Horizontal Wavenumber Estimation Based on Deconvolution in Dispersive Shallow Water

In underwater acoustics, shallow water environments (water depth D < 200m) can be regarded as dispersive waveguides while transmitting low-frequency signals (frequency f < 500 Hz), and propagated signals are described as a superposition of some order normal modes according to modal theory, each of the modes propagates with its own wavenumber. When considering a horizontal line array (HLA) and a narrowband source, the wavenumber estimation is equivalent to a spectral estimation problem, which means a sufficient number of hydrophones are required to separate the wavenumbers accurately. As for broadband sources, the wavenumber spectrums estimated at each frequency can be concatenated into a frequency-wavenumber (f-k) diagram. This paper presents the application of deconvolution to compute the high-resolution wavenumber spectrum using a small aperture array that consists of a small number of sensors. The performance of this method is demonstrated by simulations in shallow water environments and evaluated by application on real data.