Terrain based navigation for an autonomous surface vehicle with a multibeam sonar

Abstract

Bathymetric measurements are useful information both in underwater and surface navigation when a global navigation satellite system (GNSS) is denied. Autonomous surface vehicles (ASVs) equipped with a multibeam echosounder can obtain an array of bathymetric measurements, which can be used in the localization with pre-surveyed terrain maps. In this paper, terrain based navigation (TBN) for ASVs using sonar multibeam measurements is proposed and implemented within a particle filter framework. By employing terrain roughness measure, standard particle filter framework is modified to resolve the particle impoverishment problem which occurs around the areas with smooth terrain. Improved localization results are presented from the field experiments performed in an inland water environment using our surface vehicle. ？ 2019 IEEE.