Quantifying the performance of ray-based Blind Deconvolution (RBD) algorithm for long-range shipping sources

Abstract

The ray-based Blind Deconvolution (RBD) can provide an estimate of the relative arrival-times of the channel impulse responses (CIRs) between a shipping source (acting as a shallow source of opportunity) and a remote receiver array (e.g., a short bottom-mounted vertical line array (VLA)) [Durofchalk and Sabra, JASA 147(3), 1927–1938 (2020)]. Previous RBD studies primarily rely on beamforming on the (most energetic) direct path to estimate the phase of the unknown source; but in typical downward refracting environments the direct path is only observable at short ranges for a (very shallow) shipping source. Alternatively, the overall performance of the RBD algorithm when treating long range sources of opportunity remains to be evaluated in similar settings when only bottom-interacting multipath arrivals can reach the VLA, i.e., beyond the range when the direct path is observable. In this study, a combination of numerical simulations and experimental validation (using the SBCEx16 experiment conducted in the Santa Barbara shipping channel with short, bottom-mounted VLAs) is used to investigate the maximum detection range of shipping sources of opportunity via multipath arrivals as a function of frequency, bathymetry effects and anisotropy of the shipping source and the implications for estimating the CIRs using the RBD method.