A Comparative Study for Shallow Water Match-field Inversion Using Surrogate Models

The inversion for geoacoustic parameters is an important task in the study of underwater acoustic propagation and the match-field method is the most popular approach. When considering a series of uncertain environmental parameters with different ranges, the solution of match-field inversion normally involves frequent computation of replica sound fields, which results in a significant computational cost. In order to alleviate the computational burden and efficiently solve the inversion optima, surrogate-based optimization can be applied by making use of surrogate modeling techniques for sound-field calculation. Three types of surrogate modeling techniques including Polynomial Chaos Expansion (PCE), Kriging and PCE-based Kriging are applied in this study to formulate a surrogate-based match-field inversion framework. The performance of proposed method is validated on a two-layer numerical submarine model. The results show that the proposed method can greatly accelerate the match-field geoacoustic inversion process and achieve a balance between the inversion efficiency and accuracy.