A Psychometric Detection System to Create Dynamic Psychosocial Relationships Between Non-Player Characters

摘要

Creating realistic emotional responses is usually the focus of research in embodied agents, conversational agents, and social robots to make them act more human-like. This level of research is not applied to Non-Player Characters (NPCs) in games, where they can be perceived as too chatty (or not chatty enough), behave incongruously to the relationship they have with the player, or respond in generally unsuitable or undesirable ways. We present PSYCH, a framework that enables game programmers to extract psychometrics (emotions, for example) from the player and transmit them to game NPCs. The implications of our novel approach using PSYCH is that NPCs are exposed to richer data that gives greater situational and emotional context to the NPC. This results in expanded behaviors that can be translated into more unplanned game interactions. We demonstrate the capabilities of the framework by integrating it with a pre-existing psycho-social framework so that a game programmer can use PSYCH to generate different behaviors in the framework's NPCs by having the player speak simple phrases to them.