A Computational Model of Emotions for Agent-Based Crowds in Serious Games

Abstract

Crowd behaviour is an interesting social phenomenon that emerges from complex interactions of individuals. An important aspect of individual behaviour is emotion which plays a significant role in all aspects of human decision making. For example, heightened emotional states can lead people to take highly unexpected or irrational actions. One popular motivation for simulation of virtual crowds is to generate believable characters in movies and computer games. Recently the concept of serious games has been introduced in both academic and industrial circles. In this paper, we propose an emotion engine, based on modern appraisal theory, that is able to model various emotional crowd characteristics. This appraisal engine is capable of capturing the dynamics of emotional contagion and we show how different crowd composition can lead to different patterns of emotional contagion. In addition, we describe a serious game designed for training military personnel in peaceful crowd control. We evaluate this engine in the context of a property protection protest scenario where the players or soldiers are tasked to maintain a peaceful protest without violence. A systematic evaluation is presented which supports the facial validity of the emotion engine and our model of emotional contagion.