AI-Cardio: A Multi-Modal Explainable AI System for Heart Disease Detection, Prediction, and Personalized Healthcare Solutions

Abstract

Cardiovasculardisordersremainamajorglobal health concern, often requiring continuous evaluation of cardiac activity for early detection and timely medical intervention. This researchpresentsanintelligentheartmonitoringsystemthat integratesInternetofThings(IoT)technologywithartificial intelligence to support real-time cardiac assessment. The system utilizes an ESP32 microcontroller as the central processing unit connectedtoanECGsensorforcapturingelectricalheartsignals and a DHT11 sensor to record body-proximal temperature and ambient humidity. The acquired data is wirelessly transmitted to a cloud platform for persistent storage and remote accessibility. Toenhancediagnosticcapability,thesystemincorporatesa machinelearningmodeltrainedtorecognizeirregularheart patternssuchasarrhythmia,tachycardia,andbradycardia. Additionally, an AI-poweredEC Ggraphinterpretationmodule enablesuserstouploadwaveformimagesandreceiveauto- matedanalyticalinsights, improvingusabilityforindividuals withoutmedicalexpertise.ThishybridintegrationofIoTsensing andAI-assistedanalyticscreatesa cost-effective,scalable,and continuouslyoperatingmonitoringsolutionsuitableforhome healthcare, telemedicineapplications, andlong-termcardiac observation.Theproposedsystemdemonstratesitspotentialto improvepersonalhealthtracking, reduceclinical dependency, andfacilitateearlyinterventionincriticalcardiacevents.

Introduction

Overview:
Cardiovascular diseases often progress silently, making continuous cardiac monitoring essential for early detection. Traditional ECG tests and Holter monitors are hospital-bound, limiting long-term, routine monitoring, particularly for elderly or rural patients. The proposed system integrates an ESP32 microcontroller with ECG (AD8232) and environmental (DHT11) sensors to record heart signals, temperature, and humidity, transmitting data to a cloud platform for remote access and long-term logging. Machine learning and AI-assisted ECG image interpretation enable automated detection of arrhythmias and abnormal patterns, supporting timely intervention and preventive healthcare.

Objectives & Significance:

• Continuous real-time ECG monitoring with environmental context.

• Secure cloud storage for remote access and historical trend analysis.

• Automated detection of cardiac abnormalities using ML models (CNN, LSTM).

• AI-assisted ECG image interpretation for non-expert users.

• Reduced hospital visits, improved early detection, and cost-effective preventive healthcare.

System Architecture & Methodology:

• Core Components: ESP32 microcontroller, AD8232 ECG sensor, DHT11 sensor, cloud platform.

• Data Flow: Sensor data → ESP32 preprocessing → Wi-Fi transmission → Cloud logging → ML-based abnormality detection → AI-assisted ECG interpretation.

• Signal Processing: Digital filtering (IIR/FIR) removes noise; R-R interval analysis calculates heart rate.

• Machine Learning:

  • CNN extracts features from ECG signals for anomaly detection.

  • LSTM handles sequential ECG data to classify heart rhythms over time.

• Advantages: Real-time monitoring, low-cost and portable, AI-assisted diagnostics, cloud-based remote supervision, early detection of arrhythmias.

Results & Evaluation:

• The IoT-AI system outperforms traditional ECG methods in portability, continuous monitoring, cost efficiency, and remote accessibility.

• Real-time dashboards display heart rate, temperature, humidity, health scores, and alerts.

• AI-powered ECG analyzer allows upload and interpretation of ECG images for automatic detection of abnormalities.

• Users reported improved confidence in managing heart health, easy understanding of vitals, and reduced unnecessary hospital visits.

Limitations:

• Dependence on continuous Internet connectivity.

• Signal quality may be affected by motion artifacts or electrode misplacement.

• Limited biomedical parameters; AI interpretation requires clinical validation.

• Data privacy and regulatory compliance are critical challenges.

Future Work:

• Integration of additional biosensors (SpO?, blood pressure) for comprehensive monitoring.

• Enhanced deep learning models for improved diagnostic accuracy and risk prediction.

• Mobile and wearable hardware for fully portable continuous monitoring.

• Secure mobile app integration with real-time alerts.

• Large-scale deployment and interoperability with hospital systems for telemedicine applications.

