Noncontact Sleeping Heartrate Monitoring Method Using Continuous-Wave Doppler Radar Based on the Difference Quadratic Sum Demodulation and Search Algorithm

Figures

• 

  Figure 1. Schematic diagram of the distribution of the four peaks in theory.

• 

  Figure 2. The flowchart of the whole algorithm process to extract HR and RR.

• 

  Figure 3. HR peak pair cells screened out from each group of spectrums. The color block above each cell is the average of cell frequencies, and the bottom is the frequency contained in the cell.

• 

  Figure 4. The device diagram and brief block diagram of doppler radar detection system.

• 

  Figure 5. Moving cardboard simulation experiment setup.

• 

  Figure 6. The moving board maintain a mixed sinusoidal motion with amplitude of 2 mm and 0.2 mm, and frequency of 0.2 Hz and 1.2 Hz, respectively: (a) the time domain diagrams of the two-channel data received by the radar; (b) the spectrum distribution diagrams of 15–30 s.

• 

  Figure 7. The setup for sleeping and illustration of heartrate belt.

• 

  Figure 8. The spectrum of 60 s data in two sleeping states: (a) the state of lying down, where the 0.47 Hz RR-related peak and 0.68 Hz and 1.13 Hz HR-related peaks can be found; (b) large-scale movements state, where only the 0.45 Hz RR-related peak is obvious and HR-related peaks are drowned out by movement noise (the ground truth of HR is 56 bpm, i.e., 0.93 Hz).

• 

  Figure 9. HR results in sleep state for about 6 h. The above panel depicts the HR and the true value results of HR, while the bottom panel depicts the respiration detection results.

• 

  Figure 10. HR results in sleep state for about 30 min (top), motion detection results (middle), and time-domain diagrams of two orthogonal channels data (bottom). There is a corresponding relationship between the time with a large amplitude change in the time domain and the time with a large error in the obtained result. The red box shows the signal region with a posture change from supine to side, whose amplitude is greatly reduced.

• 

  Figure 11. The cumulative distribution of relative error proportions for each group.