Through the wall human heart beat detection using single channel CW radar

Figures

• 

  FIGURE 1. Representation of the theory of physiological sensing through the wall using CW Doppler Radar. Traditionally, a radar system transmits a signal, and the reflected signal is directly proportional to the small chest movements of the chest surface caused by cardiorespiratory activities.

• 

  FIGURE 2. Block diagram of the proposed system. Radar raw signal is captured both for single subjects and concurrent measurements of two subjects. Then using the MODWT method subharmonics suppression is performed to extract the fundamental frequency and then calculated its accuracy with the Biopac ECG acquisition module that acts as a reference system.

• 

  FIGURE 3. Experimental setup for human heart signal reception through the wall using Doppler Radar. (A) Single subject and brick wall (B) Single subject and wood wall (C) Two concurrent subjects and wood wall.

• 

  FIGURE 4. Radar-captured data when the subject was at 40 cm distarnce through the wall and was in seated position. Raw signal (A) from Radar (C) Biopac ECG acquisition module. Filtered signal (B) from Radar, and (D) from ECG.

• 

  FIGURE 5. Filtered waveform when the subject was at the distance of 40 cm from the radar through the wall. (A) Radar data, and (C) Biopac-ECG data. FFT of the radar filtered data (B) and ECG waveform filtered data (D). Both the filtered data and FFT of the filtered signal was drawn in time domain (E) and frequency domain (F).

• 

  FIGURE 6. Filtered heartbeat waveform from the radar (A) and Biopac ECG acquisition module (C) when the subject was at a distance of 100 cm through the wall. FFT of the radar-captured signal (B) and ECG waveform shown in (D). Both the time domain and frequency domain of radar-captured and ECG-captured signals are plotted together in (E) and (F). The FFT peak from the radar-captured signal can not be determined as the subharmonics are dominant.

• 

  FIGURE 7. The decomposed signal and its associated relative energy are shown here. The higher relative energy has been selected as the desired heartbeat signal from the Radar.

• 

  FIGURE 8. (A) Time domain signal of the MODWT-decomposed signal and ECG signal. (B) The FFT of the best decomposed signal was 1.43 Hz that closely matched with the FFT of the ECG signal.

• 

  FIGURE 9. Time domain signal of the mixture of two subjects at a distance of 40 cm through the wall (A) and its associated FFT (B). The ECG signal of the individual subject 1 (C) and subject 2 (E) and the FFT of the signal denoted that the heart rate of subject-1 was 1.657 Hz (D) and 1.47 Hz (F).

• 

  FIGURE 10. Time domain signal of the MODWT decomposed best two signals (A), (C) and its associated FFTs (B), (D). The FFT comparison between the best decomposed signals and ECG signal of subject 1 (E) and subject 2 (F) closely matches each other.

• 

  FIGURE 11. The comparative average accuracy of the direct FFT method and MODWT method at different distances of the subjects through the wall. MODWT and direct FFT methods perform quite closely up to distances of 40 cm and showed a similar accuracy of 96.84%. At a distance of 60 cm, the direct FFT method accuracy falls to 92.80% whereas the accuracy of the MODWT method is almost 95.25%. For the distance of 100 cm, the direct FFT method accuracy degrades to 86.06% and the accuracy of the MODWT method is still very high showing an accuracy of 93.45%.