Non-Contact Heart Rate Estimation in Low SNR Environments Using mmWave Radar

• Published in: 2022 United States National Committee of URSI National Radio Science Meeting (USNC-URSI NRSM) pp. 168 - 169
• Main Authors: Bauder, Chandler, Fathy, Aly E.
• Format: Conference Proceeding
• Language: English
• Published: USNC-URSI 04.01.2022
• Subjects: Data mining; Estimation; Heart rate; Radar; Radar detection; Radar signal processing; Signal processing algorithms
• DOI: 10.23919/USNC-URSINRSM57467.2022.9881404

Extracting accurate heart rate estimates of human subjects from a distance in high-noise scenarios using radar is a common problem. Often, frequency components from sources such as movement and vital signs from other subjects can overpower the weak reflected signal of the heart. In this study, we propose a signal processing scheme using an Adaptive Multi-Trace Carving algorithm (AMTC) to accurately detect the heart rate signal over time in non-ideal scenarios using a mm-wave radar. In our initial proof-of-concept results, we show a low heart rate estimation mean absolute error (MAE) of 3bpm for a single subject marching in place and less than 4.5bpm for a scenario of two human subjects at the same distance from the radar.