Abstract
Establishment of useful indices to evaluate the risk of a rear-end collision is a vital issue for developing a rear-end collision avoidance assistance system. Previous researches showed that the two rear-end collision risk indices, named RP (Risk Perception) and KdB_c, were effective for identifying the limit of no-brake situations under restricted traffic conditions. In order to use the indices appropriately, it is necessary to understand to what extent the indices reflect the risk of a rear-end collision. This paper investigates driver brake timing for avoiding a rear-end collision by analyzing real world driving data on general roads. The driving data contains wide varieties of traffic conditions. The results show that the two indices are effective when the velocity of the preceding vehicle is not less than 30km/h. However, there are situations in which non-emergency braking is done beyond the threshold values of the indices. This paper mentions possible reasons for the occurrence of those situations based on driver behavior analyses.
