This project addresses the challenge of making avatars used in tutoring systems and other learning technologies more engaging and more effective learning guides. One focus is on making the movements of an avatar more lifelike; a second focus is on understanding how to effectively integrate gestures indicating friendliness with gestures used for promoting concept learning. Design of avatar gestures draws on the literature on the roles of gesturing in communication, the literature on grounded and embodied cognition, and the literature on mathematics. The focus of this Cyberlearning: Transforming Education Exploration proposal is on designing avatar gestures for promoting learning of mathematical equivalence. Analysis of the effects of those gestures and extraction of preliminary guidelines for gesture design will lay the groundwork for extending gesturing capabilities of avatars across disciplinary content areas. Research focuses on how to automatically animate an avatar with life-like qualities and the conditions under which adding gestures to a teaching avatar promotes engagement and learning.
Avatars are used extensively in learning technologies to guide learners' actions and to provide advice. But learners can be put off by avatars that are emotionless, and avatars technology does not yet combine gesturing and "talking" in the way good tutors and teachers do as they are trying to help learners learn difficult concepts. This project addresses both of those issues -- giving avatars enough personality so that learners will engage with them more readily and willingly and giving avatars the capability of gesturing in ways that are congruent with the concepts they are expressing. There are a variety of challenges in addressing these issues, some technical, some social, and some cognitive. The project team is addressing all three types of issues, working towards design principles for avatars that move and engage more naturally and that can gesture as people do when explaining difficult content. Enhancing the interactions between computers and people in ways that promote both sustained engagement and learning will improve the effectiveness of intelligent tutoring systems and other learning technologies across disciplines and across the ages of learners.
