IEEE Access (Jan 2024)

Enhancing the Response of a Wearable Sensor for Improved Respiratory Rate (RR) Monitoring

  • A. Ali,
  • Y. Wei,
  • J. Tyson,
  • Harry Akerman,
  • A. I. R. Jackson,
  • R. Lane,
  • D. Spencer,
  • Neil M. White

DOI
https://doi.org/10.1109/ACCESS.2024.3509676
Journal volume & issue
Vol. 12
pp. 180913 – 180925

Abstract

Read online

Currently available devices for monitoring respiratory rate (RR) are cumbersome and expensive (such as capnography and plethysmography), requiring skilled clinicians to operate and analyse. In contrast, the inexpensive and lightweight ones (e.g. reported strain, stretchable belt, and capacitive-based RR sensors) are unable to provide accurate RR measurements due to being highly affected by motion and environmental noise. This study aims to develop a wearable RR sensor with a high sensitivity and accuracy (comparable with capnography) of RR and which is not adversely affected by noise in comparison to other reported sensors in the literature. We have developed a flexible textile-based wearable RR sensor that can be conveniently embedded in a patient’s gown. A design rule was adopted to empirically find the ratio of sensor electrodes with the aim to achieve the highest percentage output frequency change (%F.C). The empirical study also suggests that the %F.C can be further increased by increasing the dielectric material thickness between the sensor and ground electrodes. The enhanced textile-based sensor shows a 10.5% F.C toward the changes in the dielectric constant of a phantom, representing the human chest. The subsequent tests of the proposed RR sensor attached to the torso of the test subject shows a high-frequency variation (43 kHz) between crests (which corresponds to inhaling) and troughs (which corresponds to exhaling), with a 99.39% accuracy compared to the reported strain (a stretchable belt) and capacitive-based RR sensors. In the 35 performed RR tests, the proposed RR sensor picked up 495 crests; during the same tests, the capnograph showed a total of 498 respirations (which leads to 99.39% accuracy). This work demonstrates that the investigation in the electrode ratio and increase of dielectric layer thickness increases the proposed wearable textile base sensor %F.C, which leads to highly reliable (99.39% accuracy) RR detection with high-frequency variations and high SNR, as compared to gold standard capnograph, which is currently used in hospital for respiratory rate monitoring.

Keywords