Previous Abstract Return to Session B4 Next Abstract

Session B4: GNSS Resilience

Analysis of the Impact of a Non-Standard GPS C/A Code on Galileo Signals
Fabio Dovis, Politecnico di Torino, Italy; Davide Margaria, Beatrice Motella, Istituto Superiore Mario Boella, Italy
Location: Cypress

In this paper we analyze the impact of the transmission of Non-Standard Coarse Acquisition code broadcast by the Global Positioning System, focusing on the impact on the Galileo signals and receivers. The purpose of this document is to present a comprehensive analysis on the impact of the transmission of a Non-Standard C/A Code (NSC) by the NAVSTAR 63/SVN49 satellite of the GPS constellation started in May 2017.
Overview on Non-Standard C/A Codes and SVN49
Transmission of the Non-Standard C/A Code (NSC) for testing purposes is foreseen in the GPS Interface Control Document [1]. GPS satellites can switch off regular broadcast of C/A code and P(Y) code and transmit the NSC and Non-Standard Y Code (NSY). Such an operation is intended to protect users from receiving and utilizing erroneous satellite signals in case of “unhealthy” conditions on the spacecraft. Strictly speaking, this case cannot be formally considered as an "anomaly", because the transmission of non-standard codes is documented in [1]. Therefore, the transmission of NSC can be considered a normal operation even though it may reflect a glitch in the transmitting satellite.
In the past few years, the transmission of NSC by GPS satellites has been experienced several times. It is of particular interest the case in which the NSC has not the structure of a usual PRN C/A code (Gold code) but it takes the shape of pure square wave. The last documented transmission of this NSC has been reported in May 2017 for SVN49 [2].
The GPS spacecraft SVN49, also known as Block IIR-20(M), carries the demonstration payload for the new civil GPS L5 signal. Its launch on March 24, 2009, enabled the United States to meet the International Telecommunications Union (ITU) deadline for securing primary rights to use of the RF band by GPS. Very soon [3] issues related to the on-board signal generation were detected [4] [5]. SVN49, resumed transmissions as PRN24 on August 9, 2012. The signals were marked unhealthy and the satellite was not included in broadcast almanacs with plans to remain out of service until an L1/L2 satellite multipath issue was resolved. Since that day the satellite was being used for tests, it was set unhealthy, and it had not been included in broadcast almanacs [6].
According to the GPS Notice Advisory to Navstar Users (NANU) 2017001 [7], SVN49 resumed to broadcast using the PRN04 since “approximately 05 JAN 2017”. The satellite should have stayed unhealthy and no broadcast almanac would include SVN49/PRN04 until further notice. NANU 2017042 announced that PRN04 was allocated to SVN38 starting from “approximately 18 May 2017” [7].
As it was reported in [2], the date of the switch previously recalled for PRN04, actually matches the dates when the spikes in the L1 spectrum appeared. This is an evidence that the SVN49 started that day to broadcast a NSC.
The total power of the signal is concentrated on few spectral components, and it can be seen as a series of Continuous Wave (CW)-like in-band signals. The main harmonic components are at 511.5 kHz (i.e. 1.023/2 MHz) with respect to L1, and only harmonics spaced of 1.023 MHz are present. This aspect suggests that a periodic signal is being broadcast. In fact, the spectrum of this NSC matches the spectrum of a periodic signal with a period equal to two times the inverse of the C/A code chipping rate (i.e. Tp = 2*Tc = 2/Rc, with Rc = 1.023 Mchip/s). The theoretical power spectral density is then made of a series of harmonics components modulated by the spectral shape of the rectangular chip pulse.
In addition, as it happened in the past [8] [9], the spreading sequence of the NSC corresponds to a Binary Phase Shift Keying (BPSK)-modulated sequence with alternating logic 0s and 1s, transmitted at Rc = 1.023 Mchip/s.
It has also been verified that on the NSC a standard navigation message is modulated, with the usual 50 bit/s data rate, and a structure compliant to [1]. The presence of these navigation data modulates the CW-like spectral lines.
In terms of transmitted power, it has been assumed that the total power allocated to the NSC, is the same as any other PRN code. This assumption has been empirically confirmed, comparing the power spectral density of a simulated NSC with the live received signal.
