A radio frequency fingerprint recognition method based on convolutional neural networks and attention mechanism
To address the issues of high computational complexity and poor robustness in traditional methods based on hand-crafted feature extraction,this paper proposes a radio frequency fingerprint recognition method that combines an attention mecha-nism with a Convolutional Neural Network(CNN).This method leverages the powerful feature extraction capabilities of the model,while the attention mechanism focuses on key information.Without manual intervention,it reduces complexity while improving recog-nition efficiency and accuracy in interference-prone scenarios.Experimental results on radio frequency signal datasets from different scenarios and signal types show that,compared to CNN without the attention mechanism,the proposed method significantly improves recognition accuracy for 5 and 10 types of USRP,reaching 99.5%and 98.1%,respectively.This validates the method's effectiveness and advantage in individual emitter recognition,without the need for preprocessing or manually designed fingerprint features.
radio frequency fingerprint identificationconvolutional neural networkattention mechanismradiation source
万方数据
