| MS Seminar

Name of the Speaker: Mr. Arjun R Krishna (EE22S090)
Guide: Dr. Jayaraj Joseph
Online meeting link: http://meet.google.com/keb-mprs-skq
Date/Time: 15th May 2025 (Thursday), 3:30 PM
Title: Novel Calibration-Free Techniques for Non-Invasive Blood Pressure Measurement.

Abstract :

Accurate blood pressure (BP) measurement is critical for diagnosing and managing hypertension, a primary global health concern, the leading cause of 13% global mortality. Traditional non-invasive BP measurement techniques, such as manual auscultation and oscillometry, are limited by their dependence on operator skill, the necessity for calibration, and their susceptibility to physiological and environmental interferences. This study presents innovative, calibration-free, cuff-based BP measurement approaches that address these challenges through streamlined signal processing and automated control. The methods were verified using a public dataset of 350 participants and further assessed through device integration in a pilot cohort.

The research introduces novel technical approaches, including an automated auscultatory BP estimation algorithm that relies solely on Korotkoff sound analysis, thereby eliminating operator bias and the need for population-specific calibration. This approach achieved errors of -1.5 ± 6.6 mmHg for systolic and 3.8 ± 7.6 mmHg for diastolic pressure. A hybrid auscultatory-oscillometric analysis method, combining both signal types to improve accuracy and robustness, demonstrated errors of -1.5 ± 6.6 mmHg (systolic) and 3.5 ± 7.2 mmHg (diastolic). A calibration-free algorithm utilising brachial pulse waveform morphology further refined diastolic estimation, achieving an error of 0.8 ± 2.8 mmHg. Furthermore, a cross-correlation hybrid technique integrating auscultatory and oscillometric data enhanced accuracy, yielding mean errors of 1.4 ± 6.1 mmHg (systolic) and -2 ± 4.3 mmHg (diastolic). All algorithms strongly agreed with clinical reference standards (AAMI/ANSI/ISO) and met international guidelines, including British Hypertension Society (BHS) grade A criteria, when benchmarked against a large public dataset.