Multi-pen Sketch Recognition in a Learning Environment

Virtual physics environments are becoming increasingly popular as a teaching tool for high school level mechanical physics. While useful, these tools often offer a complex user interface, lacking the intuitive nature of the traditional whiteboard. Furthermore, the systems are often too advanced to be used by novice students for further experimentation. In this paper we describe a physics learning environment using multicolour sketch recognition techniques on digital whiteboards. The recognition system is based on a combination of Support Vector Machines and rule based methods. By assigning the various drawing modes to different physical drawing pens, we can resolve several ambiguities appearing in single pen sketching interfaces. Moreover, we argue that we can reduce the cognitive load of the user by exploiting the physical realisation of drawing modes in the form of drawing pens, instead of using textual descriptions of the modes on the screen. The system was tested using a constructive interaction method, with users completing a set task; first in a multi-pen drawing environment, and then for comparison purposes in a singlepen equivalent.