Sistema inteligente de monitoramento e automação para ambientes de animais domésticos com IOT

Abstract

The growing integration of pets as family members, coupled with the busy routines of owners, drives the search for technological solutions that ensure animal welfare. This work presents the design, validation, and evolution of a low-cost Internet of Things (IoT) based system for monitoring and automating pet environments. The prototype, validated in high-fidelity simulation, uses an ESP32 microcontroller to collect data from multiple sensors and control actuators, orchestrated by the visual tool Node-RED. As an evolution, and in response to a Software Engineering requirements analysis, the work implements two advanced extensions focused on result validation. First, a predictive health module was developed using Machine Learning (Isolation Forest). In controlled environment tests, this module achieved a Recall of 96.7% and an F1-Score of 89.2%, demonstrating superior efficacy compared to static methods in detecting behavioral anomalies. Second, the architecture was migrated to the AWS cloud (IoT Core and Lambda), where end-to-end integration tests validated the security and correct execution of voice commands via Amazon Alexa. The final objective was achieved by providing an integrated solution that ensures the animal’s feeding, hydration, and thermal comfort, confirming, through the obtained results, the technical feasibility of the proposed architecture for remote control and the provision of predictive insights into the pet’s well-being.