iWorlds: Generating artificial control systems for simulated humans using virtual worlds and intelligent environments

Abstract

This paper highlights research into using virtual worlds, intelligent environments and mixed reality to create artificial control systems for simulated humans. Following a brief explanation of this project, the beneficial contribution provided by virtual worlds, intelligent environments and mixed reality systems is outlined. Throughout this project a series of experimental iWorlds, (virtual and mixed reality intelligent environments) have been used to profile the actions and decisions of humans to generate sets of realistic behaviours and test the project research hypothesis. These worlds consist of a mixed-reality household experiment plus two supporting game-like behaviour experiments, namely ‘Bar-World’ and ‘Battleships’ which were designed to explore and reinforce aspects of the underlying theory. As a core part of this work a description of a new profile generation mechanism is presented, which comprises a genetic programming mechanism that blends real user behaviour data to produce varied artificial profiles for non-player computer characters. In support of this a set of experimental results are presented that reveal the new artificial intelligence mechanism does indeed advance the project aim to create realistic simulated humans. A discussion is included that describes the potential benefits that could be gained by incorporating these methodologies into computer games and professional intelligent environment development systems.