A medical radar system for non-contact vital sign monitoring and clinical performance evaluation in hospitalized older patients. (May 2022)

• Record Type: Journal Article
• Authors: Yen, Hoang Thi; Kurosawa, Masaki; Kirimoto, Tetsuo; Hakozaki, Yukiya; Matsui, Takemi; Sun, Guanghao
• Is Part Of: Biomedical signal processing and control. Volume 75(2022)
• Publication Date: 2022-05
• Subjects: 24 GHz radar -- Vital signs -- Signal quality -- Cheyne-Stokes respiration -- Sleep apnea; Signal processing -- Periodicals; Biomedical engineering -- Periodicals; Signal Processing, Computer-Assisted -- Periodicals; Image Processing, Computer-Assisted -- Periodicals; Biomedical Engineering -- Periodicals; 610.28
• DOI: 10.1016/j.bspc.2022.103597
• Languages: English
• ISSNs: 1746-8094

Abstract

Highlights: A novel medical radar system for non-contact vital sign monitoring. The signal quality assessment method was utilized to maintain high-quality radar signals in clinical practice setting. Proposed system enables early detection of abnormality of inpatients. Abstract: Medical radar plays an important role in non-contact vital sign monitoring. However, radar is sensitive and susceptible to noise that deteriorates signal quality, making its application unstable in medical practice settings. Previous studies have focused on detecting vital signs from radar output, but mainly used high-quality signals in laboratory settings. Therefore, signal quality evaluation to maintain good-quality signals for processing is crucial for radar applications. This paper presents a vital sign monitoring system using medical radar that automatically evaluates signal quality to eliminate unsatisfactory signals using the support-vector machine and to extract heartbeat and respiration signals using the singular value decomposition method. To evaluate the system performance, we tested it on 10 healthy subjects and compared the obtained vital signs with those measured using a contact-type ECG device and a respiration belt. Two evaluation steps were conducted. First, we measured 10 subjects for evaluating the signal quality in a laboratory setting. The results showed that the accuracy for each subject archive was 93.2%–100%. Second, we extracted vital signs including respiration rate (RR) and the heart rate (HR) following signal quality classification. The results showed RR and HR deviation from 0 to 1 bpm, and 0 to 4 bpm, respectively. Moreover, the system was applied in a real medical practice setting for continuously monitoring the bedside of hospitalized older patients at a hospital.