CFR-JVMD与多接收机协作的OFDM辐射源个体识别OA

To address the challenge that radio frequency fingerprints in specific emitter identification(SEI)are highly susceptible to multipath effects and receiver-induced distortions,this paper proposes a multi-receiver cooperative OFDM emitter identification method that integrates channel frequency response(CFR)correlation with joint variational modal decomposition(JVMD).First,by exploiting the multiplicative fa-ding characteristics of multipath channels in the frequency domain,a spectral cyclic shift ratio algorithm is developed to suppress multipath channel effects.Second,JVMD is applied to the normalized signals received by multiple receivers to jointly decompose the signals into modal components containing transmitter fingerprints and receiver fingerprints.Finally,transmitter fingerprint components are selected using a dynam-ic thresholding strategy based on non-Gaussian statistical measures and fed into a convolutional neural network(CNN)classifier to achieve spe-cific emitter identification.Experimental results demonstrate that the proposed method achieves an average identification accuracy of 97%on the ORACLE dataset and exhibits superior robustness against multipath fading and receiver distortions compared with conventional methods.