A GNSS Spoofing Detection Method Based on CNN-DOA

GNSS is an essential source of information for daily life, providing positioning and timing data. However, due to the low power of satellite navigation information at the receiving end, the open signal structure, as well as with the development of spoofing technology, the problem of spoofing and jamming at the receiving end has become increasingly severe. The GNSS spoofing detection technology based on DOA offers robust detection performance, which can adapt to various scenarios. To address the issues of low angular resolution and poor detection performance in a low SNR environment, we propose a CNN-DOA-based spoofing detection method. Firstly, we use the Toeplitz matrix reconstruction algorithm to estimate the DOA of the coherent signal. Then, we use the DOA spectrum generated by random angle distribution and random SNR as samples, with the DOA value and authenticity of the signal used as labels for CNN network training. When the power ratio of the spoofing signal to the authentic signal is greater than 1 dB (SNR = 0 dB), the detection accuracy of the spoofing signal is nearly 90%. Compared to the PI algorithm and the ADBF algorithm, the proposed model has higher resolution and robustness.