GNSS Multiparameter Spoofing Detection Method Based on Support Vector Machine

The vulnerability of the global navigation satellite system (GNSS) to spoofing limits its vast application in military safety and the national economy. Therefore, it is significantly important to rapidly and accurately detect GNSS spoofing. The traditional spoofing detection mainly detects spoofing signal intrusion by individual parameters such as signal power and signal quality. The single-parameter detection method can no longer manage complex and changing scenarios with the continuous development of spoofing technology. This paper proposes a GNSS multi-parameter joint detection method based on the support vector machine (SVM). The effects of spoofing on the signal acquisition, signal tracking, and position, velocity and time (PVT) solution processes are analyzed, and the composite signal quality monitoring (SQM), carrier-to-noise ratio C/N
₀
, pseudo-range Doppler consistency, PVT solving residuals, and clock difference and clock drift are extracted as detection features for training and testing of the support vector machine binary classification model. The F1-score of all nine-types parameter detection methods on TEXBAT and OAKBAT datasets improved by 22.95% and 19.6% (reach 93.97% and 97%), respectively, compared with only SQM parameters selected method. The obtained results demonstrate significant improvement in the spoofing detection performance compared with the traditional single-parameter methods.