Traffic safety has long been a serious concern for people living in modern cities worldwide. In particular, urban intersections are locations where traffic accidents occur very frequently, and proactive assessment of accident risk is required for taking some particular countermeasures. Human Reliability Analysis (HRA) is a sophisticate method to evaluate human error probabilities and it is widely used in various industries, but its application to accident risk in urban intersections is not straightforward. This paper discusses how to apply a second-generation HRA method to assessing accident risk in urban intersections. Contextual factors that affect driver reliability were identified, and then an assessment method of these factors for particular situations in driving through intersections was given. Field data were collected and analysed for five intersections in the Tokyo metropolitan area with relatively high accident frequencies. The obtained result was compared with accident statistics, and it was demonstrated that the proposed method could well estimate accident risk in urban intersections.
Abstract