Adaptive EPSAC predictive control of the hypnotic component in anesthesia

The applications presented in this paper are oriented towards the study of the possibilities to apply adaptive algorithms to a model based EPSAC (Extended Prediction Self-Adaptive Control) scheme for control of depth of anesthesia (DOA). The input variable is Propofol, an anesthetic substance used for both induction and maintenance control of DOA which is evaluated by means of the Bispectral index (BIS). An adaptive control algorithm based on an indirect adaptive predictive method is used. The unknown model parameters are estimated by recursive least square algorithm. The simulation tests performed on virtually generated models for a set of 12 patients show a fast response, optimal dosage and robustness of the algorithm regarding the control of the desired BIS setpoint during induction and maintainance phase.