IoT-Enabled Smart Healthcare System for Monitoring Patient Health and AI-Powered Anomaly Detection

Abstract

This is the concept of a smart health-care system combined with smart connected sensors with IoT and AI as an enabling mechanism for revolutionizing the way the health status of a patient is to be managed. Having multiple physiological parameters being monitored with this system by having a network of IoT sensors will comprise such things as heart rate, blood pressure, ECG, body temperature, and steps. The data, being collected through these diverse sensors, is processed in real-time-this means that data constantly flows from the individual nodes to the central hub with no pause. The sophisticated AI processing techniques then applied at the central hub scrutinize all the information sent to enable a more critical analysis on the information that was gathered. The role played by AI in the process is that it continues scanning the data for aberrations, any patterns or the presence of health disorders with the most advanced applications of Artificial Intelligence technology. These highly sophisticated algorithms possess the remarkable capability to analyze and identify various trends or tendencies that emerge from the comprehensive health patterns observed. Furthermore, they can play a crucial role in preventing the onset of severe health risks before they escalate into more serious issues. Since the system is targeted for the transfer of alerts to the patients\' healthcare providers in addition to the patients, it calls for prompt action in a manner that will aim to provide better wellness to the said patients. This new smart system of healthcare also avails a better mechanism for patient-centered health care, courtesy to its repeated ability to provide suggestions according to a patient\'s status of health towards providing patients with alert information of their respective sicknesses at the right and appropriate time. Besides the improvement of general health of patients, the system itself adds to the improvement and efficiency of various healthcare processing while simultaneously reducing and eliminating financial burdens associated with the need for additional inpatient services or otherwise complex laboratory tests which are required in most cases. Even more, the incredible ability of such a project to monitor the exact condition of patients in its real-time, combined with its ability to make relevant predictions about their health statuses, means that it has got the potential to revolutionize the healthcare field with an improved quality of treating patients, all while costs associated with the healthcare system decline effectively.

Introduction

1. Overview

This study presents an advanced smart healthcare system that combines IoT devices and AI algorithms to enable real-time patient monitoring and early anomaly detection. It addresses modern healthcare needs in areas such as remote care, chronic disease management, and elderly care by providing continuous tracking of vital signs and enabling timely medical intervention.

2. System Components

• IoT Devices: Wearables like smartwatches, rings, health patches, and biosensors track parameters like:

  • Heart rate

  • Blood pressure

  • SpO2 (oxygen level)

  • Body temperature

  • Glucose level

  • ECG signals

• Communication: Data is transmitted via Bluetooth, Wi-Fi, or cellular networks to a cloud-based platform, enabling remote access and scalability.

• Edge Computing: Pre-processes data locally to reduce latency and cloud load. Filters noise and extracts key insights for faster anomaly detection.

3. AI-Powered Anomaly Detection

• Algorithms (e.g., RNNs, CNNs, LSTMs) analyze real-time health data to identify abnormal patterns like:

  • Arrhythmias (irregular heart rhythms)

  • Hypoxia (low oxygen levels)

  • Fever or infections (abnormal temperatures)

• Benefits:

  • Enables early diagnosis and reduces emergency cases.

  • Reduces hospital visits.

  • Enhances healthcare in underserved and rural areas.

4. Integration with Cloud and EMRs

• Cloud platforms (e.g., AWS IoT, Azure IoT) store, analyze, and visualize patient data.

• Provides historical trend analysis, dashboards, and alerts to doctors, caregivers, and emergency services.

• Enables real-time decision-making and personalized care.

5. Proposed Framework

• Data Flow:

  1. Data Collection: From IoT devices and in-room sensors.

  2. Edge Processing: Noise filtering and preliminary analysis.

  3. Cloud Processing: AI algorithms perform anomaly detection and generate alerts.

  4. Alert System: Notifications are sent via SMS, apps, or automated calls.

• Communication Protocols: MQTT, CoAP, and HTTP ensure low-resource, reliable data transmission.

6. AI Algorithm: RNN Model

• Used for sequential health data processing.

• Steps:

  • Normalize data.

  • Train the RNN with time-series input.

  • Update weights via backpropagation.

  • Detect health anomalies based on learned patterns.

• Applications:

  • Stroke prediction

  • Blood-brain barrier analysis

  • Medication delivery systems

7. Experimental Results

• Simulation tools and ML frameworks (e.g., ML.NET) were used for testing.

• Accuracy Comparison:

  • ANN: 95.45% (Best performance for stroke prediction)

  • MLP: 91.20%

  • RNN: 89.30%

  • CNN: 75.67%

• Optimal Dataset Size:

  • Best performance achieved with 70–80% training data.

