An AUV Manipulator/Vision System for Autonomous Interventions

点击翻译按钮获取中文摘要

The PI?s request funding to develop an autonomous system for remote sampling of biological and geological specimens from the sea floor using a robotic manipulator mounted on an autonomous underwater vehicle (AUV). Visual sensing from a stereoscopic camera system called AVATAR will be integrated with the control system for the SAMURAI underwater manipulator to produce a robust autonomous sampling system. In addition, a front-end graphical user command interface will be integrated to the system to allow real-time designation of targets for autonomous sampling or other dexterous manipulation from remotely operated vehicles (ROV) and manned submersibles. The utility of this advanced autonomous sampling system will be demonstrated by teaming with investigators at the Woods Hole Oceanographic Institution by conducting operational sampling tasks of specimens in sea trials off the coast of Cape Cod and on an Atlantic scientific cruise. The SAMURAI/AVATAR system will provide marine scientists with the capability of autonomous or teleoperated interventional tasks from any hovering undersea vehicle. When ROVs or manned submersibles such as Alvin are being used, operators will be able to designate sampling targets and then go ?hands-off? while the system performs the sampling. This should dramatically reduce the workload on ROV operators in the most complex tasks, as well as potentially increase the reliability and repeatability of sample collection. Broader Impacts:The proposed system would have application to a range of scientific and practical interests ranging from undersea science to aids for the disabled. This would also create a collaboration between the space and oceanographic communities for a problem that is common to both fields. The project would also provide an educational experience for several graduate and undergraduate students. The success in this development has the potential to replace the necessity of presence (on a sampling site) of a highly skilled manual operator with a formulation of criteria for target detection and development of corresponding computer vision algorithms. Hence it allows for input from a wide range of interested persons, including those who cannot personally participate in an expedition. Development of a robust automatic sampling system is going to be useful for other vehicles operating in remote and hazardous environments - under ice, in the outer space, in speleological expeditions, in volcanic research.