Automatic Multisource Data Collection System for Preterm Infants in NICU

Abstract

Preterm infants born before 37 weeks of pregnancy require continuous monitoring of vital parameters to ensure their well-being. The preterm babies are kept in the Neonatal Intensive Care Units for their constant monitoring. This system includes various sensors interfaced with an Arduino UNO board to provide nonstop monitoring for its major physiological parameters. The blood pressure is tracked by the pressure sensor, the body temperature is monitored by the thermistor, the breath by the sound sensor, and the movements of preterm infants are recorded by the accelerometer. They communicate with the Arduino from different sensor readings through a real-time LCD display such that the readings are monitored. The ESP32-CAM module provides video output for remote monitoring of premature infants, thereby increasing the amount of supervision from clinicians. There is also a USB-to-UART serial converter for feeding data into the AI machine-learning model on the system. The AI model performs analysis on the data collected by comparing with the previous data to observe the abnormality or distress condition for the preterm kids. Thus upon detection of critical situation, the system raises an email alert to the medical team to intervene as early as possible. This monitoring system increases the efficiency of the care given in the NICU by automating real-time data collection and analysis, as well as alerting cared clinicians. By fusing sensor technology with AI-embedded decision systems, it is a leading and dependable technology for enhanced neonatal healthcare outcomes while ameliorating the risks of premature complications.