Moody5: Personality-biased agents to enhance interactive storytelling in video games
Year: 2022
Type: conference-paper
Abstract: In story-driven video games, such as Role-Playing Games (RPGs), a static and repetitive interaction with Non-Player Characters (NPCs) can negatively impact fun and immersion. On the contrary, an NPC able to show emotions and endowed with a personality that matches its behavior can keep the user engaged for a longer time. Unfortunately, this aspect has not yet been fully investigated. We propose Moody5, a preliminary solution designed to help game designers create “personality-biased” agents able to interact in sensible ways in the framework of interactive storytelling. In particular, it supports the creation of agents that exploit the Goal-Oriented Action Planning (GOAP) technique to reach their goals but whose chain of actions and quality of the interaction with the player are dynamically affected by personality and emotions. To obtain this result, we have borrowed ideas from the Big Five theory [1] and the Emotion Facial Action Coding System (EMFACS) by Ekman & Friesen [2]. We validated the proposed solution in a narrative test scenario inspired by the Harry Potter universe. Our preliminary survey demonstrated that Moody5 could improve the gameplay experience and replay value while providing a helpful Unity plugin for game developers.
Source: IEEE Conference on Games (CoG)
Institution: University of Milan
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W4296474773