|condition 1 (only translation)|
|condition 2 (translation & rotation)|
Peg in hole task: The task is to insert a peg into a hole. Subjects were asked to move the peg over the hole and insert it into the hole. The tasks were executed under two condotions. The first condition was that the peg's orientation was set as the same orientation as the hole, and the initial position of the peg was randomly located. The peg was inserted into the hole by translations only. The other condition was that the position and orientation were randomly determined. The peg must be manipulated by both translations and rotations. Performance was measured by how many times the peg was inserted within five minutes. Table 2 shows that MagellanTM was the best at condition 1, though its deviation was the worst. The 3-13 device showed the best performace at condition 2, so the rotation interface might be best considering condition 1's result.
Pentominos: Pentominos is originally a puzzle game of thinking how to build a large cube by pieces of pentomino. Subjects were asked to manipulate pentominos in the same way as demonstrated on the screen. Performance is measured by how many pieces of pentomino are placed to the demonstrated places within five minutes.Table 3 shows that MagellanTM deals with the most pieces and 3-D is next.
Despite the prototype device made by our students, the experiments showed that our 3-D input device is comparable to the commercial devices. The ball rotation interface is promising for 3-D rotation input for desktop use. We think 3-D manipulation is dependent on operators' 3-D world sensing and applications. We are planning further field tests from ergonmics.
Balaguer, J. F. and Gobbetti, E ( 1996) "3-D User Interfaces of General-Purpose 3D Animation", COMPUTER, Vol. 29, No. 8, 71-78.
Vince, J( 1995): VIRTUAL REALITY, Addison-Wesley.