Preliminary results on terrain-aided localization algorithm for glider AUV

Glider class autonomous underwater vehicles operate and collect water column data for long periods of time, even weeks, before the need for recovery. However, they possess only rudimentary navigation and localization capabilities. Gliders are under-actuated vehicles and prone to water currents disturbances hence their localization error increase significantly while dead-reckoning underwater. In this workshop, we present preliminary results of a localization algorithm aided by existing seafloor map bounding dead-reckoning drift. The glider is equipped with mechanical scanning imaging sonar producing bathymetric patches of the seafloor, which are compared with existent bathymetry of the area. Glider data were collected in a field survey at Scott Reef, Australia.