Respiratory Rate Estimation during Walking and Running Using Breathing Sounds Recorded with a Microphone

Abstract

Emerging evidence suggests that respiratory frequency (fR) is a valid marker of physical effort. This has stimulated interest in developing devices that allow athletes and exercise practitioners to monitor this vital sign. The numerous technical challenges posed by breathing monitoring in sporting scenarios (e.g., motion artifacts) require careful consideration of the variety of sensors potentially suitable for this purpose. Despite being less prone to motion artifacts than other sensors (e.g., strain sensors), microphone sensors have received limited attention so far. This paper proposes the use of a microphone embedded in a facemask for estimating fR from breath sounds during walking and running. fR was estimated in the time domain as the time elapsed between consecutive exhalation events retrieved from breathing sounds every 30 s. Data were collected from ten healthy subjects (both males and females) at rest and during walking (at 3 km/h and 6 km/h) and running (at 9 km/h and 12 km/h) activities. The reference respiratory signal was recorded with an orifice flowmeter. The mean absolute error (MAE), the mean of differences (MOD), and the limits of agreements (LOAs) were computed separately for each condition. Relatively good agreement was found between the proposed system and the reference system, with MAE and MOD values increasing with the increase in exercise intensity and ambient noise up to a maximum of 3.8 bpm (breaths per minute) and −2.0 bpm, respectively, during running at 12 km/h. When considering all the conditions together, we found an MAE of 1.7 bpm and an MOD ± LOAs of −0.24 ± 5.07 bpm. These findings suggest that microphone sensors can be considered among the suitable options for estimating fR during exercise.

Advanced Textile-Based Wearable Biosensors for Healthcare Monitoring

“…This technology can offer superior durability and protective features while minimizing the impact on the fabric's visual and tactile aspect [98,99]. Romano et al [97] published a study in which they implanted a smart microphone sensor in a face mask to estimate the respiratory rate based on breath sounds Embedding technology entails incorporating the sensor component inside the fabric to safeguard the sensor via the fabric's structure. This technology can offer superior durability and protective features while minimizing the impact on the fabric's visual and tactile aspect [98,99].…”

“…This technology can offer superior durability and protective features while minimizing the impact on the fabric's visual and tactile aspect [98,99]. Romano et al [97] published a study in which they implanted a smart microphone sensor in a face mask to estimate the respiratory rate based on breath sounds during walking and running, as shown in Figure 2b. The signal-to-noise ratio of the breath signal was enhanced, and the impact of the external environment on motion detection was minimized.…”