Oyster Farming by Autonomous Surface Vehicles : An Implementation of Control and Perception Systems

With rising population and limited land, aquaculture is an increasingly important method of meeting worldwide seafood demand. A local New England aquaculture business, Ward Aquafarms, raises oysters in rigid plastic mesh bags that float on the surface of the water. These bags can weigh up to 27kg (60lb) and must be flipped over every 10-14 days to prevent biofouling from accumulating. This process is currently done manually, but there are several problems associated with this: ergonomic difficulty, potential for injury, and the lack of workers willing to do the job. A team at MIT Sea Grant created the Oystermaran in 2021 as a solution to this problem. While this was a promising proof-of-concept, the system was entirely remote-controlled and could only flip baskets in one direction which led to excessive twisting in the connecting ropes.This paper builds on the previous work by implementing several low level safety, perception, and control systems to enable autonomous operation. The 2023 Oystermaran is shown in Figure 1. We created an oyster basket localization algorithm with a custom motor controller. We demonstrate their successful integration via a basket-centering routine that uses visual feedback to move the boat to a target point relative to a basket between the hulls.A new flipping mechanism was designed to enable flipping baskets in both directions to avoid excessive twisting. During the testing of the mechanism, we discovered that it is essential to evenly redistribute the oysters within the bag after flipping in order to keep the basket correctly balanced. Future iterations of the Oystermaran should include a flipping mechanism that addresses re-balancing the bag and extend the autonomy system to traverse throughout the basket array.This paper demonstrates implementation and integration of key perception and control systems as steps toward full autonomy in oyster farming and aquaculture.