
The driver’s vision process under nighttime observation conditions is quite complex and depends primarily on their psychophysical factors and lighting conditions. The subjective assessment of lighting conditions conducted by road safety auditors cannot be the basis for verifying the state of pedestrian crossing illumination. As the literature on the subject indicates, every person sees in a slightly different way, and the physiognomic properties of vision change with age and health status.
A number of factors influence the safety of pedestrians at a pedestrian crossing. As research in the field of road safety shows, there are various criteria for evaluating safety, e.g., parameters of accident risk or fatal injury. If an unambiguous answer cannot be defined due to too many input variables or if the changes in assessment factors are non-linear, an attempt can be made to use logic systems, e.g., fuzzy logic algorithms. By “fuzzifying” input values and operating with intervals and weights, an assessment dependent on a number of seemingly independent input variables can be obtained. Thanks to assignment to an assumed interval that best corresponds to the measured quantity, the measurement result is not treated unambiguously. According to the defined change function, the degree of membership in a given interval is determined. As a result of the designer (expert) adopting fuzzy logic rules, a resultant, precise assessment can be obtained on an assumed scale (in the presented case, between 0 and 10 points). The author of this study proposes the use of fuzzy logic algorithms to assess the parameters of pedestrian crossing illumination. It should be emphasized that in the future, it is possible to expand the model with other criteria related to road safety (e.g., pedestrian or motor traffic intensity, geometric visibility, quality of signage, etc.).
In the author’s opinion, the assessment method using operations on fuzzy sets is optimal due to the possibility of incorporating expert knowledge and further evaluation of the model for the purpose of standardizing road safety assessment.
The article contains 26,415 characters.
Source: Czasopismo Logistyka 4/2012







