A Multiple Strategy Approach to Temporal Range Estimation

Heiko Hecht
Man Vehicle Lab
Massachusetts Institute of Technology

Recent work on time-to-contact (TTC) and time-to-passage (TTP) estimation has revealed conflicting evidence. On the one hand, when tau information is available, observers often fail to use it properly. For instance, absolute size of the object, rotation, and contrast all interfere with estimation accuracy. On the other hand, judgments are remarkably robust when invariant information is no longer available. Observers seem to adjust to many disturbances, such as degraded displays, by relying on extra-sensory information.

6 Theses:

1. The empirical facts rule out a single mechanism, such as a tau-processor. Only multiple mechanisms can account for the data.

2. A two stage process (fast tau estimate that is later modified by a cognitive evaluation) as suggested by Tresilian, explains it all but is not falsifiable.

3. TTC-estimates are strongly task-dependent.

4. The visual system has a tool-box of strategies, from which it chooses depending on the task at hand.

5. The visual system is lazy, that is it always picks the "cheapest" strategy it can get away with to solve the task.

6. It is our job to identify the strategies and the criteria for strategy-selection.