New approaches to monitoring respiratory activity as part of an intelligent model for stress assessment

Abstract

Abnormal breathing patterns have been linked to many diseases and stress-related effect. Visually counting breaths per minute is the gold standard for measuring respiratory rate. In hospital research, most nurses recognize the physiological importance of respiratory rate however its measurement it is not considered mandatory. Current research studies offer viable options for continuous monitoring of respiratory activity, although with degraded performance due to artefact. This paper proposes five new respiratory rate estimation methods considering their strengths and drawbacks to determine the most suitable one for various activities. Photoplethysmography, accelerometry, infrared temperature and pressure sensors are therefore used to monitor respiratory activity. In addition, we present a method for estimating respiratory rate via thermographic video image processing. In terms of novelty and innovation, we highlight the intelligent algorithms developed for real-time respiratory rate extraction from Photoplethysmography signals, the mechanical sensor prototype based on pressure sensors, and the facial recognition, focus zone identification, and image pixel analysis algorithms for thermographic image processing. In addition, a multichannel sensing system characterized by distributed platform computation is utilized to extract physiological parameters forming the basis for the proposed Fuzzy Logic-based model to detect and classify stress levels. To validate the suggested approaches, an experimental protocol was established to monitor the volunteers’ respiratory activity in a controlled setting, as well as health monitoring throughout the induction of thermal stress and its classification, yielding excellent indications of efficiency and accuracy.