RF-Based UAV Detection and Identification Using Random Forest

As unmanned aerial vehicles (UAV s) increasingly permeate daily life due to technological advancements, they introduce significant challenges related to privacy and security. As a result, effectively detecting and identifying UAV s has become a significant research topic. This paper proposes a method for UAV RF signal detection and identification based on Principal Component Analysis (PCA) and Random Forest. Using the open DroneRF dataset, the RF signals of UAV s are first preprocessed with a Discrete Fourier Transform (DFT) to obtain the frequency spectrum data. PCA is then applied to extract features, producing a set of principal components from the spectrum, which are sub-sequently fed into the Random Forest classifier for classification. In experiments, the accuracy for UAV detection, classification, and operational mode recognition reached 99.71 %, 95.24%, and 79.19%, respectively, demonstrating better performance compared to previous studies.