Automatic object detection for AUV navigation using imaging sonar within confined environments

Abstract

Navigation is a crucial issue for the operation of autonomous underwater vehicle (AUV). Range sensors mounted on an AUV can provide real-time information about the surroundings. The use of an active sonar sensor in an enclosed shallow water environment results in received echoes that may reflect several times before attenuation. Furthermore, echo intensities from boundaries are much higher than those from targets, which impedes their detection. This paper presents an innovative method for the detection of underwater objects (in this case, spherical shapes) in such environments for AUV navigation. A mechanically scanning sonar sensor was used in laboratory based trials. Experiments have been conducted in a cylindrical water-tank with an elliptical cross section. Results indicate that the presented method leads efficient and stable detection of both quiescent and moving objects. Potential applications include the detection of targets inside shipwrecks and/or hostile environments, e.g. discovering suspicious objects in the shallow water near harbor areas or water channels connected to the sea.