A Lightweight AI-Based Approach for Drone Jamming Detection

The future integration of drones in 6G networks will significantly enhance their capabilities, enabling a wide range of new applications based on autonomous operation. However, drone networks are particularly vulnerable to jamming attacks, a type of availability attack that can disrupt network operation and hinder drone functionality. In this paper, we propose a low complexity unsupervised machine learning approach for the detection of constant and periodic jamming attacks, using the Isolation Forest algorithm. We detail the tuning of the base model as well as the integration with a Majority Rule module which significantly reduced the number of false positives caused by environmental noise, achieving high accuracy and precision. Our approach outperforms the standard Isolation Forest model in the detection of both constant and periodic jamming attacks, while still correctly identifying nominal traffic. Finally, we discuss the potential integration of the proposed solution in 6G-enabled drone networks, as a lightweight edge-based solution for enhancing security against jamming attacks.