Non-Contact Respiratory Rate Estimation in Real-Time With Modified Joint Unscented Kalman Filter

Area of Science:

• Biomedical Engineering
• Signal Processing
• Health Monitoring

Background:

• Real-time respiratory rate (RR) monitoring is crucial for early detection of respiratory illnesses like COVID-19.
• Contact-based monitoring methods are unsuitable for continuous, long-term health surveillance, particularly in home-care settings.
• Wireless radio signals offer potential for non-contact health monitoring applications.

Purpose of the Study:

• To develop a device-free, real-time respiratory rate (RR) monitoring system utilizing wireless signals.
• To adapt and improve the standard joint unscented Kalman filter (JUKF) for accurate, time-critical RR estimation.
• To address the nonlinearities in RR estimation by proposing a novel measurement error transformation technique.

Main Methods:

• Modification of the standard joint unscented Kalman filter (JUKF) for real-time RR monitoring.
• Application of a hyperbolic tangent function to transform measurement errors into parameter errors, addressing nonlinearities.
• Experimental validation using real-time measurements from healthy volunteers.

Main Results:

• The proposed modified joint unscented Kalman filter (ModJUKF) demonstrated superior accuracy compared to windowing-based methods in time-varying RR scenarios.
• The ModJUKF method achieved a 36.7% improvement in accuracy over the standard JUKF.
• Computational complexity was reduced by approximately 8.54% using the ModJUKF method.

Conclusions:

• The ModJUKF offers a highly accurate and computationally efficient solution for non-contact, real-time respiratory rate monitoring.
• This technology holds significant promise for continuous health surveillance in home-care and nursing home environments.
• The novel error transformation technique effectively handles nonlinearities in wireless-based RR estimation.