Data binning method and system for estimating respiratory airflow from body sound signal

Abstract

Data binning methods and systems for estimating a subject's respiratory airflow from a body sound signal detected by an acoustic sensor on the subject's body. The methods and systems operate in a configuration mode followed by a monitoring mode. In the configuration mode, a body sound signal and respiratory airflow are detected by an on-body acoustic sensor and a spirometer, respectively, over a common time period. Time-aligned body sound signal and respiratory airflow data points are then generated and assigned to bins each spanning a discrete signal range (e.g. discrete signal entropy range or signal amplitude range). Respiratory airflow estimation data (e.g. mean airflow and standard deviation) are then calculated for each bin and an entry for each bin associating the discrete range and the estimation data is stored in a lookup table. Then, in the monitoring mode, the lookup table is accessed using subsequent body sound signal readings (e.g. taken in the field or at home) to provide respiratory airflow estimates without further need for a spirometer.