Identification of RF Signals Based on 1-D Local Binary Pattern

Abstract

A novel technique to enhance the security at physical layer of wireless networks is introduced in this work. The technique is applied based on the usage of Radio Fingerprinting (RF) for Bluetooth (BT) signals emitted from BT transmitters. The technique which is Local Binary Pattern (LBP) is applied to extracting the features of RF signals. The method generated Histocount and Histogram of the transient portion of BT signals. Sixty records from each transmitter were subjected to the 1-D LBP technique. The transmitters, which are six transmitters, were classified and identified via extracted statistical features from the count of the histogram of the signals. The third of extracted features data are uploaded to a learning machine system for training. The rest of the data were used to test the classifiers. Four classifiers were applied to the train and test data to identify the subjected transmitters. The results demonstrate the effectiveness of the applied methods and the powerful of the implemented classifiers.