Improving Algorithms for Oscillometric Blood Pressure Estimation by Suppressing Breathing Effects

Abstract

Blood pressure estimation by the oscillometry is a practice growing in popularity. Algorithms for blood pressure estimation are diverse, however little effort has been put forth to assess their performance. This thesis first surveys and assesses the algorithms used for oscillometric blood pressure estimation. Of all the known algorithms, the results of this work revealed one procedure which performed the best. These algorithms were evaluated by readings from two trained nurses. Next, we developed algorithms for extracting and suppressing breathing effects on blood pressure estimation. Breathing causes fluctuation in blood pressure and current oscillometric devices do not account for these effects. Extracting breathing signals extends the capabilities of existing oscillometric devices, such as reporting respiratory sinus arrhythmia, without the need for hardware changes. Suppression of these effects is performed by homomorphic and adaptive filtering. Results show improvement in that estimated pressure after suppression was closer to the nurse readings.