Relative navigation of an AUV using image-and-acoustic based profiling systems

Autonomous Underwater Vehicles (AUVs) are suitable for ocean observation, since they can swim freely without an umbilical cable. There is a great demand for AUVs not only for wide area surveys, but also for condition surveys of artificial underwater structures. This paper proposes a real-time localization method for navigation that can obtain robust and autonomous observations by fusing data from multiple sensors. This method enables an AUV to localize itself in relation to underwater structures without any support systems except its own sensors. It was actually implemented using the testbed AUV "Tri-Dog 1" and verified by tank tests