A Signal Envelope Extraction Method Based on Sparse Reconstruction Optimization Algorithm

Abstract

The invention discloses a signal envelope extraction method based on a sparse reconstruction optimization algorithm, which adopts a particle swarm sparse reconstruction optimization algorithm, and takes the change factor of changing the frequency bandwidth of the DCT basis as a method based on particle swarm sparse reconstruction. The properties of each particle in the structural optimization algorithm, through multiple iterations to obtain the global optimal position of the upper envelope and the global optimal position of the lower envelope, corresponding to the upper envelope acquisition process used to change the determination of the particle The change factor of the frequency bandwidth of the DCT basis required and the change factor of the frequency bandwidth of the DCT basis required to change the lower envelope acquisition process determined by the particle, and then obtain the best DCT basis that is most suitable for the change trend of the upper envelope and the best DCT base that is most suitable for the changing trend of the lower envelope, and finally adaptively extract the best upper envelope and the best lower envelope, which improves the extraction accuracy of the upper and lower envelopes, At the same time, the anti-noise performance of the upper and lower envelope extraction is improved.

Description

technical field [0001] The invention relates to a signal processing technology, in particular to a signal envelope extraction method based on a sparse reconstruction optimization algorithm. Background technique [0002] In the field of mechanical fault diagnosis and bridge vibration detection, many vibration signals contain modulation signal components, and different faults produce different modulation forms. An important part of the analysis and processing of the modulated signal is demodulation. Demodulation analysis can obtain the envelope curve and envelope demodulation spectrum of the modulated signal, and the envelope of the modulated signal often carries useful information concentratedly. Therefore, in In many fields such as signal processing and fault diagnosis, the research of envelope demodulation method has always been the focus of many scholars. [0003] Currently commonly used envelope analysis methods are: generalized detection filter, Hilbert transform, cubic...

AI Technical Summary

Problems solved by technology

Among them, the generalized detection filter may produce frequency mixing effect in the process of envelope extraction, so it is required to select an appropriate sampling frequency to sample the signal; the method of extracting the signal envelope based on Hilbert transform has weak anti-noise, when the signal-noise When the ratio becomes smaller, the envelope error obtained by the Hilbert transform will increase accordingly; the extraction of the envelope based on the cubic spline interpolation method will cause the problem of endpoint effect, which will affect the accuracy of the empirical mode decomposition

Images