CYBERATTACKS ON ON-BOARD SYSTEMS OF UNMANNED AERIAL VEHICLES AND METHODS OF COUNTERING THEM BASED ON MACHINE LEARNING
DOI:
https://doi.org/10.35546/kntu2078-4481.2026.1.33Keywords:
unmanned aerial vehicles, UAV cybersecurity, GPS spoofing, information-extreme machine learning, threat detection, navigational securityAbstract
The article examines modern cyber threats to unmanned aerial vehicles and methods for their detection in the context of increasing UAV autonomy. Traditional approaches to detecting GPS spoofing, based on deep learning methods, show limited effectiveness in dynamic operating conditions. Such approaches depend on the stability of the environment, the availability of communication infrastructure and are characterized by high computational complexity. The integration of information-extreme machine learning allows for the detection of navigation signal substitution using compact invariant descriptors, ensuring resistance to geometric transformations and the ability to work with limited amounts of training data. A comprehensive systematization of cyber threats to autonomous UAVs is considered, including electronic warfare, physical and hardware threats, threats to multi-drone systems and attacks on machine learning models. A comparative analysis of existing GPS spoofing detection methods was conducted, which showed that approaches based on LSTM, MLP, CNN demonstrate accuracy from 78 % to 95 %, but their effectiveness decreases sharply with dynamic changes in the environment. In addition to technical aspects, the study also considers the evolution of threats, in particular: the emergence of more advanced spoofing methods, the use of directional jamming antennas to disrupt operation at high altitudes, the vulnerability of inertial navigation systems to target destabilization due to microvibrations and acoustic effects, as well as the specificity of attacks on drone swarms with asymmetric capabilities. The importance of developing integrated protective methods that can increase the reliability of navigation systems in complex geospatial conditions is emphasized. It is studied that adversarial attacks are a two-sided threat: they can deceive target recognition systems, forcing UAVs not to detect enemy vehicles or incorrectly classify obstacles. The aim of the study is to identify both the potential and limitations of existing methods for detecting cyber threats to UAVs, as well as to substantiate the prospects for the application of information-extreme machine learning, taking into account technical limitations and operational requirements. The authors note that the most effective approach is a hybrid protection model that combines IEMN with the use of radial-angular descriptors, the use of sensors and adaptive filtering. This provides higher reliability, accuracy of detecting flight path deviations and reducing errors in modern conditions of electronic warfare.
References
Impact Assessment of GNSS Spoofing Attacks on INS/GNSS Integrated Navigation System / L. Yang et al. Sensors. 2018. Vol. 18, no. 5. P. 1433. URL: https://doi.org/10.3390/s18051433
An S., Jang D. J., Lee E.-K. Adversarial Evasion Attacks on SVM-Based GPS Spoofing Detection Systems. Sensors. 2025. Vol. 25, no. 19. URL: https://doi.org/10.3390/s25196062
Black-box Adversarial Attacks on CNN-based SLAM Algorithms / M. R. Gkeka et al. arXiv preprint arXiv. 2025. URL: https://arxiv.org/abs/2505.24654
An Experimental Study of Trojan Vulnerabilities in UAV Autonomous Landing / R. Ahmari et al. ArXiv. 2025. URL: https://arxiv.org/abs/2510.20932
Behind The Wings: The Case of Reverse Engineering and Drone Hijacking in DJI Enhanced Wi-Fi Protocol / D. Pratama et al. P. 2023. URL: https://arxiv.org/abs/2309.05913
Protecting Autonomous UAVs from GPS Spoofing and Jamming: A Comparative Analysis of Detection and Mitigation Techniques / P. C. Joeaneke et al. Journal of Engineering Research and Reports. 2024. Vol. 26, no. 10. P. 71–92. URL: https://doi.org/10.9734/jerr/2024/v26i101291
Threats from and Countermeasures for Unmanned Aerial and Underwater Vehicles / W. Khawaja et al. Sensors. 2022. Vol. 22, no. 10. P. 3896. URL: https://doi.org/10.3390/s22103896
On the requirements for successful GPS spoofing attacks. ACM Transactions on Information and System Security / N. O. Tippenhauer et al. CCS ‘11: Proceedings of the 18th ACM conference on Computer and communications security. 2011. Vol. 18. P. 75–86. URL: https://doi.org/10.1145/2046707.2046719
Renu Y., Sarveshwaran V. A Review of Cyber Security Challenges and Solutions in Unmanned Aerial Vehicles (UAVs). Inteligencia Artificial. 2025. Vol. 28, no. 75. P. 199–219. URL: https://doi.org/10.4114/intartif.vol28iss75pp199-219
A Comprehensive Survey of Unmanned Aerial Systems’ Risks and Mitigation Strategies / S. Shrestha et al. 2025. ArXiv. URL: https://doi.org/10.48550/arXiv.2506.10327.
Patel A., Cherukuri A. K. Analysis of Light-Weight Cryptography Algorithms for UAV-Networks. ArXiv. 2025. URL: https://arxiv.org/abs/2504.04063
Skorobahatko M., Voitsekhovskyi A. Lightweight Cryptography in UAV systems. Theoretical and Applied Cybersecurity. 2025. Vol. 7, no. 1. URL: https://doi.org/10.20535/tacs.2664-29132025.1.326898
MARS: Defending Unmanned Aerial Vehicles From Attacks on Inertial Sensors with Model-based Anomaly Detection and Recovery / H. Meng et al. arXiv. 2025. URL: https://arxiv.org/abs/2505.00924
Data collection using unmanned aerial vehicles for Internet of Things platforms / S. Goudarzi et al. Computers & Electrical Engineering. 2019. Vol. 75. P. 1–15. URL: https://doi.org/10.1016/j.compeleceng.2019.01.028
Goodfellow I. J., Shlens J., Szegedy C. Explaining and Harnessing Adversarial Examples. arXiv. 2015. URL: https://arxiv.org/abs/1412.6572
Athalye A., Carlini N., Wagner D. Obfuscated gradients give a false sense of security: Circumventing defenses to adversarial examples. Proceedings of the 35th International Conference on Machine Learning. 2018. URL: https://arxiv.org/abs/1802.00420
Physical Adversarial Examples for Object Detectors / K. Eykholt et al. Proceedings of the 12th USENIX Workshop on Offensive Technologies. 2018. URL: https://doi.org/10.48550/arXiv.1807.07769
Sarıkaya B. S., Bahtiyar Ş. A survey on security of UAV and deep reinforcement learning. Ad Hoc Networks. 2024. P. 103642. URL: https://doi.org/10.1016/j.adhoc.2024.103642
Couturier A., Akhloufi M. A. A Review on Deep Learning for UAV Absolute Visual Localization. Drones. 2024. Vol. 8, no. 11. P. 622. URL: https://doi.org/10.3390/drones8110622
Naumenko I., Myronenko M., Savchenko T. Information-extreme machine training of on-board recognition system with optimization of RGB-component digital images. RADIOELECTRONIC AND COMPUTER SYSTEMS. 2021. No. 4. P. 59–70. URL: https://doi.org/10.32620/reks.2021.4.05
Information-extreme machine learning of a cyber attack detection system / A. Dovbysh et al. RADIOELECTRONIC AND COMPUTER SYSTEMS. 2022. No. 3. P. 121–131. URL: https://doi.org/10.32620/reks.2022.3.09
