수중 작업을 위한 무인 잠수정-매니퓰레이터 시스템의 여유 자유도 해석 및 강인 제어

Abstract

This thesis addresses a robust coordinated motion control of underwater vehicle-manipulator systems (UVMS) for autonomousmanipulation. Autonomous manipulation can overcome the limitation of remotely operated vehicles with tele-manipulators such as inaccurate trajectory tracking, difficulty in attaining precise force control, time delay in the man？machine control loop, operator fatigue, and manual coordinations of motions between the vehicle and manipulators for the task outside current workspace of manipulators. For this purpose, a UVMS should be able to determine its configurations corresponding given tasks and track the planned trajectories while subjected to external disturbances induced by ocean current, dynamic coupling caused by manipulator motions, modeling errors, and so on.Firstly, for achieving these requirements, this thesis suggests a framework for the robust coordinated motion control which consists of the redundancy resolution to resolve inverse kinematics of UVMSs, and the compensation of restoring vector. As a UVMS performs certain tasks, the restoring moments of the UVMS change, and then these changed moments lead to spontaneous motion of the end-effector. If the direction of this motion is similar to that of the end-effector？s tasks, then part of the motion can be used to control the UVMS. In other words, the restoring moments can be used to control the UVMS motion. Hence, the proposed framework considers a redundancy resolution of the UVMS where the redundant degrees of freedom are used to selectively optimize the restoring moments. For this, a performance index with variable gradient gain is newly proposed, in which the gain is determined by the result in the comparison of the task direction with the direction of the restoring moments. The compensation of the restoring forces and moments makes up for the difference between the performances due to the desired dynamics and the restoring moments. In addition, the restoring vectors are consistently updated under certain constraints. The proposed framework requires only masses, buoyant forces, and centers of gravity and buoyancy, not any hydrodynamic parameters. In addition, by the proposed framework, a UVMS can perform specific tasks with less control input and achieve smaller tracking errors compared to conventional control systems.Next, a robust tracking control of the UVMS is proposed. In addition to the feedforward of the reference motion, a sliding mode control is designed so that the controlled system has robustness against extended disturbances. From the viewpoint of the form of control, this control can be regarded as a nonlinear H_inf with disturbance observer and it is stated that the proposed control has capability to estimate and compensate disturbances. Moreover, the tracking performance of the proposed control can be improved by the gain adaptation. By this control, the UVMS shows robustness against parameter uncertainties, effects by ocean current, external disturbances.