Identification and Prioritization of Accident-Prone Segments using International Roughness Index

During last decades, owing to the increase in a number of vehicles, the rate of accident occurrence grows significantly. Efforts must be made to provide efficient tools to prioritize segments requiring safety improvement and identify influential factors on accidents. This objective of the research was to determine the safety oriented threshold of International Roughness Index (IRI) to recognize Accident-Prone Segments (APSs) using new segmentation method. The modified Floating Fixed-length Segmentation (FFLS) was performed based upon the determined safety oriented IRI threshold with respect to the available literature. Floating fixed-length patterns with lengths of 100, 200 and 500 meters were moved over an entire length of a selected highway to detect segments with IRI values higher than the threshold. To diminish the lack of heterogeneity in characteristics of segments, it was proposed to analyze adjacent road segments with a similar pattern of IRI variation, as a unit.  Owing to the limitation in road maintenance and rehabilitation costs for safety improvement, the entire APSs cannot be treated. Therefore, prioritization and selection of APSs were followed by imposing constraints upon the preservation of different percentages of the highway. Results indicated that the assumed safety oriented threshold of IRI and the modified segmentation method led to correct recognition of segments with high IRI associated with low level of safety. Application of the proposed method using 200-meter floating segment resulted in the shortest length of APSs for safety improvement. The outcomes lead to preserving the most deteriorated segments considering budget constraints. Furthermore, the validation supported the outcomes in which most of the segments were selected from sections with PCI values of 30 or 19. The latter supports the results achieved by the determined IRI threshold and segmentation method. Therefore, considering safety issues as well as maintenance operations would result in optimal use of available budget.