Sistema de monitoreo con cerca virtual para el control del rebaño en la granja de ESPOL

Abstract

CONDITION FOR PUBLICATION OF PROJECT. project addresses the need to improve herd control on the ESPOL farm in the face of high costs and limitations of physical fences. The main objective is to design and implement a virtual fence system that allows to monitor in real time the movement of animals through wireless communication and generate alerts when detecting exits from the safe zone. The integration of low-consumption technologies is proposed, along with predictive models, optimizing cattle management and reducing losses. The system was developed using LoRa CubeCell nodes installed in collars, a LoRaWAN gateway connected to ChirpStack and the Home Assistant platform for visualization. The signal data was stored in an SQL database and machine learning algorithms, specifically LSTM and Random Forest, were applied to predict the position of the animals more accurately. The tests carried out on the farm showed that the light and portable prototypes did not affect the mobility of the cattle and allowed the delimitation of virtual grazing areas, issuing immediate notifications to the staff through the web interface and mobile devices. In conclusion, the system constitutes an efficient, scalable, and low-cost alternative that improves the safety of the herd, optimizes resources, and represents a technological advance applicable to the livestock sector. Keywords: Virtual Fence, Livestock, Real time Monitoring, LoRa.