Separation of Multiple Backscattered Echoes Using Dictionary Updating Sparse Method

Xiangxia Meng,Xiukun Li,Andreas Jakobsson

DOI: https://doi.org/10.1121/1.5036086

2018-01-01

The Journal of the Acoustical Society of America

Abstract:Sparse signal representation method has been proven to be effective to estimate and separate the multiple backscattered echoes from underwater target. Usually, the transmitted signal is used as a template to generate the dictionary. It is efficient under assumptions that the knowledge of signal components is perfectly known, or the deviations from the dictionary can be ignored. However, the properties of the echoes differ from the clean reference signal because of the scattering mechanism. In contrast, it has necessity to extract these deviations to have deep insight of the target. Herein, we propose a sparse method that allows the dictionary to be updated. The initial dictionary is defined as a usual way, i.e., based on the transmitted signal. Then, we use a constrained optimization problem to generate dictionary that differs from the original one, introducing an updating procedure. We introduce two constraints, one is to penalize the updated dictionary to be sparse and the other one is to ensure the elements do not deviate too much from the original dictionary. In this way, the resulting dictionary can be used to reconstruct the backscattered echoes, which is illustrated to be efficient with high accuracy by simulated and experimental results.