Robots, AI, sensor networks and drones in aquaculture?

Precision feeding is making inroads in all livestock species and, while most advances have been made so far in terrestrial farming of chicken, dairy cattle and other livestock, farming of fish, shrimp and more is catching up fast. But how far has the use of precision feeding, analytics and optimising feeding practices come in aquaculture and what’s ahead?

Dr Francesca Antonucci and Dr Corrado Costa at the Council for Agricultural Research and Agricultural Economics Analysis (CREA) in Italy were pioneers of precision aquaculture, introducing the concept to the sector many years ago and actively researching it ever since.

Antonucci observes that particularly over the last 2 years, “this topic is attracting increasing interest within the aquacultural framework, and a review paper we published in 2019 has received increasing numbers of citations, making the work gradually more important from an application point of view.”

Precision feeding in aquaculture includes computer vision for animal monitoring, environmental monitoring tools such as sensor networks (wireless and long-range) and robotics, but also the analytics that take the sensor data and turn it into decision tools using Internet of Things (IoT) technology and more.

Putting technologies into practice

One example of how precision feeding systems are being explored, says Antonucci, is in salmon farming in coastal areas, where environmental regulations have become stricter in response to public pressure. A new paper published by a team from global engineering firm Ramboll and 2 partners shows how automated wireless sensor networks connected to different salmon farm data stations improves the sediment quality that results from discharge and excess feed.

In addition, researchers in Bangladesh have just implemented an IoT-based automated integrated rice-fish farming system that controls portable wireless sensor networks to remotely monitor environmental factors such as dissolved oxygen and turbidity. This reduces waste in feed distribution by allowing more efficient water management and more useful monitoring of growth rates to achieve higher production levels.

Antonucci also notes that, while robotics are now being widely used in many agricultural applications such as harvesting, seeding and farm animal monitoring, robotics is just starting to be explored in aquaculture, especially in the use of autonomous vehicles and IoT.

“One interesting application is found in a 2022 study by Kapetanović et al. in Croatia,” she says, “in which an autonomous surface vehicle and a remotely-operated underwater vehicle were used to collect data from both below the sea surface and from the air to analyse fish population modelling. The analyses of photo and video footage allowed non-invasive sampling of fish populations at high frequency, in real time, enabling the measurement of population density and the average size of individuals. The measurement data thus obtained are processed statistically and a model data interpretation of the growth of fish in the farm over time is created to estimate the biomass in the net pen allowing ad hoc feeding without waste and excesses.”

Productivity gains

Although research is ongoing, it is somewhat possible to predict at this point just how much precision feeding technologies can improve aquaculture productivity compared to the current productivity levels now achieved with various species – and also whether there are specific species where tech can make a bigger difference, and what technologies in particular help the most.

For her part, Antonucci points to specific non-invasive technologies that can boost productivity in the ‘big fish’ sector.

In her work with Costa, they have shown that computer image analysis can provide a non-invasive method for remote monitoring the size, shape, conformation and detailed movement of fish, which is important for measuring growth and other factors and also for determining the correct distribution of food. “Probably, in the big fish sector – that is, tuna,” says Antonucci, “this could lead to higher profits with respect to current productivity.”

The next 10 years

In the decade ahead, how much will precision feeding technologies be used around the world in various types of aquaculture? Will the use of simple precision feeding systems be common, with more advanced ones in place for farming ­specific aquaculture species?

It’s hard for anyone to predict the future, but Antonucci provides some observations about general trends. She says “the adoption of increasingly advanced technologies in the aquaculture sector, as well as in agriculture, will be directed towards preserving the environment. Reducing feed waste, not only from an economic but also from an environmental pollution point of view, will be the focal point of research for the next 10 years.”

“I think the shellfish mariculture sector, as well as those of the most-consumed fish species, will adopt increasingly autonomous and precise systems in the distribution of food, which is one of the most important parameters to monitor to optimise costs and consumption and to protect the environment.”

In their paper reviewing precision feeding in aquaculture, Fearghal O’Donncha at IBM Research-Ireland/National University of Ireland and Jon Grant at Dalhousie University in Canada (and NSERC-Cooke Industrial Research Chair in Sustainable Aquaculture) also put the focus on sustainability. “While the benefits of these advances in husbandry are apparent, their application to public facing indicators of sustainability is critical,” they explain. “The expansion of ‘big data’ in fish farming should have spin-offs for a larger conversation regarding indicators of sustainability in aquaculture.”

Sustainability headlined the acquisition in April 2022 by Denmark-based BioMar Group (a maker of high-performance aqua-diets for many farmed species) of AQ1, a company in Tasmania that produces acoustic and optical sensing feed technology for fish and shrimp farming. For 20 years, AQ1 has been developing systems that reduce feed conversion ratio, increase growth, minimise environmental outputs and improve the consistency of size and flesh characteristics at harvest. It has carried out research with almost 30 universities and companies. At the time of the acquisition, BioMar CEO Carlos Diaz explained that “in recent years, we have seen that the use of intelligent feeders in shrimp farming can optimise feed efficiency, production yield, and sustainability, when introduced alongside data-driven farm management.”

In terms of which specific technologies will expand in use, O’Donncha and Grant predict that sensor applications will grow.

“A wide variety of sensors are feasible, including optical, acoustic, and biological sensors for currents, particles, pathogens and harmful algal blooms,” they explain. “Moreover, a similarly-diverse array of image-based data are being applied to fish farming ranging from direct videography of fish to satellite remote sensing.” They add that the use of drones in data capture is an obvious application of airborne technology.

In addition, they note that the development of artificial intelligence to analyse images and interpret essential information related to fish behaviour and health is an active area of research.