Research on feature extraction of ship-radiated noise based on multi-scale reverse dispersion entropy

Abstract

Aiming at the problem of ship-radiated noise feature extraction under the complex ocean background, a novel nonlinear dynamic analysis method, named multi-scale reverse dispersion entropy (MRDE), is proposed and applied to ship-radiated noise feature extraction. Furthermore, a novel complexity feature extraction method of ship-radiated noise is proposed based on MRDE. MRDE, as an improved algorithm of reverse dispersion entropy (RDE), can describe the complexity of signal from different scales. Firstly, a comparative study of RDE and MRDE is conducted on simulated signal to verify the effectiveness of MRDE; secondly, multi-scale permutation entropy (MPE) and MRDE features of four kinds of ship-radiated noise are extracted, six feature extraction methods of ship-radiated noise based on energy, frequency and complexity are compared, and the recognition rate is calculated by k-Nearest Neighbor (KNN) algorithm; finally, in order to further improve the recognition rate for four kinds of ship-radiated noise, the MRDE features under two scales are extracted and classified. The experimental results show that the proposed feature extraction method of ship-radiated noise based on MRDE has better separability and higher recognition rate; the recognition rate of ship-radiated noise can be further improved by adding one more MRDE feature. Therefore, the proposed feature extraction method can achieve accurate classification of different kinds of ship-radiated noise.