Investigating a method for rating Sri Lankan roads through identifying the factors affecting road safety

Abstract

Road safety assessment is imminent to reduce road accidents in Sri Lanka. The existing road safety assessments in other countries are inapplicable to Sri Lankan Road conditions as there are significant differences between the road conditions of Sri Lanka and other countries. This study aims to identify the governing roadside elements which are influential to the cause of accidents and can be used for road star rating. The study was conducted in three districts of Sri Lanka: Anuradhapura, Polonnaruwa and Kurunegala. The data for road characteristics were obtained through the Google Earth Web engine whereas accident data were collected from Sri Lanka Police. The variables such as road condition, road length, road width, delineation, shoulder condition, footpath, vehicle parking length, road divide status, roadside objects length, number of lanes, number of intersections, number of pedestrian crossings, and number of bus bays were analyzed to develop a Negative Binomial regression model considering the number of accidents as the dependent variable. The results reveal that six variables: number of lanes, road condition, number of intersections, road divide status, road section length and width of lane are significant towards the occurrence of accidents. Moreover, the results demonstrate the relationship between the road characteristics and the accident number which is crucial in road designing in order to reduce road accidents. The findings affirm the possibility in developing a road safety rating mechanism for Sri Lankan streets to standardize the road network with the international standards while enhancing the road conditions with reflecting to the required safety levels. KEYWORDS: Road safety, Road star rating, Sri Lankan roads, Negative binomial regression