Detection of GNSS spoofing using the Fourier transform of correlator outputs and mitigation using an autoencoder neural network

Abstract

Present study focuses on defense against carry-off-type spoofing attacks which often cause distortion in the correlation function profile. Investigation of the frequency characteristics of the correlation function is proposed to detect the presence of the spoofing signal. Having detected the spoofing signal, it is suggested to use an autoencoder neural network to deal with the impacts of the spoofing. The autoencoder neural network removes distortions caused by the spoofing signal from the correlation function. Results demonstrate that the proposed detection method achieves a higher than 98% detection rate and autoencoder-based approach mitigates spoofing attacks by an average of 92.64%.