In this paper, we compare the efficiency in information transfer and user acceptance between the traditional humanoid robot and the projector robot.

In this paper, we investigated the effect of relative difference between the virtual object and tangible prop in its size and shape in terms of usability. We have found the obvious fact that in general that manipulation is more efficient with equally sized/shaped prop and virtual objects. In addition, when decoupled, the size difference factor did not matter. While an additional experiment (in progress) is needed to confirm the true effect of shape difference, we posit that prop design should concentrate on representing the critical shape features for a given class of objects the prop is to represent. Humans are adept at recognizing and identifying objects even if there are shown at different scales and in different angles. This ability, called ldquoconstancyrdquo is weaker in the dimension of shapes, e.g. compared to sizes, which is another reason to suspect shape to be a more critical factor for effective prop design. Another significant factor, in the design of props (not treated in this paper) is the prop to virtual object alignment (which can affect the task performance in terms of finding and feeling for a stable grasp). For instance, possible choices for spatially registering a prop to a virtual object (or vice versa) can be about their respective center of gravity, about the chosen surface, etc. Future experiments and prop design method will have to take this into account as well.