Assisted Teleoperation for Underwater Manipulation utilizing Touch Screen Inputs

Abstract

In the paper, an assistive teleoperation technique for underwater intervention is addressed. The key idea behind the proposed method is simple as that, by touching several points on the touch screen, one can move the manipulator to an object; it can reduce the burden for operators to teleop-erate every degree-of-freedom(DOF) of the manipulators. The assistance technique involves a position estimation algorithm which can estimate 6 DOF position of the object by utilizing 2D positions on the two images from cameras deployed in stereo. Control structure for the assistive teleoperation is also investigated. The structure involves an inverse kinematics scheme based on the weighted damped least squares. The scheme can handle workspace limit and singularity, and joint limits on position and velocity, which are easily violated when the desired position of end-effector is given in Cartesian space. Through the laboratory experiments, it is verified that the proposed assistance technique can reduce operation time meaningfully in comparison with the conventional teleoperation.