ABSTRACT
Hypnosis control is an essential process commonly used during medical surgeries and operations. In clinical practice, this process is achieved by an anesthesiologist who estimates the required dose for a patient and monitors the hypnotic status of the patient. For closed-loop sedation control systems to be approved for clinical practice, they have to demonstrate efficiency and robustness under parameter uncertainties and potential device faults. In this paper, new modeling and analysis of the closed-loop anesthesia administration are proposed using priced timed automata. The modeling involved the physiological system, the closed-loop controllers, and the fault scenario. The physiological model is based on a general model that accounts for parameter variability and residual errors from a broad group of data-sets. Two control techniques are analyzed: the proportional-integral-derivative controller and a variant of the sliding mode controller. The results have shown that the performance of both controllers was impacted by the sensor fault with the later one outperforming the former.Clinical relevance- The proposed in-silico methodology is used to estimate the performance degradation in closedloop anesthesia administration as a result of temporal faults. Moreover, it allows for evaluating different control techniques and help design reliable automatic control.
