Input-Feature Correlated Asynchronous Analog to Information Converter for ECG Monitoring.

Area of Science:

• Biomedical Engineering
• Signal Processing
• Integrated Circuit Design

Background:

• Wearable physiological monitoring requires efficient data handling and ultra-low power consumption.
• Existing methods for processing signals like electrocardiogram (ECG) can be power-intensive and require significant storage.
• Direct feature extraction at the sensor level can reduce data transmission and processing burdens.

Purpose of the Study:

• To present a novel architectural design for an analog-to-information (A2I) converter.
• To enable ultra-low-power, mixed-signal very-large-scale integrated (VLSI) implementation for physiological signal monitoring.
• To demonstrate source compression and direct feature extraction capabilities for signals such as ECG.

Main Methods:

• Development of a variable input-feature correlated asynchronous sampling and time-encoded digitization approach.
• Design of a complete A2I converter architecture for mixed-signal VLSI.
• Implementation of case studies focusing on QRS detection in ECG signals using the A2I converter.

Main Results:

• Demonstrated significant source compression of ECG signals.
• Achieved over 98% efficiency in detecting QRS waves in challenging ECG waveforms.
• Validated the design's suitability for ultra-low-power and minimal storage requirements.

Conclusions:

• The proposed A2I converter architecture is effective for ultra-low-power wearable monitoring of physiological signals.
• The design facilitates efficient source compression and direct feature extraction, exemplified by high-accuracy ECG QRS detection.
• This approach offers a promising solution for long-term health monitoring applications with constrained resources.