Machine Learning-Based RF Jamming Classification Techniques in Wireless Ad Hoc Networks

Abstract

Wireless networking plays a vital role in achieving ubiquitous computing where network devices embedded in environments provide continuous connectivity and services, thus improving quality of life of humans. However, current wireless networks can be easily attacked by jamming technology since wireless links are opened and shared. An attacker exploits the limitations in wireless protocols at different layers and disrupts the existing wireless communication by decreasing the signal-to-noise ratio at receiver sides through the transmission of interfering wireless signals. Thus, the design of accurate jamming detection algorithms becomes important to react to ongoing jamming attacks. With the rise of safety-critical applications, jamming attack is likely to become a con-straining issue in the future. Hence, detection of jamming attack and its classification is of great importance in order to take suitable countermeasures. We therefore proposed jamming detection and classification techniques which are based on Machine Learning. We analysed previously used jamming strategies and evaluate them with various performance metrics for jamming detection. In order to collect data, we simulated jamming in Wi-Fi ad hoc network using the ns-3 jamming model. The effectiveness of the proposed model is then evaluated and compared.