Conclusion

The proposed IoT-AI heart monitoring system successfully demonstrates an efficient and accessible approach to continu- ous cardiovascular assessment. By integrating an ESP32 mi- crocontrollerwithECGandenvironmentalsensors,thesystem enables real-time vital data acquisition and cloud-based mon- itoring beyond hospital settings. The machine learning model and AI-powered ECG image analysis enhance diagnostic sup- port, allowing early detection of cardiac abnormalities evenforuserswithoutmedicalexpertise.Thesystemalsopromotes preventive healthcare through remote supervision and timely alerts, minimizing the dependency on frequent clinical visits. Overall, the solution proves to be cost-effective, scalable, and user-friendly, making it a suitable option for home healthcare, telemedicine, and long-term patient management. Future en- hancements such as additional biosensors, wearable design, and advanced analytics can further strengthen its clinical applicability and impact on smart healthcare ecosystems.

References

[1] Ansari, Y., et al. (2023). Deep learning for ECG arrhythmia detectionandclassification:Areview.BiomedicalSignalProcessingandControl. [2] Anwar,M.A.,etal.(2023).IoTbasedsystemforheartmonitoringand arrhythmia detection using machine learning. Journal of HealthcareEngineering. [3] Ayyub,A.,etal.(2025).AcomprehensivereviewofAI-baseddetectionand diagnosis using ECG signals. Computer Methods and Programs inBiomedicine. [4] Bai, X., et al. (2024). A hybrid deep learning network for automaticdiagnosis of cardiac arrhythmia based on 12-lead ECG. ScientificReports. [5] Gupta, R. K., et al. (2025). Design and implementation of portable ECGmonitoring system using AD8232 and ESP32. International Journal ofCreative Research Thoughts, 13(5). [6] Islam,M.J.,etal.(2024).ClassificationofECGsignalsbasedonMIT-BIH dataset using BLSTM model. Mathematical Modelling ofEngineering Problems. [7] Kumar, A., et al. (2022). IoT-based ECG monitoring for arrhythmiaclassificationusingoptimizeddeeplearningclassifier.BiomedicalSignalProcessing and Control. [8] Kumar, S. S., et al. (2023). An IoT framework for detecting cardiacarrhythmias in real-time. Biomedical Signal Processing and Control. [9] Ma, L., et al. (2024). A novel real-time detection and classificationmethod for ECG images based on Mamba-RAYOLO. Sensors, 24(16). [10] Malhotra, S., et al. (2025). AD8232 bioelectric signal processing withESP32 for inexpensive ECG monitoring. arXiv Preprint. [11] Patil, S. B., et al. (2023). A health monitoring system using IoT-basedAndroid mobile application. [12] Prajitha, C., & Velu, B. M. (2022). ECG diagnosis for arrhythmiadetection with a cloud-based RNN model. Frontiers in SustainableCities. [13] Qtaish, A., & Al-Shrouf, A. (2022). A portable IoT-cloud ECG moni-toring system for healthcare. International Journal of Computer Scienceand Network Security, 22(1), 269–276. [14] Rani, J. S., et al. (2025). Heart rate arrhythmia identification withInternet of Things devices. In Proceedings of the ACM InternationalConference on Smart Computing. [15] Reddy,K.R.,&Singh,P.K.(2023).ECGarrhythmiaclassificationusingdeep learning. International Journal of Advanced Research in Science,Communication and Technology. [16] Sharma, P. S., et al. (2024). IoT based low cost ECG machine and heartmonitoring system using ESP32 and Ubidots platform. InternationalJournal of Innovative Research in Computer Technology. [17] Vyas, H., et al. (2023). A portable IoT-based health monitoring frame-work using ESP32 for isolated and home-based patient care. [18] Verma, N., et al. (2025). Smart health monitoring system with ESP32:Real-timedata&cloudbasedreports.InternationalJournalofScientificResearch in Engineering and Management. [19] Xiao, Q., et al. (2023). Deep learning-based ECG arrhythmia classifica-tion: A systematic review. Applied Sciences, 13(8).

Copyright

Copyright © 2025 Dr. Kavitha Devi CS, Dr. Vidyarani H J, Yashaswi R, Sharath N S, Tejas S S, T V Aditya, Rahul Raj. 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.