Controlling the emotional state of an embodied conversationalagent with a dynamic probabilistic fuzzy rules based system

This paper presents an innovatory system for controlling the emotional state of an embodied conversational agent. Unlike other existing models, our system is fast adaptable to different application domains and it is highly interpretable. This will provide specialists with a tool to easily test their hypotheses about the most suitable emotional attitude for an agent in a specific domain. It also provides emotional stability and dynamic and automatic behavior orientation, which will result in the design of more believable conversational agents that behave as humans do.