International Journal of Artificial Intelligence and Neural Networks
Author(s) : HASAN GULER
The main purpose of this study is to design universally accepted approach to weaning patients from ventilator. Fuzzy logic controller was used for achieving this aim. The reason for the use of fuzzy controller is that it represents expert notions employed in decision making. The developed all protocols were designed on LabVIEW software. The inputs of the developed system were PH, partial carbondioxide percent (PaCO2), saturation of oxygen (SpO2), the rate of PaO2/FiO2, saturation of oxygen (SaO2), hemoglobin (Hb) and positive end expiratory pressure (PEEP). 20 clinical scenarios were generated by using Monte-Carlo simulations and Gaussian distribution method to test performance of the developed protocol. The generated scenarios were applied to the developed protocol. Clinician was asked to evaluate the each scenario to determine weaning probability. Student t-test for p<0.05 was used to show statistical difference between the developed protocol and clinician’s evaluation. According to results, there is no statistical difference for % 97.8 probability between them.