Adaptive denoising model for ship-radiated noise based on dynamic weighted filtering

Abstract

Aiming at the problem of ship-radiated noise (S-N) denoising at low signal-to-noise ratio, adaptive denoising model for S-N based on dynamic weighted filtering is proposed. Firstly, improved variational mode decomposition with multi-strategy enhanced dung beetle optimizer (MEVMD) and the secondary MEVMD are proposed to deal with S-N which has the nonlinear and nonstationary characteristics. Secondly, density-based spatial clustering of applications with noise assisted by fluctuation dispersion entropy (FDBSCAN) is proposed to adaptively divide all modes into multiple groups according to the chaotic degree of the sequence. Finally, dynamic weighted filtering (DWF) is proposed to filter each group of mode components, and the final denoised signal is obtained through weighted reconstruction. The experiment of simulating chaotic signal proves the effectiveness of the proposed denoising model by seven evaluation indexes. The proposed denoising model has been verified in four kinds of S-N. It will contribute to the subsequent feature extraction and classification research for S-N.