DoS Attacks Detection in the Network of Drones: An Efficient Decision Tree-Based Model

Abstract

This study examines the detection of the denial of service (DoS) attacks on Wi-Fi-based unmanned aerial vehicles (UAV). The paper proposed an efficient DoS attack detection method based on Decision Tree classifier. The method consists of preprocessing, feature extraction, and DoS attack detection. The preprocessing was proved to save drones’ resources and improve the detection rate. The investigation of different classifiers, i.e., KNN, Random Forest, Logistic Regression, and Decision Tree, the latter was concluded to be the best in detecting DoS attacks of types of De-authentication and UDP/TCP flood within the shortest runtime. The evaluation further showed that proposed DoS detection method is better than the most related work where it achieved detection with F1-score of 0.989 and with the shortest latency.