Specific emitter identification based on fractal and wavelet theories
Huanhuan Wang, Tao Zhang · 2017
Considering the characteristics of communication signal from the whole and local all together, it can improve the classification accuracy. A new feature extraction algorithm of communication signals based on the fractal and wavelet theories is proposed. Employing preprocessing the received signal, the correlation dimension of empirical mode decomposition(EMD) is researched to extract feature and they are proved to be effective; As applying the wavelet method to communication signal analysis, The wavelet entropy characteristics which represent sources are extracted to be feature vectors. These features combining the correlation dimension and wavelet entropy are proved to be effective by identification experiment based on Support Vector Machine(SVM) classifier.