Context Awareness for Collaborative Learning with Uncertainty Management

In Collaborative Learning, groups of students work together using traditional and computer-based tools or applications. Participants are continuously moving and reorganizing in groups as tasks develop and the contextual information about the physical arrangement of people within groups determines the context of each sub-activity. The electronic environment needs to be in sync with the physical arrangement of the groups, but providing group context information to computer-based tools cannot effectively be done manually. This paper explores and addresses the problem of automating group awareness in CSCL applications by estimating group arrangements from location sensors and the history of interaction. We derive from case studies the requirements for context-awareness in collaborative learning, focusing on the Jigsaw technique supported by mobile devices. In our prototype system with real users, groups are detected from the location of the students within the classroom. However, this information needs filtering to avoid disturbing interruptions caused by uncertain location measures. A three- phase filtering strategy is proposed to manage uncertain contextual information by identifying sources of uncertainty, representing uncertain information, and determining how to proceed. Validation with experimental data shows the usefulness of introducing mobile devices with group-supporting applications that incorporate automatic group awareness. Results show that by managing uncertainty in the estimation of location, group membership information becomes reliable enough to satisfy the need for supporting collaborative learning with applications that are automatically group-aware, without introducing extra burdens or interruptions.