试验台
计算机科学
任务(项目管理)
水下
对象(语法)
控制器(灌溉)
有限状态机
自动控制
夹持器
人工智能
计算机视觉
实时计算
计算机硬件
控制工程
工程类
地质学
海洋学
生物
系统工程
机械工程
计算机网络
算法
农学
作者
Hong Wang,Stephen M. Rock,Michael Lees
标识
DOI:10.1109/oceans.1995.526796
摘要
Automated retrieval of instrumentation packages from the sea floor is a task well-suited for autonomous underwater vehicles. By constraining the generic object-manipulation problem to one in which the object is specifically designed to be picked up by a particular robotic system, both sensing and manipulation can be simplified. Readily available vision processing hardware along with well-established vision algorithms can be used to locate the package. The strategy for automated object retrieval by an underwater vehicle is encoded using a finite-state machine (FSM). Each phase of the automatic pickup is represented by a state. As monitors detect the successful completion or failure of a phase, events are issued that trigger the FSM to transition to the next phase of the task or to an alternative, corrective action. These events cause the vehicle control system to change control modes or desired setpoints within the activated controller. The automatic retrieval task is demonstrated experimentally with the OTTER testbed.
科研通智能强力驱动
Strongly Powered by AbleSci AI