Fuzzy Logic Control of Blood Pressure during Anesthesia

This article describes research on a fuzzy logic controller which controls mean arterial pressure (MAP) which was taken as a parameter for depth of anesthesia. The controller was designed and implemented on a personal computer. T h e design process was iterative and the reference points of the membership functions as well as the linguistic rules were determined by trial and error. The control rules made use of the error between the desired and the actual values of MAP as well as the integral of the error. The controller was tested in 11 different surgical operations and it was observed that the anesthetists supervising the controller never had to intervene or override it. Moreover, measures of the quality of keeping the MAP constant show the fuzzy controller to be superior to an ordinary human controller. Such a fuzzy logic controller has the potential for use during anesthesia with an agent like isoflurane. However , further clinical studies are needed before it becomes routine practice to use such a fuzzy logic controller for the control of anesthesia. One of the main tasks of the anesthetist during surgery is to control the depth of anesthesia. However, the depth of anesthesia is not readily measurable. In clinical practice, the depth of anesthesia is evaluated by measuring blood pressure, heart rate, and clinical signs such as pupil size, motor activity, etc. Control of depth of anesthesia has become more difficult today than at the times of ether as the classical signs of depth of anesthesia defined by Guedel are no more used. Cullen [ 11 showed a good correlation of blood pressure to the anesthetic dose given. EEG signals have been used by Schwilden [2] for the dosage of Propofol. However, in clinical routine EEG is difficult to use due to artifacts and large intra-and interindividual variations partly caused by differing a n e s t h e t i c a g e n t s. C o n t i n u o u s electromyographic recording of spontaneous activity of the upper facial muscles has been used [3] but results in unpredictable values when the patient receives muscle relaxants. Evoked responses have been used [4] for various agents; they, too, show an None of the above-mentioned methods has been established on a routine basis. Anesthesiologists still use blood pressure as the most reliable guide for dosing inhaled anesthetics. An argument heard …