Datensatz

Zusammenfassung

We investigated the usability of vestibular translatory motion cues in self stabilization. Six blindfolded observers were asked to control
the pitch or roll axis of a Stewart motion platform which followed the
movements of a virtual inverse pendulum. Specifically, we varied the
turning point of the pendulum with respect to the observer's
head. Shifting the turning point along the body axis leaves the
vestibular rotational cues identical while changing only the
translatory component of the motion.

We started with three hypotheses:

I) If the translatory components do not affect stabilization,
performance should be equal for all conditions.

II) If translation does affect stabilization, then increasing the
distance between head and turning point should systematically alter
performance.

III) In posture control, humans always rotate (tilt) around points
below the head and therefore those conditions should result in better
performance.

Subjects controlled the pendulum with a virtual force proportional to
the deflection of a joystick (acceleration based control). Five
different turning point heights with respect to the head (h=0.0m,
±0.6m, and ±1.2m) were tested in trials that lasted 120 seconds
each. To exclude learning artifacts, observers were trained in random
order at all heights four times. Observers typically reached a stable
performance after three of the four blocks (i.e., after about 30
minutes of training).

Surprisingly, the final performance (absolute error and variability)
was best for the condition where the turning point was 0.6m above the
head. Performance decreased with increasing distance from this
optimum. Therefore, none of our hypotheses seem to be true. These
results might have important implications for training in helicopter
simulators.