Controlling depth of anesthesia using PID tuning: a comparative model-based study

The control of depth of anesthesia presents a challenging and realistic problem that calls for fast and adaptable control techniques. However, patient-to-patient variability in model parameters poses the question of which control strategy can generate best results. We studied three different patient models, i.e., the slate model, the isoflurane unconsciousness model, and the isoflurane to muscle relaxation interaction model. Due to the simplicity of the PID control and easy implementation, a discrete PID controller with actuator saturation is considered in this study to supports these three pharmacokinetic models. The models being tested can give us an understanding of the effects of disposition of drug in the body by measuring the mean arterial pressure (MAP). The system characteristics are such that the PID controller should regulate the error signal of the gases sodium nitroprusside (SNP) and isoflurane using MAP readings from the patient. The system responses under the discrete PID control are presented and the results show acceptable performance.