A Comparison Among Fast Visibility Algorithms Applied to Computational Electromagnetics

Abstract

This paper presents a set of fast algorithms developed for solving the visibility problem in large complex geometric scenarios. The implemented algorithms are: Binary Space Partitioning (BSP) –based on a binary tree structure- and three new ones: Trimming Method – facets which are partially occluded are trimmed –, Cone Method – a cone emerges from the point of view to discard shadowed facets – and Pyramid Method – a pyramid is used to eliminate non-lit patches – . All the proposed algorithms are tested on a pair of scenarios for determining the field of view of electromagnetic waves in order to calculate the corresponding induced currents on the surfaces. The scenarios consist of two electrically large spheres and two PEC plates respectively where the surfaces are meshed with a variable number of flat triangular patches. The first reflection contribution is calculated using Physical Optics (PO) for both cases. The results show that the computational time can be drastically reduced by assuming small percentage of error in the computed scattered fields.