Citation

Tenkink E, van der Horst ARA. Accid. Anal. Prev. 1990; 22(3): 229-239.

Affiliation

TNO Institute for Perception, Soesterberg, The Netherlands.

Copyright

(Copyright © 1990, Elsevier Publishing)

DOI

unavailable

PMID

2393471

Abstract

The behavior of car drivers at two Dutch railroad grade crossings with automatic flashing warning lights was analyzed. Car drivers were videotaped while approaching either the red flashing lights or the white flashing "safe"-signal. Approach speeds, positions, and time intervals were semiautomatically measured from videos of more than 900 drivers: 660 while confronted with the red lights and 272 while passing the white light. Of the latter group, head movements during the approach to the crossing were also registered. Red light compliance was relatively good, as no driver was observed to cross later than 6 seconds after the onset of the red lights, despite train-arrival times of well over 60 seconds. The level of red light compliance was further quantified in terms of both the deceleration and time-to-stopping-line as accepted by drivers. From a comparison with earlier research on red light compliance at signalized road intersections it appeared that red light compliance was better at railroad crossings than at road crossings. It is concluded that faulty red light compliance is not a major cause for car-train accidents and that emphasis should be placed on the ability of the present device to attract attention and to signal unambiguously. The high degree of compliance also causes unexpected driver actions, such as emergency braking and hesitations. A yellow phase may reduce these problems. Some drivers tended to proceed immediately after a train had cleared the road instead of waiting for the end of the red signal (typically some 3 to 5 seconds after the train had passed). This tendency might reveal a major cause of dramatic errors when a second train is approaching. Immediate extinction of the red signal is suggested, or even better, a separate signal to announce the arrival of the second train. Behavior during the white signal phase also showed indications of uncertainty. In some 10% of cases drivers tended to decelerate more strongly than necessary and to make extra head movements. It is recommended that the present white flashing signal be reconsidered.