From Individual Characters to Large Crowds: Augmenting the Believability of Open-World Games through Exploring Social Emotion in Pedestrian Groups

The document proposes a novel approach to enhance the believability and player immersion in open-world games by significantly improving non-player character (NPC) crowd simulation. The core thesis is that current crowd models are insufficient, often treating NPCs as isolated individuals with limited reactions, which detracts from the realism of virtual environments. To address this, the authors advocate for integrating principles from social psychology, particularly regarding group dynamics, social emotions, and inter-character relationships, into game design.

The proposed methodology involves developing a network-type data structure to simulate complex social interactions and emotional responses within virtual crowds. This framework would move beyond individualistic emotional models, such as the PAD model, by incorporating concepts like social cognition (warmth and competence dimensions), out-group image theory, and intergroup emotions. By modeling how social networks and group dynamics influence emotional states and behaviors, the aim is to generate more realistic and nuanced crowd reactions.

Ultimately, the research seeks to bridge a critical gap in existing crowd simulation techniques, which frequently overlook the social and emotional dimensions of human groups. By applying psychological literature on social emotions and group behavior, the document outlines a path to create virtual crowds that exhibit believable collective actions and emotional responses, thereby significantly augmenting the immersive quality of open-world game experiences. The scope of this work is focused on the conceptual and computational modeling aspects for game development, without specific geographic or time period limitations mentioned.