Visual Servoing Control of Robotic System for Off-shore Oil and Gas Pipeline Inspection

Subsea pipelines are essential in off-shore oil and gas transportation. However, recent news has been reported that the leaking problem in Subsea pipelines has caused intensive losses in both the economy and the environment. This research is aimed to integrate the-state-of-art technologies in machine learning, visual servoing, Smart Touch inspection, underwater vehicles, and controls to enable automatic pipeline inspection. More specifically, this research introduces: 1) an optimal visual tracking and prediction algorithm which is robust in the complex environment with obstacles and light reflections; 2) a collision avoidance control for under-actuated robotic fish; 3) an integrated robotic system with a 4-degree-of-freedom (4-DOF) robotic arm that can automatically grab a flange and perform bolt looseness inspection using a pair of lead zirconate titanate (PZT) transducers; 4) a novel angle-based pipeline tracking control for a remotely operated vehicle (ROV) to track an underwater pipeline. In this research, visual detection and prediction, and collision avoidance are used for vision tracking and servoing control of robotic fish. The robotic arm, which can also be regarded as the manipulator for underwater vehicles, is guided by an on-board visual servoing system to achieve automatic Smart Touch pipeline inspection in which deep learning convolutional neural networks and stereo camera systems provide the 3D position of a targeted flange from images. Finally, the visual servoing control for pipeline tracking has been tested on a Blue ROV in a 10 m by 5 m swimming pool. Experimental results have demonstrated that the vehicle can move along the oil pipeline and monitor any failure and leaking problems.