Jamming attack detection in a pair of RF communicating vehicles using unsupervised machine learning

Abstract

Wireless radio frequency (RF) jamming, both intentional and unintentional, poses a serious threat for wireless networks and wireless communications in general. Vehicular ad-hoc networks (VANET) are a subset of the wireless networks that incorporate modern safety-critical applications, that are vulnerable to jamming attacks. To preserve the secure communication and to increase its robustness against that type of attacks, an accurate detection scheme must be adopted. In this paper we present a jamming detection approach for wireless vehicular networks that leverages the use of unsupervised machine learning. The proposed method, utilizes a new metric, that is the variations of the relative speed between the jammer and the receiver, along with parameters that can be obtained from the on-board wireless communication devices at the receiver vehicle. Through unsupervised learning with clustering, we are able to differentiate intentional from unintentional jamming as well as identify the unique characteristics of each jamming attack. The proposed method is applied to three different real-life scenarios with extensive simulations being presented. © 2018 Elsevier Inc.