Tetherless robot for biofouling prevention and inspection in salmon farming

Biofouling, the growth of unwanted organisms on the net, is one of the main challenges in salmon farming. Currently, biofouling is managed through frequent pressure washing of the net. Such operations are labour intensive and have negative impacts on fish health and welfare. NetClean 24/7 aimed to introduce a new biofouling management strategy where biofouling is continuously prevented instead of periodically removed by developing a concept for a tetherless, autonomous robot for biofouling prevention, and monitoring of net integrity and environmental conditions. The project consortium consisted of the project owner Remora Robotics, Aanderaa Data Instruments, and Nordlaks Oppdrett, with SINTEF Ocean and NTNU as research partners. The goal was to develop this new tetherless robot that will provide the salmon industry with an efficient and robust solution for cleaning and monitoring of fish cages. By facilitating continuous monitoring of the cage environment, the robot aims to increase the farmer's control of the farming process and providing data for decision support during operations. Through advancing automation and remote management, this novel technology desires to support the expansion of salmon aquaculture and facilitate safer operations at new and more exposed sites. NetClean 24/7 thus contributed significantly to a more sustainable and future oriented aquaculture. As part of work package 1, the autonomous functions of the robot have been analysed using the planning tool Seatonomy, in combination with interviews of the project partners. The resulting detailed review of the capabilities of the robot and its sensors, as well as its communications and error handling capabilities are now available in form of a scientific article. In addition, a more detailed mathematical model of the robot that includes the effects of the net cage on the robot has been developed. This model has served as the simulation environment for the motion planning and control algorithms developed further in this project (both published as scientific articles). Two master's students at NTNU worked in close collaboration with researchers in SINTEF on the development of motion planning and navigation components and implemented this in a simulation environment where the movement of the robot on a static net cage were simulated and path following abilities could be tested. The developed method has been adapted and extended to address fully autonomous navigation of underwater robots. In particular, adaptive motion planning and control concepts have been implemented and tested in a lab environment (resulting in two journal articles). In these trials, the robotic system was able to autonomously update the planned path and avoided collisions with static and moving obstacles. In addition, Remora Robotics has developed algorithms for initial monitoring of cleaning efficacy and energy consumption of the robot. In WP2, an interactive and user-friendly graphic interface was developed for observation and interpretation of sensor data, and visualization of the collected measurements. This was based on manufacturer specifications regarding the outputs to be expected from the robot, as well as end-user inputs from e.g. Nordlaks regarding relevant information to be displayed. In particular, a mock-up graphical user interface (GUI) was developed as web application for monitoring a fleet of cleaning robots in aquaculture pens. To expand on commonly used sensors such as oxygen or currents, Aanderaa has developed a turbidity sensor that may complete the sensor equipment of the Remora robot. As part of WP3 a concept study was performed to determine the minimum requirements and specifications for the development of a docking station with respect to solutions for charging, data exchange, and mechanical design. This also included the specifications and hardware adaptations of how to integrate commercially available solutions in fish cages, which is essential for safe unmanned operations in fish farms. In addition, an initial hydrodynamic analysis was performed to study the effects of wave on the docking station. The results are available in form of a scientific article. Building on the outputs from WPs 1-3, WP4 focused on the adaptation of the Remora robot to an advanced tetherless cleaning robot permanently installed in a cage. This cumulated in the successful validation of the elastic band method for path planning and obstacle avoidance in net pens in a field test, thus demonstrating fully autonomous navigation of the robot in an industrial scale fish cage. In summary, the novel technologies will open for new and improved services and will give farmer better control over the dynamic and complex farming environment, thus contributing to sustainable growth in aquaculture. Moreover, the results of the project will open for further research by providing unique data and pushing the borders of using more autonomous solutions in aquaculture industry.

The project significantly contributed to the development of the Remora robot over the project period by providing knowledge on automation and biofouling management. As a result, the robot now has extended autonomous capabilities. While the robot is not yet tetherless, major developments for this transition have been undertaken. The Remora robot is well on its way to provide a much-needed alternative to today's standard biofouling mitigation strategy based on pressure-washing. The project included the development of a sensor package system for capturing relevant visual and environmental data, and further development of a turbidity sensor. These outcomes can be utilized by Remora Robotics to integrate a sensor package on the robot that is able to collect data for assessment of environmental conditions such as temperature, water current velocity and direction, salinity, turbidity and O2 at given positions and depths. This will provide fish famers with important knowledge on environmental conditions in cages that may impact fish welfare. Moreover, Remora Robotics has developed algorithms for initial monitoring of cleaning efficacy and energy consumption of the ROV, and the basis for automated assessment of net integrity has been laid. Automation of the cleaning and inspection process will support farming in remote locations and under challenging weather conditions, in addition to improved control of the cleaning process. An interactive and user-friendly graphic interface was developed for observation and interpretation of sensor data, and visualization of the collected measurements. In particular, a mock-up graphical user interface (GUI) was developed as web application for monitoring a fleet of cleaning robots in aquaculture pens. This may build the basis for a customised GUI system tailored specifically to the Remora robot and its customers. The conceptual docking station design provides input on how to adapt currently available docking station designs for deep water or ocean sea floor installations to use in fish farms. This lays the groundwork for a future docking station development, which is a requirement for safely conducting unmanned operations in fish farms and similar high-energy environments. The successful validation of the elastic band method for path planning and obstacle avoidance in net pens demonstrated fully autonomous navigation of the robot in an industrial scale fish cage is possible. The novel technologies will open for new and improved services and will give farmers better control over the dynamic and complex farming situation, thus contributing to sustainable growth in aquaculture. The results of the project will benefit further research by providing unique data and pushing the borders of using more autonomous solutions in aquaculture industry.

Biofouling, the growth of unwanted organisms on the net, is one of the main challenges in salmon farming. In the Norwegian salmon industry, biofouling is managed through a regime predominantly based on frequent high-pressure net cleaning. Such operations are labour intensive and have negative impacts on fish health and welfare. NetClean 24/7 will introduce a new biofouling management strategy where biofouling is prevented instead of periodically removed by developing a concept for a tetherless, autonomous robot for biofouling prevention, and monitoring of net integrity and environmental conditions. This compact and energy efficient robot will be permanently installed inside cages with a dedicated docking station. The project outcomes will represent technological breakthroughs that combine autonomous systems with sensors and docking station technology and thus found the next generation of cleaning and monitoring tools for the fish farming industry. Cameras and environmental sensors integrated in the robot will enable daily inspections of the condition and integrity of the net and collect data to assess the environmental conditions in the cage. The docking station will recharge the robot and transfer the collected data to the control room for further analysis and visualisation. This new tetherless robot will provide the salmon industry with an efficient and robust solution for cleaning and monitoring of fish cages. By facilitating continuous monitoring of the cage environment and structural integrity at fish farms, the robot will increase the farmer's control of the farming process and providing data for decision support during operations. Through advancing automation and remote management, this novel technology will support the expansion of salmon aquaculture and facilitate safer operations at new and more exposed sites. NetClean 24/7 will therefore contribute significantly to making aquaculture more sustainable and future oriented.