5. This new signal will be part of the new IIRF modernized filing, which includes the military (M) code signal on L1 and L2 frequencies and a new civil signal on L2. The filing in 2002 requires transmission of the signals from a MEO at least in summer Beside the fact that BIIR20 was intended to save the frequency filings for the GPS L5 signal by the radiation of the first L5 signals in medium earth orbit, the main mission of this satellite is to encourage the existing constellation with an additional IIR-M payload. This main objective of BIIR-20 failed so far that the satellite could not be declared operational till this day. The reason of this issue is a signal anomaly of the legacy L1 and L2 signals. This anomaly was reported by several research groups worldwide shortly after turning on the L5 payload. Reviewing the measurement data recorded before commissioning the L5 payload this anomaly had to be existent from first signal transmission of L1 and L2 on. The signal anomaly is visible in various GPS receivers due to a systematic and elevation dependent pseudo range bias. The bias reaches a maximum level of around 1 meter near satellite zenith. Beside the receiver measurements we performed also several measurements on SVN49 legacy signals using the DLR GNSS signal monitoring and verification facility. By reviewing the L1 IQ constellation diagrams we discovered unusual distortions occurring at higher elevations. At the end of June GPS Wing released a presentation to the public which gives a first reasonable explanation of the causes [5]. Outer Ring Inner Ring Figure 1 Antenna Coupling Network [5] The antenna coupling network of the satellite is designed to radiate a uniform signal power density over the earth. Therefore the inner ring of the antenna pattern transmits most signal power with a broad beam. The outer ring has a narrow sharper pattern and is used to shape the broad beam of the inner ring. With opposite phase at high elevation the outer ring reduces the radiated power at high elevation. At low elevation the pattern is in-phase and boosts the power of the inner ring. Most of the delayed signal from J2 is coupled on the outer antenna ring and less of the signal on the inner ring (the inverse behaviour to the J1 path). After the signals are delayed and attenuated they are transmitted from J2 with a directional antenna pattern of the outer antenna ring (see Figure 2). Outer Ring THE SVN49 ANOMALY The origin of this anomaly is in strict sense related to the SVN49 L5 demo payload. This additional payload is connected to the satellite antenna using an auxiliary input port (J2) of the coupling network. Having a look on the coupling network helps to understand how this multi-path signal is generated at the satellite. The legacy navigation signals L1 and L2 are feed into the antenna using the J1 connector. Part of the signal power is injected also into the J2 input path. The signal travels around 4,11m (162 inches) and is then reflected by a mismatched L5 filter. So the L1 and L2 signal goes back the same way into the antenna coupling network. At this point the reflected signal is delayed by the time it needed to travel the 8,22m (twice the length of the used cable). In GPS Wing s Inner Ring Figure 2 The two different antenna rings [5] This explains the elevation dependency of the anomaly and why it is visible on high elevations. The power over elevation for the delayed L1 signal is shown at Figure 3.

3 multipath contribution shows up as a static elevation dependent error in the receiver measurements. With few exceptions, similar errors were observed in all receivers of the global IGS tracking network, however, the full error amplitude was only evident for sites close to the SVN49 groundtrack. Figure 3 Elevation dependent power variation of the L1 delayed signal component [5] This input port is present at all BIIRM satellites and on some satellites used for the connection of additional maybe also classified payload equipment. Investigations started after the detection of this anomaly showed that also other BIIRM satellites show similar multi-path issues but with much less power than on SVN49. The future GPS IIF satellites will have a new designed coupling network, so that a full integration of an operational L5 payload is possible without this multi-path issue. ANOMALY IMPACT ON SIGNAL TRACKING In parallel to the signal analysis conducted with DLR s 30m antenna, conventional GNSS tracking receivers have been used soon after the launch of SVN49 to monitor the new L5 signal. In the course of this activity systematic errors in the legacy signals became apparent that are today known to originate from a satellite internal multipath. For the purpose of illustration, Figure 4 shows the multipath of L1, L2 and L5 code measurements collected with a Javad Triumph tri-band multiconstellation receiver with standard correlator during a high elevation (80 ) pass over Oberpfaffenhofen. A ionosphere-corrected codeminus-carrier difference 2 i 2 k 2 fk 2f ρ i Φi ( Φi Φk ) bi Mi + ε i f (also known as multipath-combination) has been formed to identify multipath in the individual signals. Obviously both the C/A code measurements and the (semi-codeless) P(Y) measurements on L1 show a systematic code error varying between 0.5 m and +1 m with a pronounced bumb at peak elevations. On the other hand, no significant errors can be discerned in the L2C and L2 P(Y) signals. Since elevation at the receiving site is directly related to the boresight angle of the transmit antenna, the SVN49 Figure 4 Code multipath errors of SVN49 for L1 C/A, L1 P(Y), L2C, L2 P(Y) and L5 (from top down) as obtained with a Javad Triumph receiver using a standard correlator for an overhead pass. It is also important to note that most users of dualfrequency receivers would actually experience a much higher pseudorange error of about 4 m peak-to-peak amplitude since the ionosphere L1/L2 combination amplifies the L1 error by a factor of 2.5. In view of the communality of the error pattern for most receivers an elevation dependent correction function has been proposed in [6]. In parallel, efforts have been made by the operators of the GPS constellation to compensate the pseudorange error by introducing a 150 m phase center offset into the broadcast ephemerides. This method effectively reduces the single-point positioning error for dual-frequency receivers, but obviously these