A single CW is often considered one of the most harmful kinds of interference to GNSS. . CW interference is also known to trick the PLL tracking loop into locking in on the interfering signal instead of the GPS ranging signals, even after spreading by the correlation process. Intricate jamming techniques may even use two CW signals, separated by the known intermediate frequency of the target receiver. .
The impact of CW interference depends, of course, on the power of the interfering signals, but also on which lines of the GNSS spectrum are affected by it [10][11]. In addition, the actual hit of a spectral line depends on the Doppler shift experienced by the satellite signal.
The threat posed by conventional CW interference is real. Specific anti-jamming techniques, such as notch filtering, have been designed and implemented in commercial receivers, to mitigate the effects of this kind of interference [12].
Being the NSC a CW-like interference, it is then of interest to see if the NSC is actually causing a performance loss in the correlation stages of a receiver for GPS L1 C/A and for Galileo E1 OS signals (whose spectral lines are spaced of 1 and 0.25 kHz, respectively, due to the different periodicity of the spreading codes [1] [13]).
Methodology and Anticipated Results on the Interference Impact
The maximum amount of inter-satellite interference is given by the overlap of the spectrum of the NSC and the spectrum of another GNSS signal (GPS or Galileo), which share the same carrier. However, in order to consider the mutual interference between the signals, we have to consider the correlation operation that is performed by the tracking stage of a GNSS receiver.
In detail, we have to take into account that such an operation is performed along an integration time that is in most of the cases shorter or equal to the Bit duration (Tb). Then the integrate and dump operation over the integration time (Tint) further low-pass filter the CW-like spikes in the spectral representation.
Thus the cross correlation between the two signals is given by the integral of the overlap between the two spectra which have a Sinc(f) shape but different width of the main lobe (as well as different height of the side-lobes).
It is clear then, that as far as the processing of the signals in the receiver is concerned, the correlator shapes the frequency spectrum of the interference. Only high sensitivity receivers (for which Tint > Tb) get closer to the ideal case of pure spectral lines, then having zero mutual interference unless a CW-like spike is hitting one of the spectral lines of the GNSS signal.
At a first glance when the Doppler is zero, the spectral lines of the NSC fall exactly in between two spectral components of GPS spectrum. However, it has been verified by means of simulations that their cross-correlation is not null: the shaping due to the finite integration time in the tracking stage of a GNSS receiver makes the two spectra to overlap. In fact, considering the local code as periodic and shaped in frequency by the limited integration time and the spectrum we are actually considering signals that are not disjoint in the frequency domain.
Since the spectra are overlapping, the relative phase condition has to be taken into account and it will influence the amount of cross-interference between the signal in the receiver correlator.
In the same way they are overlapping also with the Galileo spectral components.
In order to assess the actual interference between NSC and GPS and Galileo signals the Interference Error Envelope (IEE) introduced in [14] has been used. The value of the envelope is a worst case assessment of the bias induced at the tracking stage by the presence of a CW interference.
In [14] it is demonstrated that the impact of CW at receiver level depends also on the configuration of the receiver itself, and it is in particular dependent on:
• Discriminator spacing
• Modulation format
• Discriminator type
• Filter bandwidth
The bias depends on the carrier shift of the CW with respect to the GNSS carrier and on the relative phase, but in order to use the tool for conservative analyses, the maximum bias is considered.
In order to assess the impact of an NSC interference on a BPSK(1) and on BOC(1,1) signals in ideal conditions (i.e. without noise, filtering, and quantization effects), a pure square wave has been simulated.
The attention has been focused on a realistic simulation of the NSC interference and then on the assessment of its impact on GNSS signals and receivers.
It must be noticed that, assuming a worst case combination of Doppler frequency shifts on both SIS and interference signals, a Doppler Frequency Range equal to [-10÷10] kHz has been used. This means that the Doppler effect has been applied to the NSC signal shifting its harmonic components with a magnitude up to 10 kHz and modifying the square wave frequency accordingly (i.e. simulating the Doppler effect both on the carrier frequency and on the code rate). On the other hand, the useful SIS under test has been kept with an ideal zero Doppler shift.
