A New Approach Based on Dynamical Model of The ECG Signal to Blood Pressure Estimation

Electrocardiogram (ECG) signal represents electrical activity of the heart. Blood pressure (BP), as the output of the heart's activity, is one of the important physiological parameters of the human body. It is demonstrated that there is a complex nonlinear relationship between the ECG signal and BP. The continuous measurement of BP can avail early detection, control and treatment of BP related diseases. To reduce difficulties and adverse effects of the traditional methods of BP measurement techniques which use a cuff, the research on cuff-less BP estimation is of significant interest in the community. This paper presents a new feature extraction algorithm based on McSharry's ECG signal dynamical modeling for estimating BP using only the ECG signal. In fact, this feature vector is formed using the morphology of the ECG signal. The proposed method, utilizing data mining techniques, achieved 1.125 mmHg for mean error and 3.125 mmHg for standard deviation of Systolic Blood Pressure (SBP).