Multifunctional Human-Computer Interaction System Based on Deep Learning-Assisted Strain Sensing Array

Area of Science:

• Biomedical Engineering
• Sensor Technology
• Artificial Intelligence

Background:

• Continuous gait monitoring is vital for health management, including post-surgery recovery and disease diagnosis.
• Current gait analysis systems are often cumbersome, requiring specialized spaces and limiting real-world application.

Purpose of the Study:

• To develop an intelligent gait monitoring and analysis system using flexible piezoelectric sensors and deep learning.
• To enable real-time, accurate gait data acquisition and motion state inference for health monitoring.

Main Methods:

• Flexible piezoelectric sensors with high sensitivity and stability were developed and integrated into shoe soles.
• Deep learning neural networks were employed to analyze the acquired gait data.
• A flexible wearable recognition system with a human-computer interaction interface was constructed.

Main Results:

• The piezoelectric sensors demonstrated high sensitivity (241.29 mV/N), quick response, and excellent stability (R^2 = 0.9946).
• The integrated system achieved 94.7% accuracy in real-time human motion state recognition.
• The system successfully tracked athletes' gait for long-term monitoring.

Conclusions:

• The developed system offers a practical solution for continuous and reliable gait analysis in daily life.
• This technology has the potential to aid personalized health management, early disease detection, and remote medical care.