Innovative manipulation techniques for underwater robotics

Abstract

This thesis is concerned with innovative manipulation techniques for underwater robotics and ocean engineering. In particular, it focuses on challenging problems of three different areas of the underwater manipulation, regarding the position feedback of the robotic arm, the visual feedback, and the manipulation of fragile objects. First, the position feedback has been addressed studying the kinematic performance of a hydraulic manipulator, in term of its accuracy and repeatability, used for underwater artifact cleaning activity. The manipulator has been re-designed during the CoMAS (Insitu conservation planning of Underwater Archaeological Artefacts) project. The results of the study have been of fundamental importance in the development of the control strategies for the control of the ROV and its manipulator. In fact, on the basis of the maximum error positions found and the kinematic performance of the arm it has been defined a “safety range” that allows avoiding collisions among the end-effector’s tool and the artifacts. Second, the visual feedback has been addressed presenting an augmented reality visualization of scene depth for aiding ROV (Remotely Operated Vehicle) pilots in underwater manipulation. The architecture and the software of the system have been developed during the CoMAS project, while in this thesis has been provided the calibration of the whole system. In particular, combining the kinematics of the robotic arm and the standard photogrammetric model of the stereo camera, it has been possible to generate a depth map that shows to the pilots the distances of the surface of the scene objects from the end-effector's pose. Experimental trials have been carried out in the laboratory and in the water tank in order to evaluate and improve the performance of the system, approaching the target softly. Despite the development of these feedbacks, currently, existing robotic manipulators are often too powerful and awkward to handle delicate or complex objects without damaging them. To figure this out, soft end effectors have been studied during my research internship at the faculty of Ocean Engineering of the University of Rhode Island (USA). The research activity has been carried out at the Robotics Laboratory for Complex Underwater Environments (R-CUE). In particular, it focuses on continues the development, prototyping, and testing of compliant jamming grippers based on Soft Robotics technologies. Specifically, the subject is divided into two main projects, studying both a universal jamming gripper and a hybrid toroidal soft gripper. The main purpose the universal jamming gripper has been to complete the integration with the existing arm, design and perform experiments with the gripper in the water tank, and propose refinements to the design of the mechanical and hydraulic system. While, the hybrid toroidal soft gripper has been designed, prototyped, and integrated with the existing arm and hydraulic system, to the end of carrying out a qualitative performance of the gripper in the water tank. The first extended trials of the hybrid toroidal soft gripper have been carried out at the Department of Mechanics, Energy, and Management (DIMEG) at the University of Calabria