Online Learning Approach for Jammer Detection in UAV Swarms Using Multi-Armed Bandits

Abstract

Integrating Unmanned Aerial Vehicles (UAVs) into the 5G cellular network and O- RAN holds great potential for the UAV and communications industries. However, UAV wireless communication systems are susceptible to malicious attempts at jamming. This paper focuses explicitly on countering jamming signals that occur in a single direction, targeting UAV systems' physical layer security (PL). To address these security concerns, we utilize Online Learning (OL) methods to enhance the security of the physical layer in UAV systems. Our proposed approach involves an intrusion detection system (IDS) based on OL continuously updating its knowledge and responding to emerging real-time attack strategies to protect UAV communication networks. The primary objective of this method is to ensure the integrity, availability, and reliability of wireless Cyber-Physical Systems (CPS) operations while ensuring the safe and efficient functioning of multi-UAV swarms within the O-RAN framework. We present the performance of the OL-based IDS, supported by a mathematical analysis that demonstrates the effectiveness of the adapted solution to the problem. Moreover, while recognizing this field's inherent challenges and complexities, this research explores potential avenues for future investigation in enhancing physical layer security in UAV systems.