Institution: Jackson State University
Department: Electrical and Computer Engineering and Computer Science
Proposed research ideas:
Multiregion FDTD (MR/FDTD) approach seeks to avoid the inefficiencies of the classical FDTD when applied to large, sparsely filled modeling problems by eliminating explicit calculation of fields in the space between subregions surrounding the modeled objects. In MR/FDTD the problem space is divided into several independent subregions distributed in otherwise free space. The fields in the subregions are determined with the use of localized FDTD lattices, and the interaction between the subregions is accounted for by performing a surface integration of equivalent sources on the subregion surfaces.
FDTD method is normally inefficient in handling large scale problems. In most cases, the domain can be divided into small subdomains separated by free space. If we can work on each domain individually while taking the effect of each one on the others using near field to near field transformation to the scattered field on the surfaces, this can significantly decrease the memory requirement and enable us to hand large scale problem.