Occupancy Grid Mapping in an Underwater Structured Environment

This paper presents practical results about the occupancy grid mapping of an underwater man-made environment using a sensor suite commonly available in nowadays Autonomous Underwater Vehicles (AUVs). The proposed algorithms are tested to be incorporated as part of the design of a new motion control system to integrate reactive obstacle avoidance with local path planning techniques to provide safe real-time guidance capabilities. The paper focus on the use of a sonar scan matching improved dead-reckoning navigation (Doppler Velocity Log (DVL) and Motion Reference Unit (MRU) based) together with an standard occupancy grid mapping algorithm. A conventional inverse sensor model for a sonar profiler is used and compared against a new inverse sensor model proposed to take advantage of the use of widely available imaging sonars. The system is validated experimentally on a dataset gathered with an AUV guided along a 600m path within a marina environment.