Machine-learning rule-based fuzzy logic control for depth of anaesthesia

A machine-learning rule-based fuzzy logic controller for depth of anaesthesia which is similar to the way an anaesthetist works is presented in this paper. The results of discussions with anaesthetists to obtain a rule base and the application of fuzzy logic to predict the primary depth of anaesthesia (PDOA) and to control drug administration are very promising. By using simple rules from machine learning trials, similar results for the prediction of PDOA were obtained and can be used to design a drug infusion controller. The robustness of the self-organising fuzzy logic control (SOFLC) algorithm is good and can supplement the anaesthetist's experience for administering drug to patients when the system is dynamic and time-varying. Using these results, the design of a hierarchical architecture for the determination of the level of depth of anaesthesia is being investigated, which will include the use of clinical signs and refinements in the control of drug administered to patients.<
>