Identifying Impersonation Attack in VANET using KNN and SVM Approach

Abstract

Vehicular ad hoc networks (VANETs) is useful communication in the vehicular systems using wireless technology. High mobility is one of important characteristic of VANETs. A message from one node to another node in VANET is transmitted with help of CAN (Controller Area Network) bus.  VANET is prone to various security attacks. Due to lack of mechanism to verify original source and destination of the message an attacker can easily inject malicious messages in the system. This attack is called as Impersonation attack. The proposed research work uses and compares KNN and SVM machine learning algorithms to overcome Impersonation attack in VANET. The research work uses data set provided by Hacking and Countermeasure Research Lab (HCRL) for experimental evaluation. Experimental results show that KNN gives 98% accuracy for detection of Impersonation attack which is high as compared to SVM approach wherein it is 93%.