Sparse Reconstruction of Gravity Plume using Autonomous Underwater Vehicle

Abstract

Ongoing researches in the polar region require smaller-scale under-ice observations for a better understanding of the atmosphere-ice-sea interaction. Autonomous Underwater Vehicles (AUVs) are the increasingly favored instruments for sensing the under-ice ocean in the polar area. Because of their mobility and carrying capacity, AUVs are able to sense the under-ice oceanographic fields, such as temperature, salinity, or velocity field. However, when an AUV observing a rapidly changing temperature field such as a gravity plume, the AUV sampling data distorts by the effect of Doppler smearing and aliasing. Moreover, blind areas of the AUV sampling exist and hinder the further usage of the data. In this paper, we propose the sparse approximation method for reconstructing dynamic temperature fields from AUV sampling data. Using sparse approximation and linear interpolation approaches, the reconstruction of a simulated dynamic temperature field of a gravity plume from simulated AUV sampling data is respectively presented. It shows that the proposed method achieved high-quality overall reconstruction results. With this method, the blind areas of the AUV sampling were complemented correctly with high accuracy.