Towards bathymetry-optimized Doppler re-navigation for AUVs

This paper describes a terrain-aided re-navigation algorithm for autonomous underwater vehicles (AUVs) built around optimizing bottom-lock Doppler velocity log (DVL) track lines relative to a ship derived bathymetric map. The goal of this work is to improve the precision of AUV DVL-based navigation for near-seafloor science by removing the low-frequency "drift" associated with a dead-reckoned (DR) Doppler navigation methodology. To do this, we use the discrepancy between vehicle-derived vs. ship-derived acoustic bathymetry as a corrective error measure in a standard nonlinear optimization framework. The advantage of this re-navigation methodology is that it exploits existing ship-derived bathymetric maps to improve vehicle navigation without requiring additional infrastructure. We demonstrate our technique for a recent AUV survey of large-scale gas blowout features located along the U.S. Atlantic margin