Millimeter-Wave Radar for Short-Term Heart Rate Measurement Using Intelligent Singular Value Decomposition Noise Reduction Algorithm

• Published in: IEEE transactions on instrumentation and measurement Vol. 73; pp. 1 - 13
• Main Authors: Lv, Wenjie, Liu, Chang, Liu, Haoqiang, Zhong, Shan, Yang, Hao, Feng, Shaohan
• Format: Journal Article
• Language: English
• Published: New York IEEE 2024; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Algorithms; Automatic multiscale peak detection (AMPD); Doppler radar; Heart beat; Heart rate; heart rate monitoring; K-means clustering; Low-pass filters; Millimeter wave radar; Millimeter waves; Monitoring; Noise; Noise measurement; Noise monitoring; Noise reduction; noise removal; Radar; Radar antennas; Radar detection; Singular value decomposition; singular value decomposition (SVD); Wave measurement
• ISSN: 0018-9456; 1557-9662
• DOI: 10.1109/TIM.2024.3436091

Heart rate measurement based on radar is an effective method for noncontact monitoring of human vital signs. However, the heart rate signal obtained from the radar is subject to noise interference due to imperfect circuit characteristics. To tackle this challenge in heart rate detection by micro-Doppler radar, this article uses a narrow-beam radar operating in the 120-GHz band to capture the heartbeat signal and adopts an intelligent singular value decomposition (SVD) algorithm to effectively eliminate noise in the heartbeat signal. This approach can adaptively select the number of large singular values corresponding to the best noise reduction effect under different signal-to-noise ratios (SNRs). In addition, an improved automatic multiscale peak detection (AMPD) algorithm is proposed for precise peak detection in the heartbeat signal. Experimental results demonstrate that the proposed algorithm significantly reduces omission or misjudgment during the peak detection process, leading to a high accuracy during short-term heart rate detection. Compared with the pulse wave measurement results, the average accuracy of the heart rate is between 96.53% and 98.18%.