Neuro-Fuzzy Design of Phased Array Applicators for Microwave Hyperthermia

A 6-mm resolution 7-tissue (skin, blood, bone, muscle, gray matter, white matter, air) anatomy based model (62times62 cells) cross section of human head is studied. Four dipoles of a 64 MHz is applied. A neuro-fuzzy modeling of the head is presented. A system that describes input-output behavior in a way similar to mathematical mapping without involving differential operator or equation is called a static system. The cell size of studying domain size is 6 mmtimes6 mm width with 2-dimension formula of FDTD method considering the Yee algorithm. The calculated SAR inside the model using transparent absorbing boundary condition (TAB) method for reflection truncation is used to develop the neuro-fuzzy system to build up a fuzzy inference system (FIS). A mathematical representation of physical system (biological model) is used to design applicators to achieve a desired pattern inside the model. Calculating the SAR at the model central axis at different phases of excitation will be done, the number of calculated patterns is 39 trials. This data is fed to the introduced modeling system to classify the inputs to get a set of clusters in order to determine the number of the rules representing the model