AUTHOR=Wang Jiahui , Cui Zhiqiang , Zhu Hanhao , Meng Lei , Song Weihua , Liu Xu 

TITLE=A back propagation neural network-based approach for inverting layered seabed acoustic parameters in shallow waters

JOURNAL=Frontiers in Marine Science

VOLUME=Volume 11 - 2024

YEAR=2024

URL=https://www.frontiersin.org/journals/marine-science/articles/10.3389/fmars.2024.1349478

DOI=10.3389/fmars.2024.1349478

ISSN=2296-7745

ABSTRACT=Most of the existing earth acoustic parameters inversion methods mainly focus on the earth acoustic parameters inversion calculation under a specific layered structure, but when processing the experimental data under unknown seabed stratification, the earth acoustic parameters inver-sion results cannot be accurately given by the stratification. To solve this problem, a new algo-rithm combining Back Propagation neural network (BPNN) is proposed to distinguish seabed stratification and inversion of acoustic parameters. This method is based on the simulated sound pressure generated by the disturbance of seabed parameters as the input data. Firstly, the feature recognition of the sound pressure data under different seabed stratification conditions is carried out, which is used to train the neural network inversion model under different stratification structures. Then, the acoustic parameters under the stratification are estimated by combining the sound field model. Finally, the simulation data and pool shrinkage data are used to compare and verify the inversion model. The results show that the neural network model based on the sub-marine sound pressure signal training under different layered structures can effectively stratify the simulation and experimental data, and give the inversion results of the acoustic parameters under the corresponding layers. The accuracy and practicability of the neural network model are further verified by the hierarchical judgment of the scale test data and the inversion calculation of acoustic parameters, which provides a new idea for the subsequent acoustic parameters inver-sion in shallow sea and has a good application scenario.