Scene-Aware Autonomous Behavior for AR Characters

Abstract

This thesis explores the development and implementation of scene-aware autonomous behaviour for characters in Augmented Reality (AR). We utilize the NPC Engine and the Intelligent Content Creation Tools for Augmented and Virtual Reality (ICCTARVR) framework to create AR characters that can autonomously interact with their environment in a meaningful and intelligent manner. The focus is on enhancing the experience of AR by enabling characters to make decisions and perform actions based on a simulated set of needs, minimizing the need for scripted characters and interactions. Building upon a previous thesis in the same realm, we implement a method for generalized user interaction in the NPC Engine and create an application in which the user interacts with and cares for three virtual pets. We conduct a user study to evaluate the quality of the pets’ behaviour and emergent narrative potential. The study found that the domain and its simulated needs allowed for engaging behaviour that showed promise as a method for creating interactive emergent narrative experiences.