面向射頻指紋信號(hào)分析與智能識(shí)別的研究綜述

打開文本圖片集

中圖分類號(hào)：TN919 文獻(xiàn)標(biāo)志碼：A 文章編號(hào)：1673-2340（2025）02-0001-21

引文格式：，等.面向射頻指紋信號(hào)分析與智能識(shí)別的研究綜述[J].南通大學(xué)學(xué)報(bào)（自然科學(xué)版），2025，24（2）：1-21.

Abstract： In the context of next-generation wireless communications and multi-source heterogeneous network systems，traditional cryptographic mechanisms and security protocolsposesignificantrisks in Interet of things （IoT）environments.There isanurgent demand for more efficientandreliable identityauthentication technologies.Radio frequency fingerprinting identification （RFFI），which leverages the inherent signal characteristicsof wireless devices，providesanovel approach toaddressing device authenticationand securitychalenges.Unlike existing reviews that focus onselected aspectsof RFFI froma broad perspective，this paper proposes asystematic and comprehensive framework. It beginsbyexplaining thefundamental principlesand characteristicsof radio frequency fingerprint （RFF）.Then，from the perspectivesof statistical featuresand deep learning （DL）-based features，the paper presents an in-depthreviewof RFFI clasification and identification methods，along with a comparative analysis of the two approaches supported by experimentalvalidation.Finally，severalpotetialesearchdiections inintellgentRFFaediscussed，ndfuureteds ofRFFtechnologyareexplored，aimingtoofferboththeoreticalinsightsandpracticalguidanceforongoingresearch andreal-worldapplications.

Keywords：radio frequency fingerprint；specificemitteridentification;deeplearning；featureextraction;Internetof things;wireless communications; intelligent identification

截至2020年，全球聯(lián)網(wǎng)的物聯(lián)網(wǎng)設(shè)備數(shù)量已超過217億臺(tái)，預(yù)計(jì)到2025年將增長(zhǎng)至412億臺(tái)。（剩余36925字）