press to zoom


Throwing is a fundamental motor skill that is difficult to do precisely with physical objects, but even harder for virtual projectiles due to a lack of physical affordances. We explore the challenges in designing the experience of augmented reality (AR) throwing by asking four participants to throw a virtual ball in an AR environment, finding that individual differences in virtual throwing occur with throwing gesture induced angle and speed differences. Informed by these findings, we aim to explore whether we can improve the naturalness of virtual throwing for users. We present a personalized autocorrective model that leverages the previous throws of the user to gradually improve their throwing accuracy in AR, followed by preliminary testing to evaluate our initial attempts that will inform future work.


In Progress