Abstract:For an objective evaluation of roadside safety, assessment model of roadside environment objective safety was established based on the probability of run-into-roadside and the roadside objective characteristics. Firstly, the indicators of the various objective characteristics that affect roadside safety were hierarchically divided, and each indicator safety level were determined on the basis of domestic and overseas research results and certain statistics data. The risk index of the roadside objective characteristics was calculated by fuzzy judgment. Then, the probability of run-into-roadside which consider objective indicators about traffic composition、highway alignment、subgrade width and pavement condition was calculated by Bayesian networks. Finally, based on planar region which was composed by the risk index as X-axis and the enforcement probability as Y-axis, roadside objective safety level was divided by comprehensive evaluation indicator. The results can provide decision for roadside safety improvement program.