Design and Testing of an Autonomous Cable Mounted Aquaculture Robot

Abstract

The ocean covers more than 70% of the Earth’s surface, with land suitable for human purposes, especially farming, making up far less than the remaining 30%. Raising livestock takes up more than 26% of the available land on the Earth’s surface and is one of the most popular, and one of the most carbon and freshwater intensive, sources of protein available (NOAA, 2020). As climate change continues to advance, and the world’s population continues to grow, so does the need for protein, and aquaculture presents an ecofriendly way to feed the world. Marine aquaculture in particular has a much smaller carbon footprint, requires less land, and uses much less fresh water than traditional farming (NOAA, 2020). In many parts of the world, the growth of aquaculture industries has already aided in alleviating food insecurity, as well as providing jobs, and based on the World Bank Report, aquaculture will continue to grow rapidly and is expected to supply more than 60% of fish destined for human consumption by 2030 (World Bank, 2013). Given the importance of the industry and its rapid growth, there is a lot of space for improvements in methodology and technology to reduce cost and increase overall yield. Overfeeding in aquaculture is one of the major causes of pollution in the industry and also leads to disease, with loss of livestock and wasted feed (which increases costs). Since excess nutrients are released into the outside environment, they can also pollute the surrounding area. Feed waste can account for 25-75% of a farm’s feed consumption, depending on the farming method (Lee, 2001). Increased feeding frequency, with less food being delivered per feed, has been shown to decrease this food waste. However, since feeding fish is labor intensive and expensive when done by hand, feeding frequency is often difficult to increase and dependent on the size of the farm, availability of labor, and species, with feeding frequencies ranging from once or twice per day to up to five times per day. Generally, however, higher feeding frequencies greatly increase the rate of growth for many species, while also reducing overfeeding (Lee, 2001). With feed also making up around 60% of production costs in aquaculture, reducing waste of excess food can lower costs significantly, as well as eliminate the danger of overfeeding related disease and stock loss. Automatic and on-demand feeders enable farmers to tune feed to consumption and have been effectively utilized to reduce labor costs as well as overfeeding, but they must be programmed and installed in every tank. The installation costs of these systems can be a financial barrier for many farmers. Our proposed robotic feeding system can meet the unique needs of a variety of aquatic animals while minimizing the labor required by farmers and the risk of overfeeding. The system traverses a fixed cable hung above pens or tanks and dispenses fish food according to preinstalled feeding instructions or by determining animal activity using a camera-based artificial intelligence system. Attaching the feed system to a cable suspended over the fish pens allows one system to feed multiple pens/tanks, and to disperse feed evenly throughout the tank with only a single unit. It also provides the opportunity for the system to refill itself automatically from a large, centrally located reservoir, enabling it to work autonomously for extended periods of time. This significantly reduces the labor and time investment needed in aquaculture operations, as well as allowing for optimal feed deployment and frequency with no increase in human labor. The addition of a camera and vision system to determine when the fish or other animals stop responding to feed (i.e. have been fed enough) decreases the amount of waste feed dispensed, and ensures that water quality is not compromised. This paper explores the design and testing of the robotic system, applications for its use, and recommends directions for future development.