外科SHIEN, JIANN-SHINGJIANN-SHINGSHIENFU, MUMUFUHUANG, SHENG- JEANSHENG- JEANHUANGKAO, MING-CHIENMING-CHIENKAO2009-09-282018-07-112009-09-282018-07-112006http://ntur.lib.ntu.edu.tw//handle/246246/92908This paper assesses the controller performance of a self- organizing fuzzy logic controller (SOFLC) in comparison with a routine clinical rule- base controller (RBC) for sedation control of intracranial pressure (ICP) pattern. Eleven patients with severe head injury undergoing different neurosurgeries in a neurosurgical intensive care unit (NICU) were divided into two groups. In all cases the sedition control periods lasted 1 h and assessments of propofol infusion rates were made at a frequency of once per 30 s. In the control group of 10 cases selected from 5 patients, a RBC was used, and in the experimental group of 10 cases selected from 6 patients, a self-organizing fuzzy logic controller was used. A SOFLC was derived from a fuzzy logic controller and allowed to generate new rules via self- learning beyond the initial fuzzy rule-base obtained from experts (i.e., neurosurgeons). The performance of the controllers was analyzed using the ICP pattern of sedation for 1 h of control. The results show that a SOFLC can provide a more stable ICP pattern by administering more propofol and changing the rate of delivery more often when rule-base modifications have been considered.en-USintracranial pressureneurosurgical intensive care unitrule -base controllerself-organizing fuzzy logic controllerjournal article