Improving Tidal Volume Estimation via Fusion of Impedance Pneumography and Accelerometry

• Published in: 2023 IEEE SENSORS pp. 1 - 4
• Main Authors: Berkebile, John A., Sanchez-Perez, Jesus Antonio, Ozmen, Goktug C., Inan, Omer T.
• Format: Conference Proceeding
• Language: English
• Published: IEEE 29.10.2023
• Subjects: Calibration; Estimation; impedance pneumography; multimodal fusion; Pulmonary diseases; Quality assessment; respiratory sensing; tidal volume; Volume measurement; Wearable computers; wearable sensing
• ISSN: 2168-9229
• DOI: 10.1109/SENSORS56945.2023.10324945

Accurate estimation of tidal volume (TV) is critically important for assessing respiratory diseases. The specialized equipment typically used to acquire TV is obtrusive and difficult to translate to monitoring outside the clinic. Impedance pneumography (IP) is a noninvasive respiratory surrogate signal obtainable by wearables that shows promise for ambulatory monitoring. However, the lack of effective globalized approaches that are not reliant on subject- or posture-specific calibration for IP-based TV estimation poses a considerable research problem. In this study, we address this problem by fusing IP and accelerometry-derived respiration (ADR) signals with ridge regression models to estimate TV using a subject- and posture-independent model. Multimodal wearable data were collected from 10 healthy subjects undergoing breathing maneuvers while sitting, lying supine, and standing. Leveraging the complementary ADR signal and demographic information, TV estimation was improved with a coefficient of determination ( R^{2} ) of 0.77 compared to a model using IP alone with an R^{2} of 0.69. The 95% limits of agreement (LOA) were similarly improved by 29.5% with the multimodal model. The results indicate that a multimodal approach to globalized TV estimation is feasible and may serve as a foundation for enabling remote TV monitoring.