Clutter Effect on the Noncontact Seismocardiogram Signals Measured using Microwave Radars

Abstract

Seismocardiogram (SCG) signals, the measurements of local chest wall vibrations in response to the heartbeat, have been used to evaluate the mechanical aspects of cardiovascular health. A noncontact sensor based on microwave Doppler radars provides a potential approach to monitor such chest wall vibrations without attaching a sensor to the body. During detection, the chest wall displacement induced by the heartbeat will exhibit in the phase variation of the microwave signal, from which SCG signals can potentially be extracted and the cardiovascular characteristics can be analyzed. However, the microwave Doppler radar approach suffers from stationary clutter, which is the reflection of the background microwave signals. Radar clutter can decrease the ratio of effective signals in the receiver, weaken the effectiveness of the signal, and deteriorate the effective signal to noise ratio. An experiment is set up to evaluate the clutter effect on the SCG signals measured by the noncontact microwave radar sensor and variation of systolic time intervals (STIs) extracted from the noncontact SCG signals under physiological perturbations. The experimental results show that, with a high clutter level, the correlation and similarity in waveform shapes of the SCG signals measured by the noncontact sensor and the contact sensor are reduced, compared with the results when the clutter level is low. In addition, the changing direction of STIs may be influenced by the high level clutter. Therefore, it is important to design a clutter removal strategy to increase the measurement accuracy of the noncontact SCG waveforms so that the cardiovascular behavior can be effectively captured.