Multiparameter Joint GNSS Spoofing Detection Based on TSVAE

Spoofing attacks pose a significant threat to the global navigation satellite system (GNSS). However, existing techniques encounter challenges in accurately identifying spoofing signals, particularly of unknown attacks in complex and changing scenarios. In this article, we propose a multiparameter GNSS joint spoofing detection method based on temporal-spatio variational auto-encoders (TSVAE). The proposed model has demonstrated high precision in detecting unknown spoofing interference by utilizing a clean signal as a training dataset. Additionally, we have refined the conventional signal quality monitoring (SQM) metrics. The improved metrics display more distinct distributional characteristics and are better suited for adapting to unsupervised learning algorithms. We further test the proposed algorithm under multipath and electromagnetic interference to evaluate its anti-interference capability. The test results on the TEXBAT and OAKBAT datasets demonstrate that the proposed algorithm exhibits exceptional detection performance and robustness.