Research on C-E Fingerprint Extraction of Emitter
Jiaying Yue, Liuyang Gao, Nae Zheng · 2019
Aiming at the problem of limited presentation ability and easy drift of results for a single feature, a joint fingerprint feature extraction algorithm based on the complexity and entropy of instantaneous parameter (C-E) is proposed. From the perspective of signal information integrity, this paper multi-extract the secondary features of instantaneous amplitude, frequency and phase, and obtain the box dimension, information dimension, information entropy, as well as the Hilbert envelope spectrum information entropy newly proposed in this paper. Finally, the features are fused into C-E features to identify the individual radiation source. Compared with single class features, the recognition rate of the joint features is greatly improved at little time cost. In simulation experiment, the accuracy increases by 15.2% and 19.7% than fractal dimension and entropy respectively. And features have good independence and noise resistance under different environments.