A Computational Respiration Factor to Detect Abnormal Respiratory Patterns Using a Hydraulic Bed Sensor for Older Adults Aging-in-place

Hydraulic bed sensors are efficient non-wearable, passive sensors for unobtrusive and continuous collection of health data for older adults aging-in-place. Continuous collection of data can speak volumes about onset and development of a disease much before it is diagnosed and that is our vision through this work. Hydraulic bed sensor data yields signal from which three major physiological components can be extracted for sleeping individuals, the ballistocardiogram component, the respiration component, and the bed restlessness component. In this work, we focus on the respiration component to detect any abnormal patterns in respiration, more specifically related to Chronic Obstructive Pulmonary Disorder (COPD). In that relation, we are proposing a computational passive bed sensor respiration factor (R factor). On the whole, we have developed a pipeline comprising a low pass Butterworth filter to extract the respiratory component from the raw signals, a respiration cycle identification algorithm, and an R factor feature extractor to detect abnormality in breathing patterns. We investigate the suitability of R factor for unobtrusive COPD detection in older adults through a case study comparing a COPD diagnosed individual and an individual with no respiratory disorder, both residents in long-term care facilities. We have discovered significant, detectable deviations in breathing patterns of the COPD diagnosed individuals compared to non-COPD breathing.

For Zoom information, contact Robert Sanders (sandersrl@missouri.edu).