An ECG data sampling method for home-use IoT ECG monitor system optimization based on brick-up metaheuristic...

Area of Science:

• Biomedical Engineering
• Computer Science
• Data Science

Background:

• The increasing use of Internet of Things (IoT) for in-home health monitoring, particularly for Electrocardiogram (ECG) data, escalates server-side data processing demands.
• Limited data transmission and storage capacity in home-use IoT systems necessitate efficient data handling for real-time analysis.
• Traditional ECG sampling rates (100-500Hz) are often reduced in home settings (e.g., 50Hz), leading to large time-series datasets requiring effective downsampling.

Purpose of the Study:

• To investigate efficient downsampling methods for ECG data in IoT health monitoring.
• To propose and evaluate a novel Brick-up Metaheuristic Optimization Algorithm (BMOA) for adaptive ECG data sampling.
• To reduce data transformation and uploading times, thereby saving costs in home-use health monitoring systems.

Main Methods:

• Developed a novel Brick-up Metaheuristic Optimization Algorithm (BMOA) for automatic and adaptive optimization of ECG data sampling.
• BMOA dynamically selects optimal components in real-time to create a tailored metaheuristic algorithm for individual users and ECG data series.
• Simulated various application scenarios using real ECG datasets and evaluated performance with a Long Short-Term Memory (LSTM) Network for classification.

Main Results:

• The BMOA demonstrated adaptive ECG data sampling capabilities across different users and scenarios.
• Classification efficiency using the Long Short-Term Memory Network was improved with BMOA-sampled data.
• Significant reduction in data storage requirements was achieved through the proposed adaptive downsampling method.

Conclusions:

• BMOA offers an effective and adaptive solution for downsampling ECG data in IoT health monitoring.
• The proposed method enhances the efficiency of ECG data analysis and classification while minimizing storage needs.
• This dynamic pre-processing approach is crucial for cost-effective and responsive in-home health monitoring systems.