Security for Software Defined Vehicular Networks

Abstract

Vehicular Ad-hoc Network (VANET) comprises of moving and stationary vehicles connected by a wireless network. VANET plays an important role in safe driving, emergency and navigation. It is a part of the Intelligent Transport System. There are two main communications components in VANET which are Vehicle-to-Vehicle communication (V2V) and Vehicle-to-Infrastructure (V2I) communication. Vehicle-to-Vehicle communication (V2V) is communication between the vehicles in the network and Vehicle-to-Infrastructure communication (V2I) is communication between vehicles and the roadside framework. The main components of the VANET include Roadside Unit (RSU), On-Board Unit (OBU) and a Trusted Authority (TA). The RSU is the fixed unit that sends and receives information from the Trusted Authority (TA) and On-Board Unit (OBU). Therefore, it acts as the communication interface between the vehicles and the Trusted Authority. VANET include nodes that are highly mobile and they have dynamic topology. The devices share sensitive information between them. VANETs are vulnerable to different types of security attacks. Hence secure communication must be enabled. The security is provided by a message encryption mechanism and an Intrusion Detection System (IDS). Messages are encrypted and decrypted using ECC. An Intrusion Detection System is proposed in this work that uses Support Vector Machine (SVM). The classifier is trained using the NSL-KDD dataset. This IDS can be placed in the controller connected with the RSUs, where necessary actions can be taken based if attack a are detected. The performance of the classifier is evaluated for different splits of the dataset and a comparative analysis is drawn from all the splits.