1987 Volume 14 Issue 21 Pages 65-80
This work intends to examine the effectiveness of a new approach, which is more fundamental than foregoing studies, to an analysis of accidents at railroad crossings. As a case study we deal with car accidents which occurred at single-track open crossings of Japanese National Railway. By applying the quantification method Types II and III analyses were made with a main focus on the type of accidents and from viewpoints of various human and physical factors. As a consequence it was apparent that the human factors gave some information, but was not much efficient. With respect to the physical factors, however, several important points for improvement of structure of crossings were suggested for the “front-crossing” accident, the most probable type of accident. Moreover it was suggested that the predictability of occurrence of accident, that is, a discrimination of types of accidents (front-crossing, other types, and no accident) was possible. Neverthless, considering that the number of samples for crossings with no accident were rather few, we must admit that the present model for the discrimination may not be so effective.