A new cognition-based chat system for avatar agents in virtual space

Short internet-based chat is typical of modern communication, especially for on-line game players. Improved chat tools are available to avatar agents in virtual space (e.g., Second Life) thanks to the fast-evolving 3D internet. The most common plain text chat is easy to use, but it is hard to understand who talks to whom. Advanced chat with 2D word balloons in 3D virtual space is difficult or unfamiliar, since it does not consider real world constraints. We propose a human-cognition based chat system for virtual avatars using geometric information. In our chat framework, if an agent (avatar) wants to talk with other people, then that agent should approach within an "audible distance" to read the text, as would be required in the real world to participate in a conversation. Additionally we propose a new model to manage the virtual chat social network over our cognition-based chatting system. Our experiment showed that the number of chat agents is saturated regardless of the number of chat agents, as occurs in real-world chat. Thus, the social-graph for chat agents is always of a manageable size in our cognition-based chat system.