Aggregated Avatars for Qualitative Feedback on Interpersonal Interaction Patterns in 3D Environments

Avatars are most often conceived of as 2D pictograms or head / torso / body representations the are human-like, animal-like or otherwise display a fictive creature, with the activity and social/psychological state of a human user mapped onto its features. However, a more general class of avatars can be identified if we relax the constraints behind this conception, whether in the sense of no longer relying on antropomorphic or zoomorphic forms, or in the sense of no longer focusing exclusively on the activities or social/psychological states of individual users. In the latter case, for example, the avatar might reflect some combined state based on the interactions of multiple users, or based on the parameters of the underlying information system as a while. In this paper, we focus on these two generalizations, which we refer to as abstract avatars and aggregated avatars. Following a discussion on what might motivate interaction designers to develop such avatars, we propose an approach to designing and evaluating specific implementations thereof. We also introduce a testbed environment which we developed in a 3D virtual reality system for the development and evaluation of abstract aggregated avatars.