WEKA-based Real-Time Attack Detection for VANET Simulations

The number of connected vehicles has significantly increased in recent years and Vehicular Ad Hoc Networks (VANETs) are one of the most important technologies developed for these Intelligent Transportation Systems (ITS). In this infrastructure, vehicles continuously broadcast data to other vehicles and road side units (RSU) which leads to complex scenarios of connected vehicles. Moreover, due to their unique nature and characteristics, such as their high mobility, VANETs are highly vulnerable to various internal and external attacks. Our aim is to implement security measures, which includes the deployment of misbehavior detection frameworks to effectively mitigate these attacks. The Framework for Misbehavior Detection (F2MD) is one of the proposed solutions developed in this context. In this paper, we present an enhanced version of F2MD that utilizes the Waikato Environment for Knowledge Analysis (WEKA) for real-time detection of attacks in VANETs. Hence, we can leverage the wide range of algorithms provided by WEKA to perform real-time prediction and evaluation tests, conduct algorithm comparisons, and visualize the results. We validate our solution by comparing a set of selected algorithms available in WEKA. The Support Vector Machine (SVM) has been identified as the optimal choice in terms of prediction’s speed and accuracy with a value approaching to 99%.