Deep Learning-Enabled Flexible PVA/CNPs Hydrogel Film Sensor for Abdominal Respiration Monitoring

Area of Science:

• Materials Science
• Biomedical Engineering
• Sensor Technology

Background:

• Flexible sensors are crucial for wearable electronics and human-machine interfaces.
• Developing sensors with high sensitivity, fast response, and durability remains a challenge.
• Biomass-derived materials offer sustainable and scalable options for sensor fabrication.

Purpose of the Study:

• To develop a flexible hydrogel film sensor with enhanced sensing performance.
• To investigate the integration of biomass-derived carbon nanoparticles (CNPs) into a poly(vinyl alcohol) (PVA) matrix.
• To demonstrate the sensor's capability in capturing various physical signals and its application in breathing phase classification.

Main Methods:

• Fabrication of a flexible hydrogel film sensor by intermixing PVA and CNPs.
• Microstructuring the sensor surface using sandpaper templates for improved performance.
• Experimental validation of the sensor for detecting human joint movements, written letter signals, and object weight differences.
• Development of a breathing phase classification framework using a 1D-CNN algorithm.

Main Results:

• The sensor exhibited high sensitivity (101 kPa^-1), rapid response/recovery time (22 ms), and excellent durability (20,000 cycles).
• Accurate detection of human joint movements, electrical signals from written letters, and weight variations.
• Successful implementation of a breathing phase classification framework using 1D-CNN, demonstrating synergistic enhancement.
• The sensor showed potential for applications in wearable medical monitoring and haptic feedback systems.

Conclusions:

• The developed flexible hydrogel film sensor demonstrates excellent sensing performance and durability.
• The integration of biomass-derived CNPs and microstructuring enhances sensor capabilities.
• The sensor shows promise for advanced applications in wearable health monitoring and intelligent human-machine interfaces.
• This work highlights the synergy between scalable materials and deep learning for next-generation sensing technologies.