Blood pressure estimation using pulse transit time models

Abstract

Blood pressure (BP) is an important indicator of human health. Common methods for measuring BP continuously are either invasive, intermittent or they require using a cumbersome cuff. Pulse Transmit Time (PTT) -based measurement can be an alternative for such methods, as it ensures continue and non-invasive monitoring. However, since the method is indirect, it requires careful modelling of PTT-BP relation. In this thesis, three approaches of BP estimation from PTT are tested: linear regression, nonlinear Moens and Korteweg model and nonlinear model developed by Gesche. In the experiments, cardiovascular pulses for PTT were sensed using two fiber optics based accelerometers developed at the University of Oulu.

To evaluate feasibility of presented models, the results were compared with reference BP values, measured using methods accepted for the commercial use. There were two groups of data. One was compared with BP measured using invasive catheter. Second group was compared with BP measured using volume clamp method. Obtained results suggest, that the presented calculation methods in present state still require further development in order to provide accurate BP values, however, they can be potentially used for observation of BP changes.