Bio-Radar Cardiac Signal Model Used for HRV Assessment and Evaluation Using Adaptive Filtering

Abstract

The ability to remotely assess vital signs using radar systems may potentialize the healthcare systems, since it does not require any direct interaction with the subject. Nonetheless, these systems present inherent challenges that could compromise the accuracy of the vital signs parameters. For instance, there is a lack of resolution on the cardiac signal, since the radar only detects the chest micro-displacement. Low cardiac resolution prevents the exact localization of signal peaks, compromising the heart rate variability (HRV) assessment. In this work, we use a customized bandpass filter (BPF) obtained through adaptive filtering, applied to the electrocardiography (ECG) signal to generate the corresponding synthetic radar signal. This provides an easy way to extract a cardiac radar signal model, free of issues that arise from random body motion (RBM) or from an eventual decrease in signals quality, which are inherent issues of long-term acquisitions. This model was further used to verify if the signal resolution provided by lower carrier radars is suitable to determine the HRV parameters accurately.