Concrete barriers are often installed to separate opposing directions of traffic at hazardous horizontal curves on undivided highways in order to avoid frequent types of crashes such as head-on collisions. Whereas this treatment is effective for preventing head-on crashes, they reduce the available sight distance (ASD). The estimation of sight distance is usually performed under a two-dimensional (2D) approach, despite the three-dimensional (3D) nature of facilities. Radius curve, clearance and trajectory location on cross section are the only variables considered under the planar approach, ignoring the spatial component. To overcome this limitation, this study applies a 3D sight distance estimation method, where the sight distance reduction caused by installing barriers is evaluated, contemplating their shape and position in a detailed fully 3D model. The procedure facilitates incorporating mass standardized constructive elements into the highway model. This method consists of a geoprocessing model on a geographic information system (GIS), where the impact of the terrain, the roadway itself, and the road furniture on sight distance is broached. The results evidenced that the ASD outcomes were different under either approach. In addition, the ASD fell below the required stopping sight distance (SSD) of the standards under either approach. A potential application of this method is the analysis of the overall safety effects and the subsequent evaluation of crash modification factors associated to this safety measure.

traffic safety, sight distance, median barriers, 3D modelling, geographic information system