Securing Wireless Communication Using RF Fingerprinting

Abstract

Recently, RF fingerprinting has become an arousing and emerging technology in identifying multiple wireless devices. The method is also believed to have a strong impact on its applications in the wireless security system. Security has always been a critical issue for wireless devices including in the application of Wireless Local Area Network (WLAN). For instance, Media Access Control (MAC) spoofing which is a malicious technique of changing a factory-assigned MAC address of a Network Interface Card (NIC) installed in a device. Due to this issue, this study suggests on making use of a network device’s unique RF fingerprint obtained from its raw baseband IQ samples to identify the transmitting radio. For WLAN, as RF fingerprinting is a physical layer security implementation, WLAN physical layer protocol data unit (PPDU) which contains L-LTF in preamble is extracted. Particularly, the RF fingerprinting process includes deep learning of convolutional neural network (CNN) as a classifier. The neural network is used to train a model by tuning and test-validation test before finalizing it as a final model for classification method for a security system.