Heartbeat: AI-Based Neonatal Cardiac Arrest Prediction Using Hybrid Deep Learning with Real-Time IoT Monitoring

Abstract

Cardiac arrest in newborn infants is a serious medical emergency requiring immediate attention. Early detection of such events can significantly improve survival rates and reduce complications. However, conventional monitoring systems are mostly reactive and depend on fixed threshold values, which limits their ability to provide early warnings. In this work, an intelligent monitoring framework is proposed that integrates Internet of Things (IoT) technology with a hybrid deep learning approach. The system continuously collects physiological parameters such as heart rate, oxygen saturation, and body temperature using embedded sensors. These measurements are transmitted through an ESP32 microcontroller to a cloud platform for real-time processing. A hybrid deep learning model combining convolutional and recurrent layers is employed to analyze time-series data and estimate the probability of cardiac arrest. The model is designed to capture both feature-level and temporal dependencies in physiological signals. The proposed system provides real-time alert mechanisms through both local and remote interfaces. The experimental results indicate that the model achieves high prediction accuracy and reliability compared to traditional approaches. This framework can be effectively deployed in neonatal care units to support timely clinical decision-making. The system is validated using real-time hardware implementation and IoT-based visualization.

Introduction

The text describes a smart neonatal monitoring system designed to detect and predict cardiac arrest in newborn infants using IoT and deep learning.

Core idea

Newborns in neonatal intensive care units are highly vulnerable, and traditional hospital monitoring systems rely on fixed threshold-based alerts, which often lead to delayed responses and false alarms. To overcome this, the paper proposes an AI-based predictive system for early detection of cardiac arrest.

Proposed solution

The system combines IoT sensors, cloud computing, and a hybrid deep learning model to enable real-time monitoring and prediction:

• Sensors measure heart rate, SpO?, and body temperature
• An ESP32 microcontroller collects and processes data
• Data is sent to the cloud for storage and analysis
• A deep learning model predicts risk of cardiac arrest
• Alerts are sent to medical staff when abnormal conditions are detected

System architecture

The system is structured into layers:

• Sensing layer: collects physiological data
• Processing layer: ESP32 handles preprocessing
• Communication layer: sends data via Wi-Fi to cloud
• Prediction layer: AI model analyzes data and predicts risk

Hardware implementation

• Uses MAX30105 sensor (heart rate, SpO?)
• Uses DS18B20 sensor (body temperature)
• ESP32 microcontroller manages data transmission
• LCD displays real-time values
• Buzzer provides local emergency alerts
• System is low-cost and tested in real-time

IoT and alert system

• Sends email notifications with AI-generated reports
• Includes both normal and emergency status updates
• Provides timestamps and vital parameters
• Supports remote monitoring by doctors and caregivers

Conclusion

The proposed system demonstrates practical feasibility through real-time hardware validation. The proposed Smart Cardiac AI Monitoring system successfully performs real-time health monitoring and detects critical conditions accurately. The system is simple, cost-effective, and suitable for continuous healthcare applications.

References

[1] E. Choi, A. Schuetz, W. F. Stewart, and J. Sun, “Using recurrent neural network models for early detection of heart failure onset,” Journal of the American Medical Informatics Association, vol. 24, no. 2, pp. 361–370, 2017. [2] K. W. Johnson et al., “Artificial intelligence in cardiology,” Journal of the American College of Cardiology, vol. 71, no. 23, pp. 2668–2679, 2018. [3] K. Gupta, N. Jiwani, G. Pau, and M. Alibakhshikenari,“A machine learning approach for early detection of cardiac arrest in newborns,” IEEE Access, 2023. [4] Espressif Systems, “ESP32 Datasheet,” 2022. [5] Maxim Integrated, “MAX30105 Pulse Oximeter Sensor Datasheet,” 2017. [6] Dallas Semiconductor, “DS18B20 Temperature Sensor Datasheet,” 2019.

Copyright

Copyright © 2026 Bhuvaneshwaran M N, Gowtham V, Sanjana D, Vincy A, Preethi A. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.