A Wearable Multisensor Patch for Breathing Pattern Recognition

Abstract

In this paper, a multisensor patch is presented for the purpose of detecting and recognising the signals produced by human breathing in response to a variety of different body movements. We show that a multisensor patch consisting of an accelerometer and a pressure sensor can simultaneously measure breathing-related inertial motion and muscle stretch with a high degree of accuracy when it is attached close to the diaphragm. To construct the multisensor patch, we relied on commercially available off-the-shelf (COTS) electronic components that were relatively inexpensive. Different breathing motions were analyzed based on the accelerometer and the pressure sensor, including inhale, exhale, normal breathing, and breath hold conditions. The breathing frequency from the accelerometer and the flexible capacitive pressure sensors was found to be 0.2 Hz, and the normal breathing rate from the accelerometer and the pressure sensor was 11 breaths/min. We demonstrate that this new functional device and related approaches allow the identification of breathing patterns that are less cumbersome and tenably more reliable than conventional measures. The proposed multisensor patch holds great potential as a sensing technology in medical applications for the early detection of respiratory changes, one of the most predictive and earliest vital signs for worsening health. The presented methodology can be adapted for mass production of reasonably priced non-invasive breathing pattern detection and off-line analysis.