Beat-to-Beat Oscillometric Blood Pressure Estimation: A Bayesian Approach With System Identification

Our study aims to advance noninvasive blood pressure (BP) monitoring through the introduction of innovative beat-to-beat oscillometric BP estimation methods. We aim to overcome current device limitations by delivering continuous and accurate BP estimates, utilizing physiologically based mathematical models.

We developed novel beat-to-beat oscillometric BP estimation methods based on physiologically grounded mathematical models of intra-arterial BP and the arterial system effect. Our approach includes a recursive Bayesian method for parameter estimation and a new system identification technique to refine initial parameter estimates. We tested our methods through simulations and real-world experiments involving 10 individuals.

Mean errors for systolic and diastolic BP were as low as -1.26 mmHg and 2.03 mmHg, respectively, with standard deviations of errors at 5.95 mmHg and 4.16 mmHg. Furthermore, our methods enabled the estimation of additional cardiovascular parameters such as heart rate, respiration rate, and mean arterial pressure.

Our novel beat-to-beat oscillometric BP estimation methods offer a significant advancement in noninvasive BP monitoring technology, addressing the limitations of current devices by providing continuous beat-to-beat BP estimates.

Our approach represents a promising direction for improving the reliability and comprehensiveness of cardiovascular parameter estimation in noninvasive BP monitoring devices, facilitating more effective patient care and monitoring.