GEP-based predictive modeling of breathing resistances of wearing respirators on human body via sEMG and RSP sensors

Abstract

Purpose Breathing resistance is the main factor that influences the wearing comfort of respirators. This paper aims to demonstrate the feasibility of using the gene expression programming (GEP) for the purpose of predicting subjective perceptions of breathing resistances of wearing respirators via surface electromyography (sEMG) and respiratory signals (RSP) sensors. Design/methodology/approach The authors developed a physiological signal monitoring system with a specific garment. The inputs included seven physical measures extracted from (RSP) and (sEMG) signals. The output was the subjective index of breathing resistances of wearing respirators derived from the category partitioning-100 scale with proven levels of reliability and validity. The prediction model was developed and validated using data collected from 30 subjects and 24 test combinations (12 respirator conditions × 2 motion conditions). The subjects evaluated 24 conditions of breathing resistances in repeated measures fashion. Findings The results show that the GEP model can provide good prediction performance (R2 = 0.71, RMSE = 0.11). This study demonstrates that subjective perceptions of breathing resistance of wearing respirators on the human body can be predicted using the GEP via sEMG and RSP in real-time, at little cost, non-invasively and automatically. Originality/value This is the first paper suggesting that subjective perceptions of subjective breathing resistances can be predicted from sEMG and RSP sensors using a GEP model, which will remain helpful to the scientific community to start further human-centered research work and product development using wearable biosensors and evolutionary algorithms.