4 represent a fairly limited user group. Standard singlefrequency receivers, in contrast, would even further degrade their navigation solution by using the modified broadcast ephemerides. Evidently, the best way of coping with the SVN49 induced signal errors consists in the use of specialized tracking techniques that are insensitive to multipath in the specific domain of range delays. Once the cause of the SVN49 signal anomaly had been traced back to a reflection in the transmission chain an immediate effort has been made by Javad GNSS to adapt their advanced multipath mitigation techniques for the Triumph receivers to the specific case of this satellite. now available with all Triumph receivers. Evidently, the multipath errors are largely, if not completely, removed on all signals including both low and high chipping rates. Thus, no need for empirical range corrections exists when using advanced correlators for SVN49 tracking and users can benefit from all GPS satellites in view despite the known signal anomaly. MEASUREMENTS WITH HIGH GAIN ANTENNA 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 of this facility is a 30 meter deep space antenna located at DLR ground station at Weilheim, Germany. Figure 6 Groundstation Weilheim, Germany The integrated measurement system fulfils highest quality standards to acquire high accuracy measurement raw data of GNSS signals to perform precise analyses [1]. After the commissioning phase of SVN49 several data takes were performed in order to provide a basis for the analysis of the aforementioned signal anomaly. Figure 7 The 30m dish at Weilheim Figure 5 Code multipath errors of SVN49 for L1 C/A, L1 P(Y), L2C, L2 P(Y) and L5 (from top down) as obtained with a Javad Triumph receiver using an optimised strobe correlator. The effectiveness of the employed technology is nicely evidenced by Fig. 2, which shows the SVN49 multipath errors obtained with an optimised strobe correlator that is This facility was used to track and analyze the first L5 MEO signals transmitted from SVN49 at beginning of April this year [2][3][4]. During these measurement campaigns we recorded also the L1 and L2 navigation signals of this satellite. After the SVN49 signal anomaly was noticed at first by our receiver measurements, we started a detailed investigation on this issue using the high gain antenna data.

5 Figure 8 Overview of the measurement system Having a look on the recorded SVN49 L1 time samples we can see an increasing distortion in the IQ constellation plot during the ascending of the satellite. The distortion has a maximum at the highest elevation and reduces again during descending of the satellite. See Figure 9, Figure 10 and Figure 11. Figure 11 SVN49 L1 IQ Constellation at 80 Elevation The IQ constellation diagram seems to be well suited to identify the parameters of this multi-path anomaly. With the IQ plot two different effects can be observed. The first effect is expansion of the constellation diagram and the second effect is an additional twist of the constellation. This was the first time we detected this specific behaviour on a GPS satellites IQ diagram. Figure 9 SVN49 L1 IQ Constellation at 40 Elevation In next step this behaviour was simulated to understand how these two effects can be explained in detail. For that purpose a time record of SVN49 with a negligible multipath component (at approximately 40 elevation angle) was used as a reference signal with a length of about 100 ms. To simulate the influence of the multi-path an attenuated, time delayed and phase shifted replica of the reference signal is generated and added to the reference signal. Using these simulations a tool was developed which produces IQ diagrams by an iterative variation of the three dimensional search space parameters - signal attenuation, delay and phase shift and compare these results due to a 2 dimensional correlation with the recorded IQ diagram for the maximum multi-path situation. With that method is possible to observe the influence of each parameter on the constellation diagram and therefore assess the parameters for the best match of the simulated IQ with the real-world SVN49 anomaly which is given by the maximum value of the normalized result of the two dimensional correlation. Figure 10 SVN49 L1 IQ Constellation at 60 Elevation The expansion of the IQ constellation is mainly caused by the additional multi-path signal power and delay. The twist of the IQ Constellation is originated from a phase shift of the multi-path component. In our first simulation these parameters were chosen: Delay from 25ns to 45ns, step-size 1ns Attenuation from 8dB to 18dB, step-size 1dB Phase-shift from 0 to 90, step-size 5

6 The result in Figure 12 for the processed simulation of the three parameters (multi-path delay of 38 ns, a multi-path attenuation of 15dB and a phase shift of 35 ) shows a very good match with the maximum distorted recorded IQ diagram (Figure 11). These values are for the simulation result of an SVN49 L1 IQ anomaly at 80 elevation angle. (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] M.Meurer, S.Erker, S.Thölert, J. Furthner, A.Hauschild, R.B.Langley, S.Carcanague, L5 Arrives, GPS World May 2009 pp [3] M.Meurer, S.Erker, S.Thölert, O.Montenbruck, A.Hauschild, R.B.Langley, GPS L5 First Light GPS World, June 2009 pp Figure 12 Simulation Result L1 IQ for a multipath delay of 38ns, a multi-path attenuation of 15dB and a phase-shift of 35. The use of the SVN49 reference signal for the simulation has one great advantage. The signal includes all other specific characteristics of the SVN49 that are not related to the multi-path anomaly, like satellite transmitter chain and satellite antenna. Only with this reference signal it is possible to create a nearly perfect match with the real SVN49 anomaly IQ. [4] S.Erker, S.Thölert, M.Meurer GPS L5 First Light GPS World, June 2009 pp [5] SVN49 Industry Presentation GPS Wing, 2009 [6] Springer T., Dilssner F.; SVN49 and Other GPS Anomalies ; Inside GNSS, July/Aug. 2009, p.32 CONCLUSION This paper gives an insight into the SVN49 anomaly. We showed an example for the impact of this effect on the signal tracking behavior of a GNSS receiver and a possible solution for this issue. In the next step we simulated the anomaly using a recorded reference signal of SVN49 by an iterative modification of the three different parameters: multi-path delay, multi-path attenuation and phase shift. We were able to create a simulated multi-path IQ constellation which shows a very good match to the measured signal. In the next step the iterative simulation will be optimized to increase the resolution of the used parameters and to get a better understanding of the anomaly ACKNOWLEDGMENTS The authors want to thank GSOC for using the 30m high gain 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