Modal depth function estimation over a partial spanning VLA
Résumé
Modal propagation in shallow water is characterized by a set of frequency dependent modal depth functions. Estimation of modal depth function is an interesting problem when no a priori knowledge of the environment is available. Classical mode estimation methods rely on SVD decomposition of the field. This imposes, between others, that the reception VLA perfectly spans the full water column. To overcome these limitations, this paper suggests using a low-frequency impulsive source. In this case, modal contribution can be separated in the time-frequency domain using appropriated signal processing tools known as warping. Once the modes are separated, modal depth function estimation becomes trivial and can be carried out over a partial water column spanning array. Unlike classical modal depth function estimation method, our algorithm is not frequency averaged and allows mode estimation at any given frequency excited by the source. The proposed method is validated on realistic noisy simulations.