Automatic pellet positioner with switched fuzzy-PD control for a smart shrimp feeder

The delivery of pellets from the surface to the bottom of a pond for shrimp feeding represents a challenge, particularly due to the critical transition between dry and wet phases. In this paper, a robotic platform is presented as a precision mechatronic actuator designed to execute feeding commands within a broader smart aquaculture framework. The system is based on a two-degree-of-freedom manipulator equipped with a winch-driven strap mechanism that enables controlled submersion and retrieval of a feeding plate. The platform exhibits nonlinear behaviour arising from variations in pellet mass, buoyancy, and hydrodynamic effects during phase transitions. While the proposed system does not perform biological feeding decision-making, it guarantees consistent and repeatable positioning performance across hybrid media interfaces. To this end, a supervisory switched fuzzy control strategy is proposed as the main contribution of this work. For benchmarking purposes, open-loop and conventional gain-scheduled PD controllers were implemented under identical conditions. Dynamic simulations and laboratory experiments incorporating plate elevation and varying load conditions were conducted to assess performance. Results indicate that, within the evaluated hyperparameter settings, the supervisory switched fuzzy strategy provides a stable and accurate response, particularly during phase transitions. The proposed control framework offers a practical solution for precision electromechanical actuators operating across hybrid air-water environments with minimal sensing requirements.