Global positioning system spoofing detection based on Support Vector Machines

Abstract

The civil Global Positioning System (GPS) is vulnerable to spoofing because of its open signal structure. The performance of previous spoofing detection methods is often limited due to spoofing's strong concealment. In this study, a method is proposed to detect spoofing by analysing the features of improved signal quality monitoring (SQM) moving variance (MV), improved SQM moving average (MA), early-late phase, carrier-tonoise ratio-MV and clock offset rate of receiver using Support Vector Machines. Then, the effectiveness of different kernel functions is compared along with other previous methods, revealing that our method outperforms previous methods when coarse Gaussian is used as kernel function. Specifically, the f1 score of the proposed method is improved by 3.22%, 12.85% and 35.72% in comparison with Back Propagation network, Ratio and Delta. The authors hope this work is beneficial for future research and for the implementation of GPS spoofing detection technology and high-performance receiver, which is of great significance to maintain the normal operation of GPS.

Enhanced Machine Learning Ensemble Approach for Securing Small Unmanned Aerial Vehicles From GPS Spoofing Attacks

“…Machine learning methods offer promise for GPS spoofing detection in small civilian UAVs by eliminating the need for extra hardware. In [31], a support vector machine (SVM)-based approach was proposed to identify UAV GPS spoofing attacks through state estimation analysis. Alternatively, [32] introduced a method relying on received signal strength measurements to establish a credible residence area, effectively discerning between authentic and spoofed GPS positions.…”