Vision-Based Localization System Suited to Resident Underwater Vehicles.

Area of Science:

• Robotics and Autonomous Systems
• Ocean Engineering
• Navigation and Control

Background:

• Increasing demand for permanently deployed resident robotic vehicles in commercial inspection, maintenance, and repair (IMR) activities.
• Challenges in maintaining accurate vehicle navigation, specifically Inertial Navigation System (INS) drift, when operating close to subsea targets.
• Need for robust and cost-effective localization solutions for autonomous underwater vehicles (AUVs).

Purpose of the Study:

• To present a novel, low-cost, and low-maintenance navigational marker system for subsea localization.
• To demonstrate the system's capability to eliminate drift in a vehicle's INS solution near IMR targets.
• To evaluate the system's performance and error propagation during offshore trials.

Main Methods:

• Development and deployment of a fixed subsea localization marker system on resident field assets.
• Integration of the marker system with on-vehicle machine vision algorithms for pose estimation.
• Utilizing high-resolution world coordinate frame registration with a high refresh rate for accurate positioning.
• Conducting offshore trials in the North Atlantic Ocean, comparing system performance against a commercial state-of-the-art INS in pure inertial mode.

Main Results:

• Successful elimination of vehicle INS drift when in proximity to the navigational marker.
• Demonstration of high-resolution pose estimation and world coordinate frame registration.
• Evaluation of system performance, including position error propagation, during trials.
• Analysis of the impact of intermittent visual position updates on Kalman filter and INS solutions.

Conclusions:

• The developed subsea localization marker system effectively enhances the navigational accuracy of resident robotic vehicles for IMR tasks.
• The system provides a reliable method for pose estimation and accurate coordinate frame registration, crucial for close-proximity operations.
• Experimental results validate the system's performance and offer insights into error mitigation strategies for AUV navigation.