Agent-based target detection in 3-dimensional environments

Abstract

Visual perception modeling is generally weak for game AI and computer generated forces (CGF), or agents, in computer games and military simulations. Several tricks and shortcuts are used in perceptual modeling. The results are, under certain conditions, unrealistic behaviors that negatively effect user immersion in games and call into question the validity of calculations in fine resolution military simulations. By determining what the computer-generated agent sees using methods similar to that used to generate the human players' screen view in 3- D virtual environments, we hope to present a method that can more accurately model human visual perception, specifically the major problem of a entity "hiding in plain sight"