Time-to-Collision Judgments of Constant and Non-Constant Velocities: Implications for Rear-End Collisions

Abstract

To avoid rear-end collisions drivers putatively judge the time to collision (TTC) with a lead car. Prior studies typically measured TTC judgments of object motion or constant velocity rather than self motion or acceleration. In contrast, rear-end collisions typically occur when a driver is accelerating or decelerating. With computer-simulations of traffic scenes, we measured TTC judgments of self motion or object motion at constant and non-constant velocities. We also measured judgments at different levels of headway and viewing duration because rear-end collisions often occur with short headways and during driver distractions. Observers typically underestimated TTC when the virtual self accelerated toward a lead car. Further, the accuracy of TTC judgments depended on the parameters of the traffic environment, including viewing duration and headway. The implication is that intelligent collision-avoidance warning systems might be more effective in reducing rear-end collisions under certain traffic conditions than others. We begin to identify such conditions here.