Generalized Predictive Control of Depth of Anesthesia by Using a Pharmocokinetic-Pharmacodynamic Model of the Patient

Monitoring and controlling the depth of anesthesia during a surgery is really important since over dosing and under dosing can be dangerous for the patients. The model which is used for describing the relationship between input anesthetic agents and output patient endpoint variables is pharmacokinetic- pharmacodynamic model which included the most significant covariates such as age and weight. Bispectral index (BIS) is one of the best criteria for evaluating the depth of Anesthesia. In this research we used BIS as a patient endpoint and Propofol as an anesthetic agent. As there is a large variety between the patients, we need a controller which should be robust against the disturbances and because the anesthesia process is nonlinear and contains a time delay, model predictive controllers (MPCs) seems act very well for it. In this paper we tried to use a constrained generalized predictive control (GPC) method for controlling the depth of Anesthesia. For comparison a PID controller has been designed. The results showed that the performance of GPC with or without presence of the noise and disturbance was much better than PID controller and also it was more robust.