Remote Presence in Intelligent Environments Frank Bentley Artificial Intelligence Laboratory Massachusetts Institute of Technology Cambridge, Massachusetts 02139 http://www.ai.mit.edu @ MIT The Problem: As the use of intelligent environments expands, the ability for people to communicate with others in geographically distant environments is becoming more and more necessary. In order to create a rich collabora- tive experience, high-resolution video and audio should be broadcast among locations to allow others to see the people communicating from remote locations. Also, the state of the environment should be transmitted to all par- ticipants. This could allow for an increased quality of verbal communication using non-verbal clues, collaboration on documents, and synchronous display of media. Motivation: As part of the Intelligent Room project at the MIT Artificial Intelligence Lab, the presence project will seek to bring multiple rooms into communication with each other. Most importantly, it will provide a means of interaction for people in geographically distant locations. This interaction will be more rich that standard video conferencing solutions because it will provide connections between aspects of the room’s state. For example, users in each room could automatically view the same web page as another user, or hear certain speech generated by the room. This sharing of state will make collaboration across distance easier for all participants and will provide them with key data without needing to interrupt the flow of verbal and visual communication. With collaboration tools in place, the Intelligent Room can move beyond an isolated point in space and will be able to connect people in interactive, collaborative meetings and discussions. As face-to-face meetings become less common and more people rely on technology for their communication needs, this type of collaboration tool will provide a virtual presence and increase the quality of remote collaboration. Previous Work: There are currently many projects devoted to collaboration in media-rich environments. One of the most well-known is the Access Grid project which grew out of Argonne National Labs in Argonne, IL and now has 49 grid node sites throughout the country [4]. The ’Grid’ provides a rich environment of shared video streams among many users of a high bandwidth network. It also currently supports a shared slide presentation utility to allow all users to view the current slide in a live presentation. Some Grid sites are working to integrate computationally expensive information not available at remote loca- tions by sharing video streams from advanced visualization systems. Projects like these expand the Grid to more than just a video link between two people and allow collaboration not possible before [3]. The Grid falls short of many people’s expectations. One of its main criticisms is that it does not create a virtual shared space. When a participant turns his head to the left, for example, he may not be looking at the same thing that is to the observer’s right. Another reason that the Grid does not create a true feeling of presence is that current uses of the Grid do not allow for detailed video focused on a single person. Therefore, it is impossible to see each participant’s individual gestures and facial expressions [5]. Finally, the Grid is very difficult to use and does not easily blend with existing communication. The sharing of documents other than PowerPoint slides is not supported which makes collaboration difficult. For these reasons, the main use of the Grid has shifted towards distance learning and away from collaboration. Although collabora- tive events occur on the grid, they are often very one sided conversations that have a presentation style. Remote participants usually only contribute in the form of questions. To facilitate true collaboration, a different type of environment is needed. Commercially available programs such as Microsoft’s NetMeeting also provide collaboration tools. NetMeeting is becoming popular in industry as a way to eliminate face to face meetings. However, NetMeeting has its short- comings. Among these is a limit of one receiving video stream per user. It is very difficult to create an illusion of presence with only one video stream. There are features of NetMeeting that should be included in the Intelligent 203