Human-Centered Safety Analysis of Prospective Road Designs

One of the most important issues in road safety management is the lack of reliable methods for predicting the likelihood of accidents. Road safety assessment systems have been developed; however, these systems only employ historical or retrospective analyses, and the human factor element is weak or missing. Effective safety management requires both holistic and prospective viewpoints, with human factors having an intrinsic role. The main goal of this paper is to contribute toward that need through the application of Bayesian belief networks and road traffic simulation for validating the safety requirements of prospective road designs. The theoretical platform of the method is the concepts of human performance and mental workload and how these affect accident likelihood. This paper presents a novel method and a tool that integrates these two mature technologies, for assessing the safety performance of road designs before they are developed. A case study is included that illustrates the application of the method and tool.