EKF-Based Vision-Assisted Target Tracking and Approaching for Autonomous UAV in Offshore Mooring Tasks

Abstract

In the offshore oil and gas industry, the most productive fields are located in increasingly distant waters, so the oil pipelines as a solution to transport the production become less attractive to build and operate. Therefore, the offshore loading operations with shuttle tanker (ST) vessels take the lead in transporting oil production. These operations occur throughout the facilities’ lifespan and represent a risky and complex task, mainly in the approach, mooring, and connection stages between the oil rig and the ST vessel. This work’s goal is to present the improvements and experimental results of a previous theoretical framework which focuses on enhancing the safety and performance of this task by using aerial robots to transfer the messenger cable between the platform and the ST on mooring and connection operations for offloading. This framework consists of a hybrid navigation methodology based on computer vision artificial intelligence for object tracking, with a sensor fusion technique using extended Kalman Filter (EKF) for autonomous unmanned aerial vehicles (UAVs) to safely navigate to and approach the destination, and therefore it could be used to assist critical stages of the offloading operation, improving safety. The proposed framework was developed in the robot operating system (ROS) and tested on a physical UAV using in scale ship models to validate the theoretical results (obtained in previous simulations). It was also subjected to a field test as proof of concept. The results demonstrated both the robustness and viability of the proposed methodology and also the cost reduction potential for real-life operations.