Multi-Modal Parametric Modeling of Airflow and Photoplethysmography (PPG) Signals Using Advanced Kalman Filtering for Real-Time Sleep Apnea Characterization

• Abiodun Okunola

  Ladoke Akintola University Technology

  Author

Sleep apnea, a prevalent sleep-related breathing disorder characterized by repeated upper airway obstruction during sleep, affects millions worldwide yet remains significantly underdiagnosed due to the complexity and cost of gold-standard polysomnography (PSG) . While home sleep apnea testing has emerged as a promising alternative, existing single-modality approaches suffer from limited accuracy, motion artifacts, and insufficient signal quality during hypopnea events . This research addresses the critical gap in multi-modal respiratory monitoring by presenting a novel framework that integrates airflow and photoplethysmography (PPG) signals through advanced Kalman filtering for real-time sleep apnea characterization. The proposed system employs a parametric state-space model where respiratory effort derived from PPG serves as a surrogate for direct airflow measurement, while a Kalman filter architecture enables optimal fusion of both modalities with adaptive noise rejection . Validation on clinical trial data from 31 subjects demonstrated that the combined approach achieved 90.3% accuracy, with sensitivity of 84.6% and specificity of 94.4%, significantly outperforming single-modality airflow-only (83.9% accuracy) and effort-only (87.1% accuracy) methods . The fused system achieved a correlation coefficient of R² = 0.92 against reference PSG measurements . This research contributes a replicable, computationally efficient framework for real-time apnea detection suitable for wearable and home-based monitoring applications, with implications for early screening, severity stratification, and longitudinal sleep health management.

• PDF

Copyright (c) 2026 Abiodun Okunola (Author)

This work is licensed under a Creative Commons Attribution 4.0 International License.