Autonomous Underwater System for Targeted Assessment and Repair

The AQUASTAR project aims to develop technology that enables automatic localization and repair of holes in aquaculture nets. This includes the use of unmanned underwater vehicles (UUVs) and the development of systems for automatic hole detection, repair tools, and control systems for both the underwater vehicle and a robotic arm that performs the actual repair. The system is designed to allow coordinated interaction between the vehicle and the manipulator, where fast and precise movements of the robotic arm compensate for slow or unpredictable movements of the vehicle, for instance due to water currents or limited control accuracy. The research activities in the project are organized around five interconnected objectives: 1) structure- and object-relative navigation, 2) optical object detection, 3) control of vehicle-manipulator systems, 4) development of repair tools for net damage, and 5) validation of fully integrated systems in field trials. So far, the project has achieved solid results in navigation and object recognition. The optical detection system for identifying holes has been developed and deployed, enabling real-time detection of small holes in aquaculture nets—a capability that has previously been difficult to access commercially. This represents a significant technological breakthrough and demonstrates that parts of the system already have practical value. Structure- and net-relative navigation has been successfully demonstrated and is currently being optimized and further verified. Control algorithms for coordinated operation between vehicle and manipulator have been developed and tested in simulation environments, and are now being refined for use on physical prototypes. The repair tool is under development, and full system integration for field testing is planned for the final phase of the project.

In Norway, the main reason for fish escape in sea-based fish cages is due to holes in the net. Fish escape has a negative impact on the environment through mixed breeding and spreading of pathogens to wild fish and on production through loss of invested resources. It is common practice to employ remotely operated vehicles (ROVs) for net inspection, but it is challenging to accurately control the ROV while coping with environmental forces and avoid causing damage to the net. The task increases in complexity when the operator is also required to carefully look for holes and keep track of already inspected areas. Holes are typically repaired by divers who sew the net, but the duration between detection and repair can be long, and the repair operations involve both high personal risks and costs. Attempts have been made to repair holes using underwater vehicles and customized solutions but none of these are known to be widely applied. The AQUASTAR project will develop innovative technology for autonomous structure inspection, feature localization and structure repair, where aquaculture is chosen as a case study. The R&D challenges thus relate to net- and hole-relative positioning, hole localization, coupled vehicle-manipulator control for conducting repair operations, and hole repair tools. Solutions to these challenges may be considered enabling technologies for further research and innovations in other maritime industries where robotic systems are required to perform inspection, maintenance, and repair operations. The AQUASTAR project aims to reduce the number of fish escapes and at the same time improve the underwater robotic inspection, maintenance, and repair services in aquaculture.