Wearable Sleep Detection System Based on Piezoelectric Signals and Convolutional Neural Network Analysis

Area of Science:

• Biomedical Engineering
• Wearable Technology
• Artificial Intelligence in Healthcare

Background:

• Accurate sleep monitoring is crucial for health management but current technologies lack comfort, accuracy, and direct physiological signal monitoring.
• Existing wearable sleep monitors often compromise on wearing comfort, data accuracy, or the ability to capture key physiological signals.

Purpose of the Study:

• To develop a comfortable, accurate, and durable wearable sleep monitoring system (SMS) for long-term health management.
• To assess sleep structure and quality by capturing eye movement signals using novel piezoelectric sensing technology.
• To provide actionable recommendations for improving sleep health through a user-friendly interface.

Main Methods:

• Developed a wearable sleep monitoring system (SMS) utilizing a piezoelectric sensing unit made from polyacrylonitrile (PAN) composite fiber film modified with PDA@ZnO.
• Employed a convolutional neural network (CNN) model for five-state sleep classification, trained and validated on extensive datasets.
• Validated the system's performance using 5-fold cross-validation and a 1000-sample held-out test set.

Main Results:

• The piezoelectric sensor demonstrated high sensitivity, comfort, and durability, enabling effective acquisition of physiological signals.
• The CNN model achieved a high accuracy of 95.90% and a Cohen's Kappa coefficient of 0.949 for five-state sleep classification.
• The system successfully provided sleep quality information via an intuitive human-computer interaction interface.

Conclusions:

• The proposed wearable sleep monitoring system offers a novel and feasible solution for accurate, long-term sleep health assessment.
• This technology enhances convenience in daily health monitoring by integrating comfortable sensing with advanced AI analysis.
• The system paves the way for improved sleep management and personalized health recommendations.