Identification of probe request attacks in WLANs using neural networks

摘要: Any sniffer can see the information sent through unprotected `probe request messages' and response in wireless local area networks (WLAN). A station (STA) send probe requests to trigger responses by simply spoofing a genuine media access control (MAC) address deceive point (AP) controlled list. Adversaries exploit these weaknesses flood APs with requests, which generate denial of service (DoS) STAs. The research examines traffic WLAN using supervised feed-forward neural network classifier identify frames from rogue frames. novel feature this approach is capture user attacker training data separately label them prior without administrator's intervention. model's performance validated self-consistency fivefold cross-validation tests. simulation comprehensive takes into account real-world environment. results show that detects attacks extremely well. This solution also an attack during early stage communication, so it prevent any other when adversary contemplates start breaking network.