Multiple models oscillometric blood pressure monitor identification

Oscillometric blood pressure monitors are fairly reliable medical equipment, inexpensive and widely used both in domicile and clinical measurements of blood pressure, however they have not yet been subject of deep investigation. The research reported in this paper presents some a priori knowledge to build a model of an oscillometric blood pressure monitor and the results of the system identification, in order to help in the search for improvements in these devices' measurement accuracy. From several studies it is known that the cuff expansion, the compressible fluid flow and the electromechanical components of the system originate nonlinearities in the system dynamics. During the identification procedures it was seen that a multiple model approach was required in order to obtain a reliable representation in all the pressure range. The study of numerous parametric linear models and interpolation functions allowed the reduction of prediction errors, resulting in a final model with good quality results but with room to improvements in the transient response.