The full paper will report the results obtained by means of simulations and related to the impact of a NSC on Galileo E1-B, E1-C, and GPS L1 C/A signals and considering all the Code Numbers described in their respective ICDs [1][13].
It can be anticipated that a remarkable difference in terms of IEE performance has been noticed, clearly showing that the Galileo E1 OS signals seems more affected by the NCS transmission than the GPS L1 C/A (i.e. worst case errors around 0.8 m for Galileo versus about 0.013 m for GPS, using Tint = 4 ms in both the cases).
The larger impact on the Galileo signals, empirically demonstrated by means of simulations, can be explained as basically due to the BOC structure of the Galileo OS signal, that turn out to be more sensitive to the NSC within the stages of the receiver signal processing.
A comparison of the impact of the NSC with respect to the impact of a pure CW signal is also provided, showing how when a pure CW is present, Galileo is less sensitive to its presence with respect to the GPS. This result is mainly due to the distribution of the power over a larger number of spectral components, in case of Galileo. This effect does not happen in the case of the NSC, where due to the presence in the spectrum of several Sinc(f) components causes, as a final result, a larger bias.
Additional results will be presented in the full paper in order to complement the investigation, by analyzing the impact of the correlator spacing and of the front-end bandwidth on the performance.
References
[1] Global Positioning System Directorate, System Engineering and Integration, Interface Specification, NAVSTAR GPS Space Segment / Navigation User Segment Interfaces, IS-GPS-200 Rev. H, IRN003, July 28, 2016. Available at: http://www.gps.gov/technical/icwg/IRN-IS-200H-001+002+003_rollup.pdf
[2] F. Dovis et al. “Anomalous GPS Signals from SVN49”, GPS World July 2017. Available at: http://gpsworld.com/anomalous-gps-signals-reported-from-svn49/
[3] E. Gakstatter. “ESRI Conference and SVN-49 Troubles”, GPS World July 2009.
[4] T. Springer, and F. Dilssner. "SVN49 and other GPS anomalies." Inside GNSS2009 (2009): 32-36.
[5] R. B. Langley. “Expert Advice: Cause Identified for Pseudorange Error from New GPS Satellite SVN-49” GPS World (2009): 8-12.
[6] GPS World staff. “SNV49 Off the Air?”, GPS World March 2013.
[7] GPS navigation Center, https://www.navcen.uscg.gov/?pageName=gpsAlmanacs
[8] Z. Zhu, S. Gunawardena, M. Uijt de Haag, F. van Graas and M. Braasch, “GNSS Watch Dog: A GPS Anomalous Event Monitor”, Inside GNSS, Vol. 3, No. 7, Fall 2008, pp. 18–28.
[9] Z. Zhu, S. Gunawardena, M. Uijt de Haag and F. van Graas, “Satellite Anomaly and Interference Detection Using the GPS Anomalous Event Monitor”, in Proc. of the 63rd Annual Meeting of The Institute of Navigation, Cambridge, Massachusetts, April 23–25, 2007, pp. 389–396.
[10] Ouzeau, C., Macabiau, C., Roturier, B., and Mabilleau, M. “Performance of multicorrelators GNSS interference detection algorithms for Civil Aviation”  ION NTM 2008, National Technical Meeting of The Institute of Navigation.
[11] Motella B., Savasta S., Margaria D., Dovis F. “A method to assess robustness of GPS C/A code in presence of CW interferences”, International Journal Of Navigation And Observation, vol. 2010, pp. 1-8. ISSN 1687-5990.
[12] Fabio Dovis (editor), GNSS Interference Threats and Countermeasures, Artech House (2015).
[13] European GNSS (Galileo) Open Service. Signal-In-Space Interface Control Document. OS SIS ICD, Issue 1.3, Dec. 2016.
[14] B. Motella, S. Savasta, D. Margaria and F. Dovis, "Method for Assessing the Interference Impact on GNSS Receivers," in IEEE Transactions on Aerospace and Electronic Systems, vol. 47, no. 2, pp. 1416-1432, April 2011.



Previous Abstract Return to Session B4 Next Abstract