Dr. Heiko Hecht
Massachusetts Institute of Technology,
Man-Vehicle Lab
will speak on
Visual Judgement of Collisions and Near Collisions:
The Sorry State of Tau Theory
Abstract:
I will present a critical assessment of an elegant theory and prime
example of ecological optics, namely tau theory. The theory claims that
time-to-contact (TTC) is optically specified by the relative change over
time of local expansion, information to which observers are highly sensitive
(Lee, 1976). Likewise, time-to-passage (TTP) can be optically specified
by a global flow field even in the absence of local expansion or size cues.
Kaiser and Mowafy (1993) demonstrated that observers are in fact sensitive
to this global flow information. Typically, local and global variables
have been studied in isolation and under a set of constancy assumptions
(e.g. constant approach velocity of the object). In realistic situations,
however, both local and global cues are usually present, but constancies
are often violated. We investigated performance in situations where
local but not global expansion cues were disturbed, or where the constant
velocity assumption was violated. TTP/TTC judgments had to be made
based on expanding or non-expanding targets. In some cases, the targets
underwent local rotation, such that global information remained intact
but local information was disturbed. Observers were not able to ignore
local disturbance, even when global information adequately specified TTP.
Also, observers did not appear to utilize acceleration information. Neither
were they able to update TTC estimates when the stimulus was intermittently
occluded. Different remedies for the ailing tau theory will be discussed,
ranging from ad hoc patches to thorough demolition.
The lecture will take place in the Lecture Hall, Room 203, 44 Cummington
St.
on Wednesday, February 21, 2001
at 4:00 pm
Hosted by the
Brain and Vision Research Laboratory