Real-time detection of GPS jamming and spoofing attacks on unmanned aerial vehicles using machine learning
Computers and Industrial Engineering, cilt.219, 2026 (SCI-Expanded, Scopus)
- Yayın Türü: Makale / Tam Makale
- Cilt numarası: 219
- Basım Tarihi: 2026
- Doi Numarası: 10.1016/j.cie.2026.112207
- Dergi Adı: Computers and Industrial Engineering
- Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus, ABI/INFORM, Aerospace Database, Applied Science & Technology Source, Compendex, INSPEC, DIALNET, Business Source Ultimate (EBSCO), Engineering Source (EBSCO), Technology Collection (ProQuest)
- Anahtar Kelimeler: Cybersecurity, Edge computing, Flight control systems, GPS jamming, GPS spoofing, Machine learning, Real-time detection, Software-defined radio, Unmanned aerial vehicles, XGBoost
- İstanbul Gelişim Üniversitesi Adresli: Evet
Özet
Unmanned Aerial Vehicles (UAVs) are extensively deployed across diverse application domains, including reconnaissance, surveillance, logistics operations, and agricultural monitoring. However, their high dependency on Global Navigation Satellite System (GNSS)-based navigation systems exposes them to significant cybersecurity vulnerabilities. This study presents the development of an artificial intelligence-based detection model capable of identifying GPS jamming and spoofing attacks within a controlled experimental environment. Attack simulations were successfully conducted using the HackRF One Software-Defined Radio (SDR) device, while GNSS data were acquired through a NEO-M8N GPS module integrated with a Pixhawk flight control unit. The data collection process was executed through custom scripts on an Ubuntu-based system, and the acquired datasets were analyzed using seven distinct machine learning algorithms to compare their performance metrics. Experimental results demonstrate that the XGBoost algorithm achieved superior performance with 96.4% accuracy, 96.4% recall, 96.5% precision, and an F1-score of 96.4%. The model's training time of approximately 1.03 s and test inference time of 0.01 s indicate suitability for real-time applications. The developed AI-based system was successfully integrated with the flight control unit via a Raspberry Pi single-board computer and validated through real flight scenarios, enabling instantaneous detection and response during attack events. The proposed approach offers a cost-effective, reliable, and field-deployable solution aimed at enhancing the cybersecurity of both civilian and military UAV operations. This research contributes to the growing body of knowledge on securing autonomous systems against emerging electromagnetic threats and provides a practical framework for real-time GNSS anomaly detection.