Remote sensing of patterns of cardiac activity on an ambulatory subject using millimeter-wave interferometry and statistical methods.

Using a 94-GHz millimeter-wave interferometer, we are able to calculate the relative displacement of an object. When aimed at the chest of a human subject, we measure the minute motions of the chest due to cardiac activity. After processing the data using a wavelet multiresolution decomposition, we are able to obtain a signal with peaks at heartbeat temporal locations. In order for these heartbeat temporal locations to be accurate, the reflected signal must not be very noisy. Since there is noise in all but the most ideal conditions, we created a statistical algorithm in order to compensate for unconfident temporal locations as computed by the wavelet transform. By analyzing the statistics of the peak locations, we fill in missing heartbeat temporal locations and eliminate superfluous ones. Along with this, we adapt the processing procedure to the current signal, as opposed to using the same method for all signals. With this method, we are able to find the heart rate of ambulatory subjects without any physical contact.