Towards autonomous underwater iceberg profiling using a mechanical scanning sonar on a underwater Slocum glider

Abstract

A Slocum underwater glider is been modified to map the underside of icebergs for monitoring iceberg deterioration off the coast of Newfoundland, Canada. The vehicle is equipped with a mechanical scanning sonar to map the iceberg surface, and a thruster for level-flight at a higher surging speed. In this paper we are presenting a profile-following controller that uses the sonar ranges to compute desired headings guiding the Slocum glider traveling safely around icebergs. A vehicle-attached occupancy map (VOM) is updated using sonar measured ranges with a dynamic inverse-sonar model. A desired path is then generated from the VOM by applying polynomial regression on the occupied cells. The line-of-sight guidance law is implemented to compute the desired heading to follow the desired path. The algorithm is initially evaluated in a simulation environment. The vehicle operation is simulated on a real-time hardware simulator, while the sonar is modeled in ray-tracing method. The iceberg is derived from an iceberg database with additional translational and rotational motion emulating a floating iceberg. After that, the guidance system is applied on a set of field data collected in 2015. During the trial, the Slocum glider was deployed to profile an underwater ramp feature in Conception Bay, Newfoundland, Canada. The feasibility of the porposed controller is indicated by the outcomes from this paper.