Improving the Self-Explaining Performance of Czech National Roads

Improving the road network according to the principles of self-explaining roads is a promising way to increase the level of safety; however, there are no universal guidelines on how to measure and improve the self-explaining performance of existing roads. To apply this approach on Czech national roads, the present study was conducted, consisting of five steps: (a) automated segmentation into tangents and horizontal curves; (b) collection of floating car data and calculation of speed; (c) development of multivariate speed models for estimation of speed, including on segments not covered by floating car data; (d) networkwide application of the models and evaluation of speed consistency, such as differences in speeds on tangents and following curves; and (e) identification of substandard curves, and categorization and proposal of optimization for consistent placement of traffic control devices or reconstructions. The paper describes all the steps as well as several checks conducted along the way, such as comparison of profile speed and floating car speed, interpretation of regression models, and validation of predicted speed consistency against long-term average crash frequency. The methodology has been certified for use in practice and will be applied by the Czech national road agency.