Analysis To Communication Signal Subtle Features Based On Empirical Mode Decomposition

The methods for analysis to communication signal subtle features have become research hot spot in communication signal processing. Although the methods for analysis to communication signal subtle features have some development, they still confront with great difficulties in extraction of communication signal subtle features and selection of features. This dissertation focuses on the methods for analysis to communication signal subtle features based on Empirical Mode Decomposition(EMD). The main works of this dissertation are listed as follows:(1) A method based on EMD for extraction of Time-Frequency-Energy distribution features of transient signal is proposed. First, Time-Frequency-Energy distribution of transient signal is acquired by EMD and Hilbert Transform. Then calculate singular values of the Time-Frequency-Energy distribution by singular value decomposition(SVD), which are regarded as the subtle features of transient. Experiments have proved that the feature we extract has a stronger adaptability in identify communication emitters.(2) A method based on EMD for extraction of information dimension features of transient signal components is proposed. First, the mode components are acquired by EMD. And then, information dimension of each mode component is calculated as a kind of transient signal subtle feature. Experiments have proved the feature that we extract has a stronger adaptability in identify communication emitters.(3) A method based on EMD for extraction of stray components frequency spectra features of steady signal is proposed. First, the mode components of steady signal are acquired by EMD. And the main signal component and noise component are removed and calculate the frequency spectra of remain components by FFT. The frequency spectra of remain components are regarded as the subtle feature of steady signal. Experiments have proved the feature that we extract has a stronger adaptability in identify communication emitters.(4) A method based on EMD for extraction of stray components square integral bispectra(SIB) features of steady signal is proposed. First, the mode components of steady signal are acquired by EMD. And the main signal component and noise component are removed and calculate SIB of remain components. The SIB of remain components are regarded as the subtle feature of steady signal. Experiments have proved the feature that we extract has a stronger adaptability in identify communication emitters.(5) An system used for identification to communication emitters is designed. Experiments are carried out to test and verify all the methods proposed in this dissertation. And the results have indicated that the transient signal features and the steady signal features proposed in this dissertation has a stronger adaptability in identify communication emitters.