Nonlinear State Estimation Technique Implementation for Human Heart Model

Abstract

Human heart is a vital organ therefore proper diagnosis of heart activities is essential. Various parameter estimation techniques have been developed to estimate heart parameters. In this work, we use Ensemble Kalman Filter (EnKF) and Particle Filter (PF) for dynamic assimilation of human heart parameters. EnKF and PF are modified filters specifically designed for state prediction of nonlinear systems with large data samples. A third order mathematical heart model is used to estimate three heart parameters that includes movements of heart muscle fiber, tension in heart muscle and electrochemical activity of the heart. EnKF and PF are applied to heart model and different case studies are performed to observe the prediction accuracy by comparing sum squared error values. Case studies are performed with variable state and measurement noise values. The proposed approach demonstrates promising results in accurately predicting the human heart parameters.