hide
Free keywords:
-
Abstract:
We are investigating the nature and quality of mental spatial representations by studying spatial behaviour in real and virtual environments. Last year we presented data on bearing estimations in a real environment. New bearing data obtained in a corresponding virtual environment have allowed us to study knowledge transfer between real and virtual environments.
From each of eleven familiar locations in the centre of Tübingen, subjects had to estimate the direction of the ten other occluded locations. Whereas the subjects in the outdoor experiment had walked to each subsequent location, in the laboratory the same ten subjects were seated in front of a half-cylindrical projection screen (7 m diameter). On this screen, 180 deg × 50 deg segments of panoramic photographs taken at each location were displayed. Pointing in this virtual environment was accomplished by rotating the projected image until the estimated direction matched that of a demarcation line on the projection screen.
The mean absolute angular pointing error was 11.0 ± 0.3° in the real environment (RE) and 12.9 ± 0.4° in the virtual environment (VE). Mean signed pointing error (a measure for the systematic deviation) varied strongly between places, but was generally smaller when pointing to a place (RE: -10°...13°; VE: -10°...14°) than when pointing from a place (RE: -4°...4°; VE: -5°...5°). The pattern of systematic errors was very similar in both environments and was stable over repetitions of the experiment.
We conclude that: (i) the cognitive maps that underlie our subjects' pointing behaviour were close to the real map but not identical to it; (ii) subjects' pointing behaviour in the laboratory setting was nearly identical to that in the real world, suggesting an effortless transfer of knowledge obtained in the RE to the VE.