RF Fingerprint Recognition Method Based on DBN-SVM

Authors:

Abstract
Radio frequency fingerprint identification technology is an emerging technology to ensure the security of wireless network. Aiming at the classification of multiple transmitters with highly similar structures and emitting the same RF signal, this paper proposes a radio frequency fingerprinting method based on deep belief network and support vector machine (DBN-SVM) model. First, the signals of different devices are preprocessed, and then the DBN model optimized by Quantum Particle Swarm Optimization (QPSO) is used for feature extraction. Finally, the features extracted by the DBN are put into the hierarchical SVM (H-SVM) for classification to obtain the classification results. The experimental results show that the accuracy of the RF fingerprinting algorithm based on DBN-SVM reaches 92.53 % in complex electromagnetic environment and real channel conditions.

This publication has 9 references indexed in Scilit: