Ultra-Low-Power Wearable Vibration Sensor with Highly Accurate Embedded Classifier.

Area of Science:

• Biomedical Engineering
• Wearable Technology
• Signal Processing

Background:

• Wearable sensors require reduced power consumption for extended use and smaller form factors.
• Real-time signal analysis via continuous wireless communication significantly drains power in wearable sensors.

Purpose of the Study:

• To propose an ultra-low-power wearable vibration sensor with an embedded signal classifier.
• To demonstrate the sensor's efficacy in accurate cough detection.

Main Methods:

• Development of a wearable vibration sensor featuring an embedded signal classifier.
• Implementation of a low-calculation-load algorithm for signal classification.
• Real-time cough detection using the developed sensor system.

Main Results:

• Achieved 93.0% accuracy in cough detection, surpassing conventional methods by 24.3%.
• Reduced average power consumption of the wearable sensor by 8.8 times.
• Demonstrated high accuracy and low computational load of the embedded signal classifier.

Conclusions:

• The proposed ultra-low-power wearable sensor with an embedded classifier offers a viable solution for continuous physiological monitoring.
• This technology enables automatic monitoring of cough symptoms in patients, improving healthcare accessibility.
• The sensor's low power consumption and high accuracy are critical for practical wearable health applications.