Finite-Difference Time-Domain Modeling of a High-Resolution Human Phantom

Abstract

In this thesis, I study electromagnetic (EM) mechanismas of wave interaction with human tissues in 10 GHz to 100 GHz band. The finite-difference time-domain (FDTD) method is employed among various EM analysis methods. The FDTD method is easy to implant, powerful, robust, and has the advantage of obtaining a broadband response in a single simulation. For the FDTD analysis in the frequency band of interest, the minimum voxel size required by FDTD should be satisfied. Therefore, I develop a voxel resolution enhancement method to create a human phantom suitable for FDTD analysis in the frequency range of interest. The enhancement method for voxel resolution consists of a voxel size reduction method and a voxel smoothing method. In this thesis, I consider the Ella model of Virtual Family to obtain a high resolution human phantom. Dispersion modeling is also required for FDTD analysis of a high resolution human phantom. In this work, quadratic complex rational function dispersion modeling is performed for 75 human tissues. FDTD analysis of a human resolution human phantom requires overwhelming computational time and memory because of the increased number of FDTD cells. Therefore, massage passing interface parallel processing programming is applied to increase computational efficiency.