Scaling of relative velocity between vehicles

Data are presented on the ability of drivers to perceive and scale the relative velocity between their own and a lead vehicle. Experiments were carried out on four groups of subjects using Ekman's ratio-rating method. Only when the subtended angular velocity of the lead vehicle exceeded about 0.003 rad/s were the subjects able to scale the relative velocity. The threshold subtended angular velocity obtained in the experiments was very much affected by the ability of subjects to use the concept of a ratio-engineering students found this a simpler task than did subjects from the general population. The result of this was that the values used by engineering students were closer to the real values. The relative velocity was perceived non-linearly, with a Stevens' power law exponent of about 0.8. It was found that linear models gave as good a fit to the data. The implications of the results of the experiments are (i) traffic flow models that include human visual characteristics must consider the ''dead zones'' in response produced by thresholds of subtended angle change and subtended angular velocity; (ii) it may be necessary to consider the non-linear relationship between perceived relative velocity and actual relative velocity; (iii) in overtaking, the driver will not be able to scale the speed of the oncoming vehicle as the subtended angular velocity will be below the threshold value at the time of making the decision to overtake. Copyright (C) 1996 Elsevier Science Ltd.