Robust Spoofing Detection and Mitigation based on Direction of Arrival Estimation.

Kurzfassung

The threat of spoofing attacks is a serious problem for civil GNSS applications with safety content, such as airplane landing or ship navigation in a harbor. Also many strategically important infrastructures, such as electric power grids or mobile communications networks, are becoming increasingly dependent on GNSS services. In contrast to military GNSS users which solve the problem to a large extent by utilizing encrypted signals, civil GNSS receivers have to live today and most probably in the near and mid future with unencrypted signals of open GNSS services. Therefore, such receivers have to be protected by additional receiver-sided techniques, which are able to detect and mitigate spoofing attacks. Adequate solutions for the GNSS spoofing problem are the subject of intensive research. A number of receiver-autonomous spoofing detection techniques have been proposed. In order to detect the presence of a spoofing attack these techniques rely on the observation of the signal power, the Doppler frequency offset, the PRN code delay and its change rate, the correlation function shape as well as the cross-correlation of the signal components at different carrier frequencies. The advanced protection against even the most sophisticated spoofing attacks can be provided by the use of multiple antennas. This comes from the fact that the differential carrier phases of a signal, observed at different antennas, depend on the direction of arrival of the signal. Using this, a receiver with an antenna array is able to estimate the directions of arrival of the GNSS signals and detect the spoofing attack, if a large part of the signals come from a single direction. Moreover the malicious signals can be mitigated by generating a spatial zero in the array antenna reception pattern in the direction of the spoofing source. The use of the multi-antenna based approach for spoofing detection and mitigation was investigated by the authors in [1][2]. A technique for joint spoofing detection and antenna attitude estimation by using estimated signal directions of arrival was developed. This technique was implemented in an experimental receiver [3] with an adaptive antenna array where the direction-of-arrival (DOA) is estimated in each tracking channel at the post-correlation stage. On the one hand DOA information is used to constrain the digital beamforming process. On the other hand the proposed technique uses this information also for detection of spoofing attacks. The detection is based on testing the observed DOAs of the satellite signals against the predicted DOAs. The latter are obtained in the local east-north-up (ENU) coordinates while solving the PVT problem and using the computed satellite positions and user position solution. Because the attitude of the antenna array in the local ENU coordinates is not necessarily known, the spoofing detection is therefore treated as a joint detection (i.e. of spoofing attack) and estimation (i.e. of attitude) problem. It was practically demonstrated that the observed DOAs can be used to identify the direction to the spoofing source and produce a spatial null in the array reception pattern for mitigating this type of radio frequency interference. However the technique developed in [1][2] still suffers from the effects of short-term inaccuracies in the direction of arrival estimation occurring during the spoofing attack. On the one hand the algorithm itself can be improved by advanced techniques. On the other hand this problem can be effectively solved by using sequential estimation approach for the array attitude combined with an adequate user motion model. The results of practical tests reported in [2] also indicate that the DOA estimation performance under spoofing attack should be improved in order to maintain reliable spoofing detection and pointing of spatial null toward the spoofing source. This is especially important at the initial stage of a spoofing attack where the counterfeit signals just appear with relative weak power levels and therefore are difficult to detect. The paper presents a new sequential approach for solving the problem of DOA-based attitude estimation. The approach is described in details and its performance is analyzed with the help of computer simulations. The performance improvement for spoofing detection with respect to the former snapshot approach from [1] and the approach using a simple Kalman filter [2] are discussed. Since accurate DOA information is of large importance for the spoofing detection with the proposed approach, some effort is spent in the frame of this study on improving the direction of arrival estimation. The effect of mutual coupling of the array elements is accounted for by using two-stage DOA estimation and applying the corresponding corrections, especially for signals arriving at low elevation angles. It is shown that the corrections can be derived either from anechoic antenna measurements or through the array calibrations using live satellite signals. Both types of corrections are analyzed and compared with respect to their performance. Further, the subspace orthogonal projection method is applied to improve the DOA estimation in the situations where both authentic and counterfeit signals with the same spreading code are simultaneously observed at correlator outputs of a tracking channel. If both signals have significantly different powers, the DOA estimation with commonly used techniques such as MUSIC or ESPRIT often fails to resolve the weaker signal [2]. Therefore at the second estimation run the orthogonal projection is applied in order to minimize the effect of the stronger signal and allow for resolving the direction of arrival of the weaker signal. In order to obtain representative practical results the performance of the spoofing detection and mitigation with the sequential estimation approach is assessed by post-processing raw signal data collected during field trials in scenarios with a GPS repeater. The DLR’s experimental array receiver platform GALANT [3] with a 2-by-2 rectangular array is used for collecting the raw data. [1] M. Meurer, A. Konovaltsev, M. Cuntz, and C. Hättich, “Robust Joint Multi-Antenna Spoofing Detection and Attitude Estimation using Direction Assisted Multiple Hypotheses RAIM,” in Proc. ION GNSS 2012, 2012. [2] A. Konovaltsev, M. Cuntz, C. Haettich, and M. Meurer, “Autonomous Spoofing Detection and Mitigation in a GNSS Receiver with an Adaptive Antenna Array,” in Proc. ION GNSS+ 2013, 2013, p. 12. [3] M. Cuntz, A. Konovaltsev, M. Heckler, A. Hornbostel, L. Kurz, G. Kappen, and T. Noll, “Lessons Learnt: The Development of a Robust Multi-Antenna GNSS Receiver,” in ION GNSS 2010, 2010.