Radio Frequency Fingerprint Identification of Drones Based on Variational Mode Decomposition

Recent technological advances have made drones one of the most popular unmanned vehicles for commercial and noncommercial uses. However, there are challenging concerns about their malicious uses. This issue shows the necessity of using a drone detection and identification system with high accuracy and range. In this research, we introduce a drone detection and identification system using fingerprint Identification of radio frequency (RF) signals of drones. Firstly, the drone signal is detected using an energy detector. Then, the RF signal variational mode decomposition (VMD) is calculated. In the following, highorder statistical features are extracted from VMD band-limited modes. The classification process has been applied using directed acyclic graph SVM (DAGSVM). According to the results, the proposed method can detect and identify different drone types with 93% and 86% accuracy for standard and practical datasets, respectively.