Fuzzy Model for Safety Evaluation Process of New and Old Roads

Abstract

More than 50% of traffic fatalities occur on two-lane rural roads, and more than half of these fatalities occur on curved roadway sections. A large body of research can be used to analyze and evaluate the fundamental relationships between accident situation, highway geometric design, driving behavior, and driving dynamics. These factors form the basis for the development of three quantitative safety criteria used to evaluate the hazards of two-lane rural roads with respect to new designs; redesigns; restoration, rehabilitation, or resurfacing projects; and existing alignments. The safety criteria support the design engineer in classifying new or old roadway sections according to good (sound), fair (tolerable), and poor (dangerous) design practices. On the basis of observation of the actual variation in the accident rate with respect to road alignment, a fuzzy model was developed to classify roadway elements by using these safety criteria to obtain a more careful evaluation of highway design inconsistencies. For each criterion, the inconsistencies were included in three fuzzy sets (good, fair, poor) with differing degrees of membership. By defining linear membership functions, it was possible to obtain good results to classify road sections and then to determine a prioritization scale of maintenance interventions. The procedure can be applied to large databases of road networks to identify the more dangerous design elements that need interventions to improve highway safety and to allocate resources under limited budget conditions.