Aquaculture is an industrial sector of immense growth, with the production taking place further from the coast, in more exposed areas, in larger cages, and with more biomass. Underwater autonomy is at the forefront for maintaining this growth in the near future in a sustainable way, by improving employee and environmental safety, as well as efficiency. Fundamental problems, for autonomous aquaculture operations involving underwater robots, include navigating to target positions and inspecting target structures with strong safety guarantees. In a domain with poor sensing conditions, disturbances, currents, dynamic and deformable surroundings, and uncertain sensing and controls, most state-of-the-art methodologies struggle immensely.
In SINTEF Ocean, having experience with the challenges of underwater robotics, especially in the context of industrial scale fish farming in Norway, we have focused on addressing the fundamental problem of robust autonomy. Our goal is to both provide novel general methodologies that scientifically improve the state-of-the-art in multiple fronts, and adapt solutions tailored to aquaculture industrial problems. The past years we have developed motion planning solutions that address a subset of important challenges that enable robots to move and monitor their surroundings with efficiency and strong safety guarantees, under the presence of dynamic and deformable obstacles, unpredictable disturbances, highly uncertain motion and localization capacity, including the intersection of all aforementioned challenges.
Moreover, while considering conventional robotic or remotely operated platforms used in fish farms currently (Fig. 1), we develop novel solutions and methodologies that enable safe operations of bioinspired articulated robotic platforms (Fig. 2); an emerging trend that may offer new capacity for livestock monitoring and behaviour understanding.