Exploiting acquisition measurements and spatial processing for improved GNSS spoofing detection and classification with snapshot receivers

Spoofing attacks mislead global navigation satellite system (GNSS) receivers into reporting false position, velocity, and time (PVT) solutions. However, not all spoofing attacks are the same: they range from simple meaconers to advanced synchronized spoofers. Classifying the attack improves situational awareness and helps the receiver operator take appropriate counteractions. This paper presents spoofing detection and spoofing attack classification using lightweight spatial and acquisition metrics from a snapshot receiver. Further, appropriate machine learning (ML) techniques provide high performance and intuition. The results show that spatial information (i.e., multi-antenna comparisons) has superior spoofing detection, but incorporating the acquisition metrics (i.e., temporal information) facilitates spoofing attack type classification. Finally, only basic acquisition metrics are required for classification, demonstrating lightweight feature selection and training. This work shows that spoofing detection and attack type classification is possible and emphasizes that a low feature space is required.