Deconvolved Conventional Beamforming for a Horizontal Line Array

T. C. Yang
DOI: https://doi.org/10.1109/joe.2017.2680818
IF: 3.883
2017-01-01
IEEE Journal of Oceanic Engineering
Abstract:Horizontal line arrays are often used in underwater environments to detect/separate a weak signal and estimate its direction of arrival from many loud interfering sources and ambient noise. Conventional beamforming is robust but suffers from fat beams and high-level sidelobes. High-resolution beamforming, based on the inverse of the signal covariance matrix, such as minimum-variance distortionless response (MVDR), yields narrow beamwidths and low sidelobe levels but is sensitive to signal mismatch and requires many snapshots of data. This paper applies a deconvolution algorithm used in image deblurring to the conventional beam power of a uniform line array (spaced at half-wavelength) to avoid the instability problems of common deconvolution methods. The deconvolved beam power yields narrowbeams, and low sidelobe levels similar to or better than high-resolution beamforming and at the same time retains the robustness of conventional beamforming. It yields a higher output signal-to-noise ratio than conventional (and MVDR) beamforming for isotropic noise. Performance is evaluated with simulated and real data.
What problem does this paper attempt to address?