Classification and Source Location Indication of Jamming Attacks Targeting UAVs via Multi-output Multiclass Machine Learning Modeling

Abstract

This paper introduces machine learning (ML) as a solution for the detection and range localization of jamming attacks targeting the global positioning system (GPS) technology, with applications to unmanned aerial vehicles (UAVs). Different multi-output multiclass ML models are trained with GPS-specific sample datasets obtained from exhaustive feature extraction and data collection routines that followed a set of realistic experimentations of attack scenarios. The resulting models enable the classification of four attack types (i.e., barrage, single-tone, successive-pulse, protocol-aware), the jamming direction, and the distance from the jamming source by yielding a detection rate (DR), misdetection rate (MDR), false alarm rate (FAR), and F-score (FS) of 98.9%, 1.39%, 0.28%, and 0.989, respectively.