Utvikling av teknologi for autonom, biointeraktiv og høykvalitets datafangst i merdrommet

Finfish aquaculture is an important global contributor to the production of seafood for human consumption. However, this industry is also known for a substantial HSE risks and a high frequency of work-related injuries. Increasing the automation level of high-risk operations within aquaculture could therefore lead to economical as well as social and ethical benefits. Increased automation can contribute to improving the level of control humans have over aquaculture operations. A high level of autonomy will also ensure that data on the condition of fish in farms can be obtained with a higher degree of repeatability and objectivity. The aim of this project was to develop underwater technology with autonomous functionality for mission planning to achieve high-quality data acquisition from the cage volume. One of the most important capabilities within this context is to operate in interaction with the biomass and the deformable structures. The developed technology provides continuous and close follow-up of the current situation inside the cage. The project addressed several challenges within the aquaculture industry related to poor accuracy and representative sampling of important variables to describe the farm situation in detail and as a whole. The project was conducted by a consortium consisting of WaterLinked AS, Sealab AS, Norsk Havservice AS, NTNU, HES-SO and SINTEF Ocean AS. Among the key elements of the project is a robust high-bandwidth and low-cost communication system. The project has developed and validated a low cost hydroacoustic subsea communication and position reference system for use in fish cages. The unique feature of the developed system is its small physical size and the highly robust datalink provided to the user. The developed underwater positioning system has been used in combination with numerical methods and estimation techniques to realize a relative position reference system where the main challenge was to develop a realistic real-time map of the fish cage. The obtained results from field demonstrations for both real-time map estimation and underwater positioning of the vehicle showed good accuracy and suitability for autonomous navigation of underwater vehicles in cages. The main element of the project is to capture high-quality vision data. To obtain such data using currently available state-of-the-art systems (e.g. based mostly on stationary sensors) is a highly demanding process, and fails in many cases to obtain data describing the dynamic farming environment with sufficient resolution and accuracy. An autonomous underwater vehicle being equipped with a 3D vision system will be able to collect data from the entire volume of the cage. This project developed a 3D vision system which enables the acquisition of high-quality data with the overall goal to identify fish conditions and perform cage inspection during daily operations, as well as the robotic vision for an underwater vehicle. Algorithms were developed to estimate the distance and the orientation relative to the inspection object of interest. The autonomous navigation of underwater vehicles in fish cages is quite challenging. The research challenges are related to operational analysis of the level of automation and the development of bio-interactive control concepts for motion planning of underwater vehicles that allow inspection of the area of interest while avoiding collision with the infrastructure inside the cage and avoid 'scaring' the fish during operations (e.g. keeping a constant distance to the fish, slowing down when realizing a flight response ). Results for analysing autonomous operations and solutions for autonomous navigation of the vehicle inside the fish cage were obtained together with the technical specifications of the necessary equipment for sensing, localization, high-quality data capture and communication. A general bio-interactive control framework for autonomous operations in highly complex and dynamically changing environments was developed in reference to interaction with fish, deformable flexible structures and environmental disturbances. Extensive simulations were conducted to illustrate the performance of the proposed control framework and the efficacy of the control concept was investigated though field trials in a fish farm. For the underwater vehicle to be a permanent resident, a subsea docking station system for automatic launch and recovery, as well as inductive battery charging and transmission of the large data needs to be developed. In this project, a conceptual study was conducted to derive the requirements and specifications for such a docking system. The project results clearly show that underwater vehicles have great potential for high-quality data capture in fish cages, but should be investigated further to confirm their feasibility for commercial usage.

This collaboration between researchers, technology providers and a service company enabled a deep understanding of the challenges and benefits of using underwater technology for high-quality data capture for assessment of the fish farm state within three main areas: A) fish, B) infrastructure, and C) production environment. Furthermore, by developing new products and methods, the project contributes to improved HSE, fish handling/management and production efficiency/sustainability. The resulting fundamental knowledge and technology for autonomous operations in a dynamically changing environment are also applicable to other marine industries. By encouraging future use of autonomous underwater vehicles, the project outcomes will improve the level of control humans have over operations in aquaculture. As such, they contribute significantly to making aquaculture a safer and more sustainable environment for humans as well as the cultured organisms.

Prosjektet CageReporter har som overordnet idé å benytte autonome og trådløse farkoster, som bærer av sensorsystemer for datafangst, og hvor data overføres fra oppdrettsmerder til en sentral enhet og operatør. Farkosten vil benytte thrustere for aktiv bevegelsesstyring, og innhente data om tilstanden i merden mens den beveger seg i vannvolumet som er avgrenset av notposen, heretter kalt merdrommet. Farkosttypen blir ofte referert til som AUV (Autonomous Underwater Vehicle), og kjennetegnes ved at den i stor grad er selvgående, men at den også har interaksjon med operatør ved hjelp av trådløs undervannskommunikasjon. Prosjektet skal utvikle løsninger som gir farkosten autonom funksjonalitet for å kunne operere i samspill med biomassen (biointeraktiv) og anleggskonstruksjonene, og som i kombinasjon med sanntids kvalitetskontroll skal sørge for innhenting av høykvalitets data. En grunnleggende egenskap ved prosjektideen er bruk av residente farkoster, fast tilstedeværende i hver enkelt merd, for kontinuerlig datafangst. Prosjektideen vil basere seg på bruk av lavkostnadsteknologi for trådløs undervannskommunikasjon, posisjonering og kamerasystemer for 3D-syn. Prosjektet adresserer næringens mange utfordringer, knyttet til manglende nøyaktighet og kontroll over oppdrettssituasjonen, ved bruk av teknologi for å innhente høyoppløselige data i tid og rom som kan benyttes til å kvantifisere tilstanden i merden, gruppert innen hovedbruksområdene: A) FISKENS TILSTAND, B) ANLEGGSINSPEKSJON og C) PRODUKSJONSMILJØ. Eksempler på konkrete bruksområder er deteksjon av unormal fiskeadferd, notinspeksjon og kartlegging av vannkvalitet (produksjonsmiljøet). CageReporter vil sørge for kontinuerlig og tett oppfølging av tilstanden i merden, og være oppdretterens øyne i merdrommet. I prosjektet vil det utføres forskning innen 4 hovedområder: - Undervannskommunikasjon og posisjonsreferansesystem - Datafangst og sanntidsanalyse av datakvalitet - Autonome systemer - Undervannsdocking