Collaborative Intrusion Detection System for Network and Flight Security in Unmanned Aerial Vehicles Group

Unmanned Aerial Vehicles (UAVs) have been gaining popularity in various areas, such as military, civil and commercial. However, these vehicles are exposed to cyber threats that can compromise their security and privacy and even result in physical damage. Such threats include signal interception, unauthorized access, data theft, and even remote control of the UAV. Therefore, UAV manufacturers and users need to be aware of these threats and adopt appropriate measures to strengthen the security and integrity of the systems. One of the defense strategies is the implementation of an Intrusion Detection System (IDS), which monitors the system for suspicious behavior. When an abnormality is identified, the IDS sends a notification to the control station, allowing appropriate decision-making. IDS aimed at UAVs often focus on detecting attacks on specific data sources, without considering the application in a group scenario. In this context, this paper presents a collaborative intrusion detection system for group security of UAVs. The system is capable of identifying threats both on the network and in-flight, using supervised and unsupervised learning. Attacks detected on the network include blackhole, grayhole, and flooding, while in-flight threats include GPS spoofing and jamming, with tests carried out using real UAVs. Federated learning is incorporated into the system to preserve data privacy and promote collaboration in training between UAVs. In addition, geographic and physical characteristics are con-sidered to ensure that the IDS operates independently of the specific hardware of the UAVs. The development also focuses on implementing a lightweight IDS, ensuring efficiency and optimized operation.