Abstract
Based on studies of perception of time-to-collision, one might predict that drivers' control of speed and braking depends upon visual expansion fields as defined by Tau-dot. We tested this hypothesis by measuring braking behavior under real-world road condition using a ‘vicarious driving’ task. Observers rode over a 28-mile course as passengers in an instrumented car that recorded speed, application of a simulated brake pedal, and passage of designated roadside landmarks. Observers applied the “brake” at the last moment to prevent collision at the plane of the landmark, at speeds ranging from ∼45 to 110 km/h.
Brake response distances were calculated on the basis of vehicle speed and the time difference between “brake” application and passage through the target plane. Tau-dot values were calculated by the following equation:
Tau-dot = (−1(zD/V^2))
where: z= eye-to-target distance, V=velocity, D=deceleration
These theoretical estimates of perceived time-to-collision were converted to stopping distances.
Results replicated earlier findings that braking-response distances increased as a function speed^2. This supports the hypothesis that drivers can perceive their vehicle's kinetic energy. Although the data fit tau-dot predictions at low speeds [<70 km/h], response distances are consistently shorter than predicted at higher speeds. This discrepancy suggests that other information in addition to tau-dot plays a role in control of vehicle speed and braking. Additional information may be available from non-visual sources (e.g., auditory, proprioceptive, and vestibular correlates of inertial forces) as well as other aspects of the dynamic visual array. Current experiments are testing analogous braking behavior in a fixed-base simulator with a video presentation of the road course, which eliminates non-visual information.