  • Lower leads to underfitting; higher may cause overfitting.

• Performance Metrics:

  • Accuracy, Precision, Recall, F1 Score

  • ANN outperformed others in all key metrics.

Conclusion

Among all the crucial and life-threatening health disorders globally, stroke disease has remained at the top of concern and emphasizes the need for early detection along with timely diagnosis to reduce the risk factors and improve patient outcomes. The paper introduces a Data Transmission-based IoT-enabled healthcare system that has been developed for continuous monitoring, real-time data acquisition, and analysis of physiological signals for the effective prediction of stroke-related disorders. The system efficiently utilizes Internet of Things (IoT) sensors that are tailored for collecting and gathering all essential health parameters that are essential in monitoring well-being. The parameters include significant metrics like heart rate, blood pressure, body temperature, and levels of blood oxygen (SpO2). These vital signs, which play an important role in assessing an individual\'s health, are monitored continuously over time and transmitted in real-time to the designated processing unit where they undergo further detailed analysis. It is possible to predict critical risk factors of stroke patients using deep learning algorithms, such as ANN, CNN, and RNN, even in emergency conditions. The results of the investigation prove that the proposed model is superior to the existing one. In other words, ANN models achieved 93.21% accuracy, which means they work well with producing accurate predictions. Moreover, a training dataset size between 75% and 80% is enough to provide reliable results without any underfitting or overfitting risk and ensures the high performance of models. This system can detect symptoms of health disorders way before the rest due to inclusion of the latest Data Transmission technology accompanied by deep learning. This would not only give the system a high level of accuracy and precision in its predictions of diseases but also enables intervention promptly and before time. That is why it is extremely helpful in healthcare systems and more so when patients are handled promptly with issues related to stroke disorders.

References

[1] G Prasad, P Gujjar, HNN Kumar, MA Kumar… - Cyber Trafficking, Threat Behavior, and Malicious …, 2023 [2] Imoize, Agbotiname Lucky, et al., eds. Handbook of Security and Privacy of AI-Enabled Healthcare Systems and Internet of Medical Things. CRC Press, 2023. [3] Kalusivalingam, Aravind Kumar, et al. \"Enhancing Patient Care Through IoT-Enabled Remote Monitoring and AI-Driven Virtual Health Assistants: Implementing Machine Learning Algorithms and Natural Language Processing.\" International Journal of AI and ML 2.3 (2021). [4] Tiwari, Meena, and Akhilesh A. Waoo. \"Transforming Healthcare: The Synergistic Fusion of AI and IoT for Intelligent, Personalized Well-Being.\" Revolutionizing Healthcare: AI Integration with IoT for Enhanced Patient Outcomes. Cham: Springer Nature Switzerland, 2024. 109-149. [5] Bhambri, Pankaj, and Alex Khang. \"Managing and Monitoring Patient\'s Healthcare Using AI and IoT Technologies.\" In Driving Smart Medical Diagnosis Through AI-Powered Technologies and Applications, pp. 1-23. IGI Global, 2024. [6] SEDDIKI, M., BOUHEDDA, M., TOBBAL, A., REBOUH, S., & ANIBA, Y. (2024). AI-Driven IoT Healthcare System for Real-Time ECG Analysis and Comprehensive Patient Monitoring. Science, Engineering and Technology, 4(2), 61-81. [7] Bajpayi, P., Sharma, S., & Gaur, M. S. (2024, March). AI Driven IoT Healthcare Devices Security Vulnerability Management. In 2024 2nd International Conference on Disruptive Technologies (ICDT) (pp. 366-373). IEEE. [8] Kumar A, Jain R, Gupta M, Islam SM, editors. 6G-enabled IoT and AI for smart healthcare: Challenges, impact, and analysis. CRC Press; 2023 Jun 21. [9] Awad, Ali Ismail, et al. \"AI-powered biometrics for Internet of Things security: A review and future vision.\" Journal of Information Security and Applications 82 (2024): 103748. [10] David, Jameson. \"Revolutionizing Biomedical Engineering: AI-Driven Medical Devices in Cloud-Based Healthcare Systems.\" (2024). [11] Rangasamy, Sangeetha, et al. \"Revolutionizing Healthcare with IoT: Connecting the Dots for Better Patient Outcomes.\" AI and IoT Technology and Applications for Smart Healthcare Systems. Auerbach Publications, 2024. 204-223. [12] Daniel, Thomas. \"Cloud Computing and Data Security Strategies in AI-Driven Healthcare Devices.\" (2024).

Copyright

Copyright © 2025 Ms. Neha Arora, Umme Rumana, Laxmi S Dyapanagoudra, Hassain Miyan, Aakif Khan. 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.