Fuzzy logic knowledge elicitation for model-based ventilator management in the ICU

Ventilator control in the intensive care unit has been an important feedback control problem in critical care. Although attempts have been made in the past using model based or knowledge based techniques, automation is not widely used. We propose a method which combines the quantitative approach in most model based methods and a qualitative approach using fuzzy logic. The paper presents two methods of fuzzy membership function derivation. The fuzzy sets will be used for the model based expert advisory system for the ventilators. The two methods are compared and one of them will be used to derive fuzzy membership functions for other variables and parameters in this project.