Model-based oscillometric blood pressure estimation

Oscillometry is the most common measurement method used in automated electronic blood pressure (BP) monitors. A variety of oscillometric BP algorithms exist in the literature. However, most of these algorithms are without physiological and theoretical foundation. Moreover, most of the existing oscillometric algorithms estimate the BP from the envelope of the oscillometric pulses and ignore the wealth of information that the oscillometric pulses contain. More information could be obtained from the amplitude and time characteristics of the oscillometric pulses at different cuff pressures if an accurate mathematical model is developed. This paper reviews three novel model-based oscillometric BP estimation methods developed by our research group. These methods include (i) mathematical modeling of the oscillometric waveform envelope and BP estimation using neural networks, (ii) mathematical modeling of the oscillometric waveform and parameter estimation using extended Kalman filter, and (iii) mathematical modeling of the pulse transit time (PTT) and estimation of BP based on PTT-cuff pressure dependency. The performance of the proposed methods was evaluated on simulated and actual data in terms of mean error, mean absolute error, and standard deviation of error.