INTELLIGENT INTRUSION DETECTION AND ROBUST NULL DEFENSE FOR WIRELESS NETWORKS

摘要

In the past few years, plenty of research effort has been devoted to using cross layer design to enhance the performance of wireless communication systems. In this study, the major research effort is devoted to consider jointly adaptive neuro-fuzzy inference system (ANFIS) intrusion detection in the MAC layer and multimodal digital beamforming (DBF) in the PHY layer, with focus on improving the average detection delay (ADD) of denial-of-service (DoS) attacks and demonstrating the defensive function of the robust null extension mode of a multimodal DBF by taking advantage of a two-state Markov chain model. A packet acquisition and analysis system (PAAS) is designed to collect and analyze the friend, intrusion and interference information generated from the wireless networks. A prototype of the ANFIS-based intrusion detection system (ANFIS-IDS) is implemented, trained and tested against a real de-authentication DoS attack to empirically demonstrate the improved performance of the proposed ANFIS-IDS, compared with the IDS using non-parametric sequential change point detection (NPSCPD) algorithm.