EDTree: Emotional Dialogue Trees for Game Based Training.

Abstract

Immersion and interactivity are a major focus when creat- ing gaming applications, as technology has improved and enabled the creation of larger and more detailed virtual environments the need for more engaging NPCs (non-playable characters) is also required. Many games utilise a form of dialogue tree when conversing with characters within a gaming application, allowing the user to choose their question- s/responses. While this method does provide a dynamic conversation system, it is quite a one-sided level of interactivity with the NPC simply responding to the current question without it a ecting the conversation on a whole. We present a novel dialogue system that explores the emo- tional state of the NPC to provide a more complex form of dialogue tree, termed EDTree (Emotional Dialogue Tree). Based on user actions, the interactions between the user and the NPC are enriched by the emo- tional state of the NPC. Utilising this system will provide an immersive experience based around improved believability of virtual characters. To demonstrate the e ectiveness of our approach, we show an example of a training system that explores the use of gaming technology and the proposed EDTree.