Respiration and heartbeat signal separation algorithm using UWB radar platform

The human chest wall fretting signal detected by ultra-wideband radar combines chest wall periodic motion caused by breathing and heartbeat. And the frequency and the amplitude of the respiration signal change at any time. We use the empirical model that chest wall motion caused by respiration movement changes sinusoidally to improve the separation effect of respiration and heartbeat waveforms. Based on the hypothesis, we propose an algorithm named adaptive sine wave fitting combined with the baseline drift elimination algorithm to eliminate the respiratory and its high-order harmonic in the time domain. Through our algorithm, the heartbeat spectrum is separable. And the average error of heart rate reduces to 1.586%, which is 24.770% of the high pass filter (HF) method and 13.188% of the CEEMD method. There is also some improvement in the matching effect between the heartbeat and the ECG waveform. Furthermore, we verify the feasibility and accuracy of the algorithm through different breathing patterns. And we find that the SWF algorithm performs more stable in the chest and abdominal breathing mode than the HF method and the CEEMD method.