Session: 13-01-02 Blue Economy II

Paper Number: 125653

125653 - High-Fidelity Simulation Platform for Autonomous Fish Net-Pen Visual Inspection With Unmanned Underwater Vehicles in Offshore Aquaculture 

Unmanned Underwater Vehicles (UUVs) have become essential tools in offshore aquaculture, predominantly due to the challenging working environments, safety concerns for human workers, labor shortages, and the potential for improved productivity. However, the ambition of achieving fully autonomous UUV deployment faces several barriers ranging from the development and validation of real-time controllers to the preparation of field trials and actual deployment — a complex and costly process involving multiple stakeholders such as policymakers, infrastructure owners, management, development engineers, and deployment engineers — to obtain process clearance and assess operating costs. In this context, high-fidelity simulation emerges as a crucial component, serving as an informed decision-making tool to the various stakeholders, while also offering a cost-effective and efficient means to pre-validate the advantages and disadvantages of UUV deployment in the face of uncertain and harsh underwater environments.

One such use case is the autonomous fish net-pen visual inspection with UUVs in offshore aquaculture, and the importance of high-fidelity simulation in this context is underscored by the numerous challenges and complexities inherent in the adoption of this technology. These challenges include challenging operational constraints of fish farms and underwater conditions, such as the nature of close infrastructure inspection, deep waters, strong currents, low visibility, and varying water quality which are difficult to model mathematically and need to be taken into consideration in the development of the UUV localization, navigation, and control systems that can operate well in real-time. Other challenges include operational constraints of UUVs, such as limited battery life, real-time data processing, and adaptability to unforeseen obstacles or technical malfunctions.

Furthermore, the preparation phase for UUV deployment involves a complex interplay of multiple stakeholders, each with its own set of considerations and interests. Policymakers must ensure that regulatory and environmental standards are met, infrastructure owners must make substantial investments in UUV infrastructure, management is tasked with overseeing operational logistics, and development and deployment engineers must work together to optimize UUV systems for effective operation. The collaboration of these diverse stakeholders, along with the consideration of operating costs, introduces additional complexities to UUV deployment planning. Consequently, high-fidelity simulation offers a valuable means of bringing all involved parties onto the same page, enabling them to visualize and understand the challenges and benefits of UUV deployment in uncertain and harsh environments.

Considering these challenges, this paper presents the development of a high-fidelity simulation for autonomous fish net-pen visual inspection with UUVs, adapted from the UUV simulator which is based on the Robot Operating System (ROS) and Gazebo Physics Engine. Thanks to ROS's open-source software platform with robust community support and a comprehensive ecosystem, the high-fidelity simulation approach provides a practical experimental simulation study. It also incorporates a wide range of libraries and tools that can be seamlessly integrated throughout all stages, from prototyping and development to deployment, offering a cost-effective and time-efficient solution. Critical factors for the application of high-fidelity simulation of UUVs in the practical working environment and operational constraints are also discussed. Therefore, this work highlights the indispensable role of high-fidelity simulation in uniting stakeholders, mitigating operational challenges, and pre-validating the merits of UUV deployment, ultimately contributing to the sustainability and success of offshore aquaculture operations.

Presenting Author: Thein Than Tun Auckland University of Technology (AUT)

Presenting Author Biography: Thein Than Tun is a PhD candidate from the School of Engineering, Computer and Mathematical Sciences at Auckland University of Technology (AUT), New Zealand. His PhD research focuses on energy-optimal control of unmanned underwater vehicles (UUVs) in offshore aquaculture. Previously in Singapore, he worked in the robotics industry and research laboratory at the Singapore University of Technology and Design (SUTD). He completed a bachelor of engineering in Engineering Product Development (EPD) from SUTD in 2016 and a Diploma in Mechatronic Engineering from Ngee Ann Polytechnic, Singapore, in 2012.

Authors:

Thein Than Tun Auckland University of Technology (AUT)
Loulin Huang Auckland University of Technology (AUT)
Mark Preece The New Zealand King Salmon Company