Diseño y simulación de un prototipo de alimentación automática para camarones de cultivo usando Teachable Machine

Abstract

The main objective of this Degree Work is the simulation of an automatic feeding prototype for farmed shrimp, based on the detection of audios generated by the shrimp at the time of feeding. To achieve this purpose, the Teachable Machine tool is used, which allows artificial intelligence models to be trained to recognize specific acoustic patterns. The proposed system uses a sensor called a hydrophone that is responsible for capturing the sounds produced by the shrimp during feeding and these sounds are processed by the Teachable Machine software, which is able to distinguish between the times when the shrimp are eating and the times when they are not. Once the corresponding audio is detected, the system activates an automatic feeding mechanism that releases the adequate amount of food. The simulation of the prototype demonstrates that the use of artificial intelligence and audio analysis can offer an innovative and efficient solution for the aquaculture sector, allowing for more accurate and automated feeding, which has significant potential to improve productivity in shrimp farming.