Adaptive Pi Controller for Postoperative Blood Pressure Control

Postoperative hypertension is common among cardiac patients. Intensive care by medical personnel is necessary to avoid complication due to elevated blood pressure after operation. Automatic adaptive drug delivery control system is designed to effectively inject vasodilator drug into patient’s cardiovascular system to reduce elevated blood pressure. The mentioned vasodilator drug used is sodium nitroprusside (SNP) which is short potent in effect, and thus infusion rate is the prime factor of its effectiveness. Adaptive tuning of this control system allows fast action to accommodate dynamic changes in patient’s body. Introduction The phenomenon of postoperative hypertension is normal among cardiac patients. Especially patients after coronary artery bypass grafting procedures. The elevated blood pressure must be reduced as soon as possible in order to avoid, for instant, deterioration of suture lines that will lead to increase bleeding, or even losing a life. Continuous infusion of vasodilator drug, such as sodium nitroprusside (SNP), would reduce the elevated blood pressure, almost immediate. Intensive care by clinical personnel is necessary. Most of the time, the injection of SNP is done manually by clinical personnel, for instant, nurses, which is always poor in quality. In order to overcome this problem, a computerized adaptive PI controller system has been proposed. The objective of this closed-loop control system is to control the infusion rate of vasodilator drug by using the feedback mean arterial pressure (MAP) from the cardiac patient. This control system emphasize on achieving a settling time less than 10 minutes and having an overshoot of less than 10mmHg of MAP. This adaptive system was designed with 3 different controllers: high infusion rate, normal infusion rate and low infusion rate. This is to correspond with 3 plant models of low sensitivity, normal sensitivity and high sensitivity, which are the models of actual patient’s sensitivity with respect to SNP. By comparison of MAP samples, this control system is able to determine the sensitivity of the patient and thus assign a suitable infusion rate controller. This system will bring the MAP of the patient as close as possible to the desired acceptable MAP so as to avoid complication due to postoperative hypertension. The transfer function and mathematical model used in this paper was originally proposed by Slate[1] and relates the reduction of MAP with the infusion rate. An algorithm presented on this paper, proposed by K.Y.Zhu, J.Ma and Krishnan, S.M[2], is to adjust the reference plant model, so that the plant model can be identified and suitable controller parameters can be evaluated. The infusion rate and change in MAP also within the preset limit. Problem formulation A dynamic mathematical model of MAP is presented by Slate[1]: MAP = Po P∆ + Pd + n (1) Where MAP is the mean arterial pressure, Po is the initial blood pressure, also called the background pressure, assumed constant, P∆ is the change in blood pressure due to infusion of SNP, Pd is the change in pressure due to the rennin reflex action which is the body’s reaction to the use of a vasodilator drug, n is noise. In this paper, Po is assumed to be constant. A continuous-time deterministic model describing the relationship between the change in blood pressure and SNP infusion rate is [1]: ) ( 1 ) 1 ( ) ( s I s e Ke s P s T s T c i