Advanced Control Strategies for the Regulation of Hypnosis with Propofol

Anesthesia implies maintaining a triad of hypnosis, analgesia, and neuromuscular blockade by infusing several drugs which specifically act on each of the above aspects. This work focuses on controlling the hypnosis by automatic infusion of propofol using bispectral index (BIS) as the primary controlled variable. A fourth-order nonlinear pharmacokinetic (PK)-pharmacodynamic (PD) representation is used for the hypnosis dynamics of patients. A reliable PK-PD model with associated parameters is obtained from the literature and the closed-loop responses of four types of control strategies (model predictive control, internal model control, controller with modeling error compensation, and proportional-integral-derivative (PID) control) are compared. Robust performance of the four controllers is tested for a broad range of patients by considering variability in PK-PD parameters. Also, the relative performance of the four controllers is studied for different setpoints, noise, and disturbances in BIS signal. Numerical simulations show that all the advanced controllers performed better than the PID controller with the model predictive controller showing the best performance.