The Impact of Road Characteristics Assessment on The Traffic Accidents Occurrence

Abstract

This paper proposes a model for assessment of driver’s propensity toward traffic accidents based on the implementation of fuzzy logic. The proposed Fuzzy inference system is based on three input variables and one output. Input variables are: The assessment of dangerous places on the observed road section (nine dangerous spots are selected; however, the scores from these nine spots are sublimated in one variable), Assessment of road characteristics (seven road characteristics are chosen; however, the scores from these seven spots are sublimated in one variable) and Frequency (Input variable Frequency is based on the criterion how many times a week or a month the examinee drives on the observed road section). Output variable is The number of accidents that a driver had experienced. The model is tested on the sample of 305 drivers and most of them are professional drivers. The results are perceived through the cumulative error that fuzzy inference system makes in description of empirical data.