Application of the Kalman Filter for Faster Strong Coupling of Cardiovascular Simulations
Application of the Kalman Filter for Faster Strong Coupling of Cardiovascular Simulations
复制标题
DOI：
10.1109/jbhi.2015.2436212
复制
发表时间：
2016-07
影响因子：
7.7
通讯作者：
Yukiko Hasegawa;T. Shimayoshi;A. Amano;T. Matsuda
中科院分区：
文献类型：
--
作者：
Yukiko Hasegawa;T. Shimayoshi;A. Amano;T. Matsuda
本文提出了一种降低多尺度心血管仿真模型强耦合计算成本的方法。在该模型中，心肌细胞、左室结构动力学和循环血流动力学的各个模型模块是耦合的。强耦合方法计算稳定、准确，但需要迭代计算，计算量大。如果预测者提供准确的初始近似，迭代计算就可以减少。该方法使用卡尔曼滤波，通过滤除过去值中包含的噪声来估计准确的预测。通过应用于先前发表的多尺度心血管模型来评估所提出方法的性能。与无预测法和拉格朗日外推法相比，该方法的迭代次数分别减少了90%和62%。即使在参数变化和左室有限元模型单元数增加的情况下，所提方法所需的迭代次数也明显低于无预测时的迭代次数。这些结果表明了该方法的鲁棒性、可扩展性和有效性。
In this paper, we propose a method for reducing the computational cost of strong coupling for multiscale cardiovascular simulation models. In such a model, individual model modules of myocardial cell, left ventricular structural dynamics, and circulatory hemodynamics are coupled. The strong coupling method enables stable and accurate calculation, but requires iterative calculations which are computationally expensive. The iterative calculations can be reduced, if accurate initial approximations are made available by predictors. The proposed method uses the Kalman filter to estimate accurate predictions by filtering out noise included in past values. The performance of the proposed method was assessed with an application to a previously published multiscale cardiovascular model. The proposed method reduced the number of iterations by 90% and 62% compared with no prediction and Lagrange extrapolation, respectively. Even when the parameters were varied and number of elements of the left ventricular finite-element model increased, the number of iterations required by the proposed method was significantly lower than that without prediction. These results indicate the robustness, scalability, and validity of the proposed method.