AI-based Approach for Radio Frequency Jamming Attack Detection in Unmanned Aerial Vehicles

Unmanned Aerial Vehicles (UAV) are highly versatile systems with applications expanding in various fields, for instance, Surveillance, Reconnaissance, Disaster Response, Agriculture, and many more. Although UAVs have many important advantages over conventional manned aircraft, such as cheaper operating costs and a lower risk for human pilots, their vulnerability to Radio Frequency (RF)-jamming poses a substantial risk to navigation and communication systems by disrupting signals. UAVs depend on remote control and autonomous operations. These attacks are concerning in many application areas. Communication and navigation play crucial roles in these autonomous systems. This paper aims to explore different deep-learning algorithms for the detection of RF-jamming attacks in UAVs. A comparison analysis is conducted to evaluate five distinct architectures using conventional evaluation metrics criteria comprising accuracy, precision, recall, confusion matrices, and the area under the ROC curve. These comparative analyses helped in selecting an accurate architecture for the definition, RNN architecture gave an impressive accuracy of 93% and demonstrated a superior performance than other architectures. This paper aims to strengthen RF-jamming detection systems in UAVs to enhance their safety and operational reliability in several scenarios.