Parameter estimation in the canine cardiovascular system

A new method of parameter estimation has been developed to estimate a newly chosen group of eight parameters of the canine arterial system. These parameters, which are physically meaningful, include arterial radii, Young's modulus for the aortic wall, aortic length, and peripheral resistances. The method is model-based, and depends upon minimization of a criterion which combines weighted integrals of absolute values of pressure and flow with errors in certain average and peak values of waveforms. A minimization algorithm which is a modification of the Hooke and Jeeves pattern search method was used. The entire scheme was implemented on a hybrid computer. This method of parameter estimation was developed and checked by use of model-to-model studies. Then dog-to-model parameter estimation runs were made, using data recorded in a dog experiment. Estimated parameters were found to check direct measurements where these were possible, and variations in estimated parameters from data obtained after various drug interventions were found to fall into expected patterns. Future applications of the methods developed would appear to have promise for obtaining rapid estimates of human parameters from noninvasive measurements made in clinical situations.