Improving the Control of Preterm Infant Mass Skin Temperature Using Adaptive Neuro Fuzzy Inference System

This paper presents a system for controlling an infant skin temperature using ArtificialNeuro-Fuzzy Inference System (ANFIS). Mathematical and Simulink models of the infant and the incubator system were developed. The ANFIS model integrates both the Fuzzy Logic Control (FLC) rules and the Artificial Neural Network (ANN) to produce optimum control for the variation of the skin temperature of the preterm infant inside the incubator. Simulation was carried out using the ANFIS model, open loop incubator and FLC in two instances. A0.9kg and 2.5kg infant masses and postnatal ages of 5 days were used in the first and second instances. Result showedthat in the first instance, the ANFIS model made a 2.2% and 0.82% improvement over the open loop incubator and the FLC respectively. While a 6.5% and 2.3% improvementwere also made by the ANFIS model over the open loop incubator and the FLC in the second instance.