A Novel Feature Extraction Method for Underwater Acoustic Target Based on Parameter Optimized VMD and 1(1/2)-D Spectrum

Abstract

Due to the difficulty of correctly extracting the features of underwater target radiated noise in the complicated environment, a novel feature extraction method for underwater acoustic target based on parameter optimized variational mode decomposition (VMD) and 1(1/2)-dimensional spectrum is proposed. Firstly, the number of modal components and quadratic penalty term are optimized by particle swarm optimization (PSO) algorithm, and the results of PSO algorithm are used as the input parameters of VMD algorithm. Then, the radiated noise signal of underwater acoustic target is decomposed into several intrinsic mode functions (IMFs) by VMD algorithm. Secondly, the envelope entropy of each IMF is calculated, and the IMF corresponding to the minimum envelope entropy is selected as the optimal component. The Hilbert envelope demodulation analysis is performed on the optimal component to obtain the envelope signal. Finally, the envelope signal is analyzed by 1(1/2)-dimensional spectrum analysis method to extract modulation features. The simulated and the ship's radiated noise signals are analyzed by the proposed method, and the results demonstrate that the modulation features of signals can be extracted successfully, which indicates that the proposed method is effective.