Abstract
Information about an object's shape is needed to obtain a stable grip. In virtual environments it is difficult to simulate the stability of the grip in a realistic way. We are interested in whether subjects use visual shape information when haptic feedback is constant. We simulated thirteen 10cm tall elliptical cylinders at eye height. The length of their principal axes varied from 43.6mm to 64mm while keeping the volume constant. The orientation about the vertical axis was either 0, 30 or 60. At the location of the simulated cylinder stood a real cylinder of 52mm diameter of the same height. A mirror prevented subjects from seeing their hands. 12 subjects were instructed to reach and grasp the simulated cylinder with a precision grip. The simulated cylinder disappeared just before they touched the real cylinder. They had to lift the real cylinder and return it. All subjects were aware that the real cylinder they were grasping was always the same. We found that 7 out of 12 subjects systematically adjusted the grip orientation towards the simulated orientations of one of the principal axes. The gain was typically 0.6 except for aspect ratios close to unity. The other subjects always used the same grip orientation. The maximum grip aperture was adjusted to the simulated length of grip axis with a gain of 0.2. In our previous study with real cylinders we found gains of 0.8 and 0.6 for the scaling of the grip orientation and grip aperture, respectively. We conclude that visual shape information is used when haptic feedback is constant, but the effect on the preshaping of the hand is small when compared to the effect for real objects.
NWO MagW 410-203-04