Experimental evaluation of retroreflective markings on trains and grade crossings

Abstract

Every year in the United States, hundreds of accidents occur at grade crossings due to motor vehicles colliding with trains. Furthermore, a large majority of these accidents take place at night in rural areas. One proposed solution to prevent such accidents involves mounting retroreflective material on the sides of trains so that a vehicle's headlights will illuminate the reflectors and make the train more conspicuous. The objective of this research was to determine which train mounted reflector pattern gives an approaching driver the best train recognition. Four reflector patterns based on previous research were selected for this study, and a computer based nighttime driving simulator was developed for this research. In the first experiment, over a thousand scenes containing the view of a road intersection and a grade crossing were displayed to the subject. The subject's recognition of different reflector patterns was recorded and subsequently analyzed using the Receiver Operating Characteristic (ROC) based on Signal Detection Theory (SDT). The second experiment involved a driving task in which the subject encountered numerous grade crossings, and the recognition distance between the train and subject's position was recorded and analyzed.