Feature extraction of underwater acoustic signal using mode decomposition and measuring complexity

Abstract

Feature extraction in complex ambient noise is an important technology in underwater acoustic signal processing. A new feature extraction method using mode decomposition and measuring complexity is proposed to extract the feature information of underwater acoustic signal. In this paper, several set of intrinsic mode functions (IMF) are generated by decomposing three sorts of ship-radiated noise (SRN) using the ensemble empirical mode decomposition (EEMD) method. Then, select one IMF that contains the most dominant information. In the end, a new version of permutation entropy (PE) of the chosen IMF is calculated. By comparing the new PE of various sorts of SRN, it is easy to distinguish three sorts of SRN. The result of the simulations demonstrate the proposed feature extraction method is superior to the existing methods, which can effectively improve the recognition ability for various sorts of ships.