Evaluation of RIP sensor calibration stability for daily estimation of lung volume

Abstract

This work supports the development of a non-invasive wearable system for the tracking of cigarette smoking in the natural environment. Accurate estimation of lung volume, as well as preservation of calibration over the period of wear, are required for such a system. In this paper, the preservation of calibration of a Respiratory Inductive Plethysmograph (RIP) in a 24-hour period was evaluated. Attachment options, such as an RIP belt held in place by tension (belt), RIP belt held in place by adhesive (belt+adhesive), and RIP belt sewn into a T-shirt (belt+shirt) were studied and compared using data collected from three subjects performing usual daily activities. An 800 ml bag was used as the gold standard for evaluation. The average volume obtained for the belt, belt+adhesive, and belt+shirt options were 0.89+0.28 L, 1.02+0.27 L, and 0.79+0.05 L, respectively. Additionally, comparing these against a reference volume of 800 ml, the average RMSE was 0.24+0.05 L, 0.24+0.23 L, and 0.03+0.03 L, respectively. Results show that the RIP belt sewn into a T-shirt provides a better preservation of the calibration throughout the 24-hour period.