This article discusses the issues related to the safety for the transport of dangerous goods by road. Research on accidents in transport unambiguously points to the human factor, which is the most responsible for causing accidents. Determining the causes of driver unreliability in the human−vehicle−environment system requires thorough research. Unfortunately, in this case, experimental research with human involvement is limited in scope. This leaves modeling and simulation of the behavior of the human factor, i.e., the driver transporting dangerous goods. The human being, because of its complexity, is a challenging element to parameterize. The literature presents various attempts to model human actions. Herein, the authors used heuristic methods, specifically fuzzy set techniques, to build a human factor model. In these models, human actions were specified using a verbal or linguistic description. The specificity of the fuzzy sets allowed for “naturally” limiting the “precision” in describing human behavior. The model was built based on the author’s questionnaire and expert research, based on which individual features were selected. Then, the traits were assigned appropriate states. The output parameter of the model was λL—the intensity of human error. The obtained values of the intensity of the accident caused by the driver’s error were implemented into the author’s method of risk assessment. They constituted one of the factors determining the probability of an accident in the transport of dangerous goods, which allowed for determining the optimal route for the transport of these goods characterized by the lowest risk of an undesirable event on the route. The article presents the model’s assumptions, structure, and the features included in the model, all of which have the most significant influence on shaping the intensity of human error. The results of the simulation studies showed a diversified effect of the analyzed characteristics on the driver’s efficiency.
Lomax-Rayleigh distribution: traditional and heuristic methods of estimation
