심층강화학습(DQN) 기반 주파수 호핑 드론 스웜 대응을 위한 적응형 정밀 재밍 전략

한민석¹ ; 황재룡²^{, *}

1해군사관학교 전자제어공학 부교수
2해군 중령/해군사관학교 사이버과학과 교수

Adaptive Surgical Jamming Strategy against Frequency Hopping Drone Swarms Using Deep Q-Networks

Min-Seok Han¹ ; Jaeryong Hwang²^{, *}

1Associate professor, Dept. of Electronics and Control Engineering, Republic of Korea Naval Academy
2LCR, ROK Navy/Professor, Dept. of Cyber Science, Republic of Korea Naval Academy

Correspondence to: *Jaeryong Hwang Dept. of Cyber Science, Republic of Korea Naval Academy 1 Jungwon-ro, Jinhae-gu, Changwon-si, Gyungsangnam-do, 51704, Republic of Korea Tel: +82-55-907-5350 E-mail: jhwang@navy.ac.kr

초록

본 논문은 주파수 호핑(frequency hopping)을 통해 생존성을 높인 드론 스웜(drone swarm) 위협에 대응하기 위해, 심층 강화학습(DQN) 기반의 적응형 정밀 재밍(adaptive surgical jamming) 전략을 제안한다. 2.4 GHz 대역에서 운용되는 50대의 드론 환경을 모델링하고, 에이전트가 재밍 주파수와 전력뿐만 아니라 재밍 모드(정밀/광역)를 상황에 맞춰 동적으로 결정하도록 학습시켰다. 시뮬레이션 결과, 제안된 기법은 목표 주파수를 국소적으로 타격하는 정밀성을 바탕으로 1.8 J/dB의 우수한 에너지 효율성을 달성하였다. 이는 기존의 광대역 재밍(barrage jamming, 2.9 J/dB) 및 Q-learning(2.3 J/dB) 방식 대비 전력 소모를 획기적으로 줄이면서도 드론 통신망을 효과적으로 무력화할 수 있음을 입증한다.

Abstract

This paper proposes an Adaptive Surgical Jamming Strategy based on Deep Q-Networks (DQN) to effectively counter drone swarm threats that employ frequency hopping techniques for survivability. We modeled a dynamic RF environment featuring 50 drones operating within the 2.4 GHz - 2.5 GHz band , where the intelligent agent learns to optimize jamming frequency, power levels, and modes (surgical vs. barrage). Simulation results demonstrate that the proposed DQN-based method achieves superior energy efficiency of 1.8 J/dB by precisely engaging targets. This performance significantly outperforms conventional Barrage Jamming (2.9 J/dB) and standard Q-learning (2.3 J/dB) approaches, proving the strategy's capability to neutralize hostile communication networks while minimizing power consumption.

Keywords:

Deep Reinforcement Learning, Drone Swarm, Frequency Hopping, Surgical Jamming, Energy Efficiency

키워드:

심층 강화학습, 드론 스웜, 주파수 호핑, 정밀 재밍, 에너지 효율성

Acknowledgments

이 논문은 2025년도 해군사관학교 해양연구소 지원을 받아 수행된 논문임.

References

J. Peng, Z. Zhang, Q. Wu, and B. Zhang, “Anti-jamming Communications in UAV Swarms: A Reinforcement Learning Approach,” IEEE Access, Vol. 7, pp. 180532–180543, 2019. [https://doi.org/10.1109/ACCESS.2019.2958328]
Y. Li et al., “On the Performance of Deep Reinforcement Learning-based Anti-jamming Method Confronting Intelligent Jammer,” Appl. Sci., Vol. 9, No. 7, 1361, Mar. 2019. [https://doi.org/10.3390/app9071361]
R. Xue and M. Zhao, “Cognitive-based High Robustness Frequency Hopping Strategy for UAV Swarms in Complex Electromagnetic Environment,” Wireless Commun. Mobile Comput., Vol. 2022, 4139345, Jul. 2022. [https://doi.org/10.1155/2022/4139345]
A. Pourranjbar, G. Kaddoum, A. Ferdowsi, and W. Saad, “Reinforcement Learning for Deceiving Reactive Jammers in Wireless Networks,” IEEE Trans. Commun., Vol. 69, No. 6, pp. 3682–3697, Jun. 2021. [https://doi.org/10.1109/TCOMM.2021.3062854]
Z. Xing, Y. Qin, C. Du, W. Wang, and Z. Zhang, “Deep Reinforcement Learning-driven Jamming-enhanced Secure Unmanned Aerial Vehicle Communications,” Sensors, Vol. 24, No. 22, 7328, Nov. 2024. [https://doi.org/10.3390/s24227328]