Machine Learning-based RF Jamming Detection in Wireless Networks

Abstract

Due to the open and shared nature of wireless medium, wireless networks are vulnerable to Radio Frequency (RF) jamming attacks since an attacker can easily emit an interference signal to prevent legitimate access to the medium or disrupt the reception of signal. An attacker may utilize different jamming strategies by exploiting the vulnerabilities of wireless protocols at different layers. Therefore, the detection and classification of jamming attacks is important to take proper countermeasures. In this paper, we firstly discuss some classic jamming strategies and some performance metrics for jamming detection. We then implement the jamming attack modules based on the NS-3 simulator, and study the effects of different types of jamming strategies. We propose some jamming detection schemes based on a variety of machine learning algorithms. The effectiveness of the proposed jamming detection schemes are valuated and optimized based on the collected data for different jamming attacks.