Non-Contact Detection of Vital Signs Based on Improved Adaptive EEMD Algorithm (July 2022)

Didi Xu, Weihua Yu*, Changjiang Deng, Zhongxia Simon He

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

12 Citations (Scopus)

Abstract

Non-contact vital sign detection technology has brought a more comfortable experience to the detection process of human respiratory and heartbeat signals. Ensemble empirical mode decomposition (EEMD) is a noise-assisted adaptive data analysis method which can be used to decompose the echo data of frequency modulated continuous wave (FMCW) radar and extract the heartbeat and respiratory signals. The key of EEMD is to add Gaussian white noise into the signal to overcome the mode aliasing problem caused by original empirical mode decomposition (EMD). Based on the characteristics of clutter and noise distribution in public places, this paper proposed a static clutter filtering method for eliminating ambient clutter and an improved EEMD method based on stable alpha noise distribution. The symmetrical alpha stable distribution is used to replace Gaussian distribution, and the improved EEMD is used for the separation of respiratory and heartbeat signals. The experimental results show that the static clutter filtering technology can effectively filter the surrounding static clutter and highlight the periodic moving targets. Within the detection range of 0.5 m~2.5 m, the improved EEMD method can better distinguish the heartbeat, respiration, and their harmonics, and accurately estimate the heart rate.

Original languageEnglish
Article number6423
JournalSensors
Volume22
Issue number17
DOIs
Publication statusPublished - Sept 2022

Keywords

  • ensemble empirical mode decomposition (EEMD)
  • frequency modulated continuous wave (FMCW)
  • non-contact vital signs detection
  • static clutter filtering

Fingerprint

Dive into the research topics of 'Non-Contact Detection of Vital Signs Based on Improved Adaptive EEMD Algorithm (July 2022)'. Together they form a unique fingerprint.

Cite this