Reliable Respiratory Rate Estimation from a Bed Pressure Array

Abstract

Unobtrusive sleep monitoring allows older adults to have continuous monitoring during the night in their own homes. We propose a method to reliably estimate respiratory rate using a bed-based pressure sensor array. Movements are detected prior to respiratory rate estimation and suppressed. The amount of movement during an estimate and a weighting for the estimate are used to create a reliability metric. The reliability metric is scored out of 100 for each sensor where high scores denote more reliable data. Once respiratory rates were calculated, the mean reliability metric determined the estimate reliability. Nocturnal data from a male and female participant was analyzed. Results show better accuracy and validity than both analysis without movement suppression and analysis with movement suppression but without postprocessing data fusion. While more than 50% of estimates include movement corruption, only 15% are unreliable and, moreover, removal of unreliable estimates significantly reduces estimate variance and provides validity estimation.

A Pressure Sensitive Home Environment

“…The metric is based on the autocorrelation function's peak to valley difference and the percentage of samples in the 300-sample segment that are classified as rest. Details of the reliability metric can be found in [7]. The range of the reliability metric is between 0 and 100.…”

“…Respiratory estimates from each of the sensors were clustered and fused as proposed in [7]. This method is a weighted voted mechanism, with weightings based on the reliability metric.…”

“…This method is a weighted voted mechanism, with weightings based on the reliability metric. The validity of each respiratory rate estimate was determined from the reliability metrics of the estimate [7].…”

“…Specifically, we show how processing of bed-based pressure sensor data has been used to monitor an older adult during the night. Previously, we have shown methods of movement detection [6] and reliable respiratory rate estimation [7]. Here we further this research into a nocturnal monitoring system and visual environment.…”

Measuring chest movement using an array of unobstusive pressure sensors

“…Previous work used the fiber optic pressure sensor array for the analysis of bed transfer sequences [8], force and contact location estimation [9], [10], as well as for the measurement of breathing rate [11].…”