Design and Implementation of an Apparatus for Respiratory Parameters Estimation Based on Acoustic Methods

Breathing is one of the most fundamental characteristics in the continuation of life. Today, various contact and non-contact methods are used to monitor patients' breathing and diagnose lung diseases, each of which has its limitations. In this study, we designed a device to record respiratory sounds from the upper airways to evaluate the relationship between sounds and respiratory flow rate. Audio signals were recorded by microphones embedded in the breathing mask, and then amplitude and average power of the signal were extracted in different frequency bands. Using an artificial neural network, a correlation with a coefficient of 0.9 was observed between acoustic characteristics and respiratory parameters, including peak flow and average flow. This relationship can be used to extract the respiratory pattern, monitor personal health, and identify respiratory diseases. Some of the mentionable advantages of this method are reduction in user restrictions and price and also establishing a non-contact and non-invasive method.