Desarrollo de un sistema inteligente de alerta temprana para monitoreo de variable críticas de salud con notificación vía mensajería en cabinas insonorizadas

Abstract

This early warning system aims to strengthen the safety and health of controlled environments, such as soundproof booths, by continuously monitoring critical data variables such as temperature, humidity, and oxygen levels in parts per million (PPM). The goal is to develop an electronic early warning system that integrates an open-source preventive program and sensors to collect and send information to a program that records, analyzes, and predicts a critical event. Therefore, a prototype was designed using Arduino, along with sensors such as the DHT22, MQ-135, and the HC-05 module. Python code was also designed with the TensorFlow library for prediction and alert sending to Telegram. Tests showed low margins of error: 8.98% for temperature, 2.27% for humidity, and 0.10% for oxygen levels, confirming the system's reliability in issuing real-time alerts. Furthermore, the hardware integration and Telegram notification setup were successful. The conclusion is that the prototype meets the proposed objectives, as a predictive model was designed that anticipates critical events with high accuracy.