Modeling, estimation, and control of human circulatory system with a left ventricular assist device

Yi Wu, Paul E. Allaire, Gang Tao, Don Olsen

Research output: Contribution to journalArticlepeer-review

50 Scopus citations

Abstract

In this paper, a state-space model is developed through theoretical analysis and numerical solutions to approximate the response of the human circulatory system. This system model has one critical time-varying parameter: the resistance of peripheral blood vessels. A parameter estimation scheme is derived to estimate this parameter, and the parameter estimate is used to implement an adaptive observer to estimate the aortic pressure for physiological control. An optimal adaptive controller is proposed to control the estimated aortic pressure to track a reference signal updated by a nonlinear function of the pump head to meet the physiological need. A Matlab simulation program and an experimental mock human circulatory loop are employed as test environments for the human circulatory systems with a left ventricular assist device and their physiological controllers. Different physiological conditions, such as the variation of left ventricular failures, variation of activities, and collapse of the left ventricle, are evaluated to test the designed physiological control system. Simulation and experimental results consistently show that the aortic pressure estimation error is small, and that the abnormal hemodynamic variables of a congestive heart failure patient are restored back to the normal physiological range.

Original languageEnglish (US)
Pages (from-to)754-767
Number of pages14
JournalIEEE Transactions on Control Systems Technology
Volume15
Issue number4
DOIs
StatePublished - Jul 2007

All Science Journal Classification (ASJC) codes

  • Control and Systems Engineering
  • Electrical and Electronic Engineering

Fingerprint

Dive into the research topics of 'Modeling, estimation, and control of human circulatory system with a left ventricular assist device'. Together they form a unique fingerprint.

Cite this