An orthogonal subspace method for source depth discrimination in shallow water using a horizontal line array

Source depth discrimination with a horizontal line array in a shallow water environment generally involves mode filtering. Since a limited horizontal aperture can make the observation matrix an ill-conditioned matrix, which affects the mode filtering, the application of the Schmidt Orthogonalization to the problem of source depth discrimination with a horizontal line array is investigated. It aims to construct orthogonal trapped and nontrapped component subspaces while preserving an almost complete mode space and reducing energy loss. An orthogonal subspace discriminator is proposed, and the corresponding decision metric relies on estimating the proportion of the trapped component. The performance of each method is predicted by receiver operating characteristic curves obtained by simulation, enabling one to compare the proposed method with the comparison methods. The impact of horizontal aperture, frequency, signal-to-noise ratio, and environmental mismatch on performance is numerically studied. The results show that the proposed method performs better than comparison methods at smaller apertures and is more robust against environmental mismatch at lower frequencies. Finally, the method is demonstrated on experimental data acquired with a bottom-mounted horizontal line array deployed in the South China Sea.