Trajectory-Driven Deep Learning for UAV Location Integrity Checks

Abstract

While unmanned aerial vehicles (UAVs) are increasingly utilized in many domains, there is a growing concern about location integrity for securely deploying and managing the vehicles. A body of studies tackled this problem, e.g., using hardware sensors, cryptographic mechanisms, and machine learning (ML) approaches, but they concentrate primarily on GPS signal-related information (e.g., jamming and noise). In this study, we take a different approach that performs the checks by analyzing actual movement information represented with a sequence of flight records. This sequence-based approach keeps track of location updates across the flight path (‘trajectory’) rather than relying on point-wise signal-specific features to test the validity of the location information based on a single observation. Specifically, we define a set of attributes effectively capturing the movement of aerial vehicles over time, resulting in eight features with no signal-dependent information. We then present our deep sequence method, implemented on top of either a recurrent neural network (RNN) or a Transformer with a backend classifier, performing integrity checks with the newly defined feature set. Our extensive experimental results support the feasibility of our trajectory-based analysis approach, showing up to 98.9% classification performance with negligible false positive rates (lower than 1%) for ensuring location consistency (even without referring to the GPS signal-specific information).

Keywords

• Unmanned aerial vehicles
• location integrity
• location spoofing
• trajectory
• feature extraction
• sequence-based analysis