IoT enabled health monitoring system using rider optimization algorithm and joint process estimation

Area of Science:

• Biomedical Engineering
• Sensor Technology
• Machine Learning for Healthcare

Background:

• Timely detection of abnormal health conditions is critical for effective medical intervention and improved patient outcomes.
• Existing real-time health monitoring systems often face limitations in achieving high accuracy, sensitivity, and specificity.
• There is a need for enhanced performance in real-time sensor data analysis for health monitoring.

Purpose of the Study:

• To develop an improved real-time health monitoring system utilizing sensor data.
• To propose and evaluate the Joint Process Estimator Rider Optimization Algorithm (JPEROA) for classification tasks in health monitoring.
• To compare the performance of the proposed JPEROA method against established machine learning algorithms.

Main Methods:

• Real-time physiological data (heart rate, blood oxygen, body temperature) collected using MAX 30102 and LM35 sensors.
• Data transmission and analysis via the ThingSpeak Internet of Things (IoT) cloud platform.
• Standardization of sensed features for uniform scaling.
• Implementation of the Joint Process Estimator Rider Optimization Algorithm (JPEROA) with Deep Stack Auto-encoder for classification, estimating line and delay coefficients.
• Comparative analysis with Support Vector Machine, Random Forest, Gradient Boosting, Naive Bayes, and Multilayer Perceptron.
• Validation using the PTB Diagnostic dataset.

Main Results:

• The proposed JPEROA method achieved a maximum accuracy of 0.9625.
• The system demonstrated a maximum sensitivity of 0.975 and a specificity of 0.95.
• JPEROA outperformed other evaluated machine learning algorithms in accuracy, sensitivity, and specificity.

Conclusions:

• The JPEROA algorithm, integrated with a Deep Stack Auto-encoder, offers a significant advancement in real-time health monitoring.
• The proposed system effectively enhances the accuracy, sensitivity, and specificity of detecting abnormal health conditions.
• This approach holds promise for improving patient outcomes through more reliable and precise health monitoring.