Previous Abstract Return to Session F6 Next Abstract

Session F6: GNSS Signal Processing in Degraded Environments 2

STFT-based Method Applied to GNSS Spoofing Estimation, Mitigation and Discrimination from Multipath
Hanchuan Jiang, Chao Sun, Hongbo Zhao, Wenquan Feng, Beihang University, China
Location: Monroe
Alternate Number 1

Spoofing attacks greatly undermine the reliability of the global navigation satellite system (GNSS), especially for safety critical applications. As the low power level and opening signal structure, GNSS open service signals are very vulnerable to the spoofing attacks. Thus, the demand for proper anti-spoofing techniques becomes a must in current and future GNSS receivers for robust, accurate and reliable positioning.
Typical anti-spoofing schemes can be categorized into two categories: spoofing detection and spoofing mitigation. Through spoofing detection process, the receiver recognizes spoofing but does not deal with it. However, in order to ensure the acquisition and tracking of the navigation signal, spoofing mitigation process is necessary. Prior to the spoofing mitigation, we have to estimate the spoofing parameters, and distinguish the spoofing signal from its authentic counterpart, as well as the possible environment effect, such as multipath. However, the current spoofing parameter estimation methods, such as the maximum likelihood approach or the adaptive filtering method, either need a priori assumptions about the number of spoofing signals or require high computational load, which limits the applicability of such methods.
Motivated by developing a more feasible spoofing estimation method, we propose a new spoofing parameter estimation and mitigation technique using the short-time Fourier transform (STFT). This STFT-based method has been successfully applied to the area of signal distortion detection and multipath detection by C. Sun et al.. As structures of the spoofed and multipath signals are very similar, it also has the potential to be an effective anti-spoofing method. The proposed method can be summarized into two stages: cross power spectral density (CPSD) calculation and spoofing parameter estimation.
The received signal is transferred to the frequency domain to obtain CPSD and separate out the channel transfer function for spoofing parameter estimation. During this process, we take the operations of segmentation and averaging to reduce both noise effect and computational load. Then, we take the inverse short-time Fourier transform (ISTFT) of CTF to obtain the channel impulse response function in the time domain. By searching for and analyzing the time offset and amplitude of each impulse, the spoofing parameters can be estimated. After that, the autocorrelation function of spoofing signal is subtracted in the tracking loop and the effect of the spoofing attack is effectively removed.
In addition, when considering multipath environments, this paper introduces a way to classify the authentic, spoofing and multipath components in the composite signal by analyzing the characteristics of the estimated parameters. The distinction between spoofing and multipath can base on but not limited to the following principles:
1) The reflected ray lags behind the line of sight (LOS), but spoofing signal can be ahead of the authentic signal
2) Multipath amplitude is correlated with the delay, so the rays with longer delay usually have weaker power
3) Multipath tends to vary fast and randomly but spoofing tends to vary slowly and regularly
The performance of the proposed STFT-based method was evaluated both theoretically and employing real GPS signals. The spoofing dataset used here is the so-called Texas Spoofing Test Battery (TEXBAT), which is a test battery of real cases publicly provided by the University of Texas at Austin. One multipath environment scenario recorded as a drive in the city Osaka, Japan was used. A NordNav receiver software was used to track and post-process the GPS signals. Preliminary experimental results indicate that the proposed STFT anti-spoofing method is capable of estimating the spoofing parameters and then removing the spoofing effect from the tracking loop. Even in degraded environments, the proposed method is able to identify a potential spoofing attack and minimize the false alarm rate caused by multipath.



Previous Abstract Return to Session F6 Next Abstract