Heart Rate Measurement using a 60 GHz Pulsed Coherent Radar Sensor

Abstract: Heart rate is today measured in various ways, but they all include contact with the skin. Measuring heart rate contactless would be a more efficient method in healthcare and would also benefit people suffering from skin problems. It has been proven that it is possible to use radar to measure heart rate. The aim of this thesis is to investigate how to extract the heart rate with a 60 GHz pulsed coherent radar sensor. The heart rate was measured by aiming the radar sensor at the chest. This study included use of different radar configurations, distances to the participants and tilting angles of the radar sensor. An electrocardiogram was used as a reference signal during the collection of these measurements. During the preprocessing phase, two different approaches of removing the movement caused by respiration were tested: A bandpass filter and principal component analysis. Furthermore, two different methods for heart rate estimation were tested: One method based on detecting minima of the signal and the other method consisted of a recurrent neural network. During the evaluation, the heart rate provided by the sensor signal was compared to the heart rate of the ECG signal. The conclusion of this work was that the best measurement setup was achieved when measuring with a tilted sensor with a beam aimed at the participant’s chest. The sensor was placed in front of the patient at a distance of 0.4 meters. To remove the respiration, applying an adaptive bandpass filter to the unwrapped phase turned out to be the best solution. The minima detector got the most accurate result when comparing to the reference signal. The majority of heart rates estimated based on these steps had a small difference when compared to the reference signal.