Smart Mat for Respiratory Activity Detection: Study in a Clinical Setting

Abstract

We discuss in this paper a study of a smart and unobtrusive mattress in a clinical setting on a population with cardiorespiratory problems. Up to recently, the vast majority of studies with unobtrusive sensors are done with healthy populations. The unobtrusive monitoring of the Respiratory Rate (RR) is essential for proposing better diagnoses. Thus, new industrial and research activity on smart mattresses is targeting respiratory rate in an Internet-of-Things (IoT) context. In our work, we are interested in the performances of a microbend fiber optic sensor (FOS) mattress on 81 subjects admitted in the Cardiac Intensive Care Unit (CICU) by estimating the RR from their ballistocardiograms (BCG). Our study proposes a new RR estimator, based on harmonic plus noise models (HNM) and compares it with known estimators such as MODWT and CLIE. The goal is to examine, using a more representative and bigger dataset, the performances of these methods and of the smart mattress in general. Results of applying these three estimators on the BCG show that MODWT is more accurate with an average mean absolute error (MAE) of 1.97 ± 2.12 BPM. However, the HNM estimator has space for improvements with estimation errors of 2.91 ± 4.07 BPM. The smart mattress works well within a standard RR range of 10-20 breaths-per-minute (BPM) but gets less accurate with a bigger range of estimation. These results highlight the need to test these sensors in much more realistic contexts.

Recent Advances and Applications of Textile Technology in Patient Monitoring

“…Figure 2a highlights the MAEs of four out of the nine articles on respiratory rate estimation using smart mats or bedsheets [50,55,58,59]. Similarly, 3 of the 12 heart rate articles investigated the MAEs of their smart mats or bed sheets [50,51,54], shown in Figure 2b.…”

“…The black dashed lines represent the average MAE and number of subjects of all included studies. It should be noted that the vertical black dashed line in Figure 2a is skewed to the higher end of the number of subjects due to the large sample size in [59]. The gray region shows the region of interest, which has a large sample size and low MAE values.…”

“…Though [50] did not fall within the gray region of Figure 2a, it was the closest study to this region due to the low MAE value of 0.38 breaths per minute and the second highest sample size of 10 subjects. Meanwhile, [59] consisted of the largest sample size with 80 subjects but also achieved the highest MAE of 4.57 breaths per minute. The minimum MAE relating to the heart rate was achieved in [50], with an MAE value of 0.55 beats per minute and the largest sample size of 10 subjects, falling within the gray region of Figure 2b.…”

“…The most common detection method was peak detection, which identifies the peaks of the signal to determine each breath or beat [15,50,[56][57][58]. Maximal overlap discrete wavelet transform (MODWT), wavelet analysis, cepstrum, clustering, and empirical mode of decomposition (EMD) are also popular detection algorithms for detecting the signals of the heart rate and respiratory rate, which can be seen in [12,14,51,59]. Peng et al [14] compared the wavelet analysis and EMD for estimating the respiratory rate and heart rate.…”