Interaction Prediction for Content Synchronization of Net-based Shared Workspaces

Digital collaborative environments enable spatially separated users to access and modify shared data over network. However, transmission delays of the network lead to inconsistent data and reduce the efficiency of collaboration due to interaction conflicts. In this paper, we present a predictive screen-locking algorithm to avoid interaction collisions on net-based shared interactive screens. A model-based predictor calculates the user's next interaction given his past one. The algorithms locks critical objects to the remote station which is less likely to interact with the object. Although the predictor continuously adapts to the user's interaction behavior, an initial interaction model is needed when the collaboration session is started. Hence, we deduce a reasonable, probabilistic interaction model from a large screen collaboration user study.