Research On Denoising And Separation Technology Of Multicomponent Signals Basedon Variational Mode Decomposition

Nowadays,with the emergence of new system radars,the modulation form of radar signals is constantly changing.The transmission of battlefield commands and the changes of battlefield situation are increasingly dependent on the propagation of electromagnetic signals.In the increasingly complex electromagnetic environment,the number of signals simultaneously received by the radar receiver is also increasing,and even the spectrum of different signals crosses each other.For multi-component signals,the decomposition of each signal component is a key issue in signal processing in the new era.This dissertation mainly studies the noise reduction and separation techniques of multi-component radar signals based on the variational mode decomposition algorithm.Firstly,this dissertation analyzes the distribution characteristics of multicomponent radar signals,and summarizes the mathematical models of radar signals with different modulation forms.Based on this,the time-frequency distribution characteristics of multi-component signals are obtained.The research results show that the multi-component signal is caused by the superposition of different signal sources in the time domain,and the timefrequency distribution characteristics are the superposition of the timefrequency distribution characteristics of each signal component.When the spectrum of different signal components crosses,the time-frequency distribution will be significantly distorted at the intersection of the spectrum.Then,based on the study of variational mode decomposition,this dissertation denoises multi-component radar signals.According to the characteristics of the band-limited intrinsic mode functions obtained by VMD,the noise components are denoised by the interval thresholding,and the signal components are denoised by discrete wavelet transform.In addition,this dissertation also proposes a combination of VMD and least mean square algorithm in the process of noise reduction,and gives the corresponding system block diagram.Finally,since modern radar signals are mainly AM-FM signals with large bandwidth,the intrinsic mode functions defined in the variational mode decomposition are narrow-band.Therefore,this dissertation uses variational nonlinear chirp mode decomposition to separate multi-component radar signals.Based on the research of this algorithm,the ridge tracking variational nonlinear chirp mode decomposition is proposed,which improves the signal separation ability of the original algorithm under low SNR.In this dissertation,based on the study of variational mode decomposition,an adaptive noise reduction algorithm based on VMD is proposed for multicomponent radar signals.The VMD-ITn-DWTs can improve the signal-to-noise ratio by about 11 dB when the signal-to-noise ratio is-6dB in the process of performing noise reduction on the radar simulation signal.In terms of signal separation,the proposed ridge tracking variational nonlinear chirp mode decomposition,under the same SNR condition,the correlation error of the instantaneous frequency of each signal component can be reduced by about 30% compared with the original algorithm.