The tau hypothesis predicts that observers initiate collision avoidance actions based on an integration of speed and distance. We constructed an experiment to test whether human subjects are sensitive to the critical boundary of imminent collision, even when this boundary is not defined by an integration of speed and distance. Subjects were asked to perform a simulated flight towards a barrier surface and initiate an evasive ballistic ascent response at the last possible moment before collision. The between-subjects manipulation was the flight ascent dynamics. In one flight dynamic condition, the boundary of collision was defined by an invariant distance to arrival across the different forward speed conditions. In the other flight dynamics condition, the boundary was defined by an invariant time to arrival across forward speeds. Subject performance in the two groups indicated a sensitivity to the appropriate information defining the critical boundaries, although there was also a conservative distance bias and response variability that increased with forward speed. This bias is explained as a functional scaling of the response with the response variability present.
Available at: http://works.bepress.com/john_flach/115/