Heartbeat detection with Doppler radar based on estimation of average R-R interval using Viterbi algorithm

Vital signs are related to mental stress conditions. In particular, the variation of R-R interval (the peak-to-peak interval of heartbeats) reflects mental stress. For measuring heart rate, non-contact detection using Doppler radar has been studied in many works. In this paper, we propose an algorithm to detect periodic heartbeats for some periods in which the R-R intervals are almost equal. In the long term, the R-R intervals gradually vary, however, the R-R intervals are almost equal within a short period. To detect those periods, we assume that the R-R intervals always have almost equal intervals. Then, we detect periodic peaks appearing at almost the equal intervals by using the Viterbi algorithm. After detecting periodic peaks, we detect the periods during which the R-R intervals are almost equal, based on the variance of the detected peak-to-peak intervals. Moreover, we estimate an average R-R interval and find peaks that appear at the interval nearly equal to the estimated average R-R interval within each detected period. Experimental results show that the proposed algorithm improves RMSEs (root mean square error) of the R-R interval compared with the conventional methods.