High Accuracy Respiration and Heart Rate Detection Based on Artificial Neural Network Regression.

Area of Science:

• Biomedical Engineering
• Signal Processing
• Artificial Intelligence

Background:

• Contactless monitoring of vital signs is crucial for remote healthcare.
• Existing methods may have limitations in accuracy or convenience.

Purpose of the Study:

• To develop and validate a 24GHz Doppler radar system for accurate, non-contact measurement of heart and respiratory rates.
• To evaluate the performance of a combined Convolutional Neural Network (CNN) and Long Short-Term Memory (LSTM) neural network for vital sign regression analysis.

Main Methods:

• A 24GHz Doppler radar system was utilized for data acquisition.
• A CNN+LSTM neural network architecture was employed for regression analysis of radar signals.
• Detection accuracies for heart rate and respiration rate were calculated.

Main Results:

• The system achieved high detection accuracies: 99% for heart rate and 98% for respiration rate.
• The CNN+LSTM model demonstrated effective regression analysis for contactless vital sign monitoring.
• The radar system proved capable of precise vital sign detection.

Conclusions:

• The developed 24GHz Doppler radar system offers accurate, non-contact vital sign monitoring.
• This technology has significant potential for heart rate variability warning systems and convenient home healthcare.
• The system enables fast and reliable monitoring for daily care applications.