Wide area seafloor observation using an autonomous landing vehicle with adaptive resolution capability

Underwater vehicles are currently being extensively used for observation and study of the seafloor. Detailed seafloor analysis requires wide area observations with high resolution information obtained from in-situ analysis from sensors or by sampling. Certain sensors require high altitude and high scanning where as others need proximity to the seafloor or contact with stable footing to perform integrated measurements over a period of time. Such wide area high resolution surveys cannot be performed by a single cruising or hovering type vehicle and often requires multiple deployments of different types of vehicles. In this research the authors have designed and developed a new class of AUV along with an intelligent survey technique in which an underwater vehicle can generate centimeter order wide area maps of the seafloor by cruising at high speeds using an acoustic sensor. At intermediate locations of interest identified by processing the acoustic data autonomously during the mission, higher order resolution information can be obtained by lowering scanning speed and altitude using a laser profiling system. The vehicle can also land at some of these locations to obtain integrated measurements at the same position or magnified images. A slightly negatively buoyant underwater vehicle has been designed and developed with systems necessary to perform such a survey. An algorithm to detect areas of interest from acoustic scanned bathymetry has been implemented and tested. An autonomous landing system has been developed which uses laser scanned bathymetry to calculate a landing vector coordinate for safe landing on the seafloor. Experiments were conducted at a tank facility to demonstrate the multi-resolution survey scheme. An artificially generated seafloor scenario was acoustically scanned from a high altitude and speed to identify an area of interest in real-time. This area was then scanned at an higher resolution by lowering altitude and speed to perform autonomous landing by detecting suitable landing sites using the landing algorithm. The results from these experiments have been included in the paper.