Matrix Pencil Method for Vital Sign Detection from Signals Acquired by Microwave Sensors.

Area of Science:

• Biomedical Engineering
• Signal Processing
• Microwave Technology

Background:

• Microwave sensors offer high-temporal resolution for contactless vital sign monitoring.
• Accurate and efficient signal processing is crucial for these sensors.
• Existing methods may struggle with noisy data common in real-world applications.

Purpose of the Study:

• To evaluate the Matrix Pencil Method (MPM) for extracting vital signs from microwave signals.
• To compare MPM's performance against Bandpass Filtering (BPF) and Variational Mode Decomposition (VMD).
• To assess the suitability of MPM for noisy, non-contact vital sign monitoring.

Main Methods:

• Application of the Matrix Pencil Method (MPM) to decompose microwave signals into damping exponentials.
• Utilizing the GUARDIAN dataset (continuous wave microwave sensor) and ultra-wideband (UWB) sensor data.
• Implementation and comparison with Bandpass Filtering (BPF) and Variational Mode Decomposition (VMD).

Main Results:

• MPM successfully separates vital signs like breathing and heartbeat with high precision.
• All tested methods (MPM, VMD, BPF) perform adequately with strong, pure signals.
• MPM demonstrates superior performance in noisy signal conditions compared to VMD and BPF.

Conclusions:

• MPM is a powerful signal processing technique for non-contact microwave vital sign monitoring.
• Its robustness in noisy environments makes it ideal for practical applications.
• MPM enables precise separation of physiological signals from complex microwave data.