Performance Analysis and Comparison of Anomaly-based Intrusion Detection in Vehicular Ad Hoc Networks

E. A. Shams, A. H. Ulusoy, A. Rizaner

Performance Analysis and Comparison of Anomaly-based Intrusion Detection in Vehicular Ad Hoc Networks

Číslo: 4/2020
Periodikum: Radioengineering Journal
DOI: 10.13164/re.2020.0664

Klíčová slova: Vehicular ad hoc networks, support vector machines, denial of service attack, intrusion detection, machine learning

Pro získání musíte mít účet v Citace PRO.

Přečíst po přihlášení

Anotace: Security and safety applications of Vehicular Ad hoc Networks (VANETs) are developed to improve the traffic flow. While safety applications in VANETs provide warnings and information for the vehicle and other units in the area, malicious behaviors can render this very purpose meaningless. Intrusion Detection Systems (IDSs) are key features for identifying the presence of faulty or malicious behaviors. Support Vector Machine (SVM) is an efficient tool for anomaly detection and it can be employed for intrusion detection based on the metrics of a known attack or normal behavior. Dropping and or delaying network packets are two of the most common variants among other methods in Denial of Service (DoS) attacks. Hence an IDS which can detect both variants can detect similar types of DoS attacks. The result of the study is obtained by designing and implementing an SVM detection module into computer-generated simulation, which depicts a successful outcome in detection of mentioned DoS attack variants.