Multi-Receiver Ensemble Machine-Learning Techniques for GNSS Spoofing Detection with Generalisability Approach

Abstract

Global Navigation Satellite System (GNSS) spoofing incidents are on the rise. Accurate spoofing detection methods are crucial for maintaining the system integrity, especially since spoofing events may occur unexpectedly and may impact different receivers in varying ways. In this study, we present a novel approach to GNSS spoofing detection utilizing ensemble-learning techniques with data obtained from multiple receivers at Jammertest event from Norway, in 2023. Leveraging the Jammertest dataset, which has been underutilized in non-time series analyses, we conducted several experiments to extract the relevant features and develop an effective spoofing detection model using seven receivers. Our results demonstrate a promising accuracy, with performance reaching approximately 98% detection accuracy.