Regimes de fluxo veicular e acidentes de trânsito em interseções semaforizadas de Fortaleza

Abstract

The advent of new technologies for monitoring and controlling traffic allows the development of more robust road safety studies, from traffic data collection at short temporal intervals (1 to 15 minutes), making possible the evaluation of the effect of variables that oscillate within thesame day (vehicle volume, speed, degree of saturation, traffic light cycle) in the frequency and severity of traffic crashes. Traditionally, studies in road safety field are based on entities, assessing the relationship between the annual frequency of crashes and the average annual characteristics of intersections or road segments. However, this work was developed using disaggregated data, applying an approach based on the traffic condition, by grouping similar traffic variables, rather than annual average aspects of each entity. The use of disaggregated data requires a more reliable accuracy of crashes reported time, aiming a trustable association between the crash and its traffic precursor condition. One of the goals of this study was to develop a speed disturbance detection algorithm based on speed camera data, in order to compare the difference between the disturbance time and the time that crashes were reported. The results showed a 19 minutes average time difference and pointed out that crashes with victims presented smaller average differences in comparison to the crashes with only property damages. Using information aggregated in intervals of 15 minutes provided by the SCOOT system, a database with 4.1M traffic conditions was consolidated. A cluster analysis was performed to group similar traffic conditions. The main results of evaluation of the relationship between the clusters and the crashes at urban signalized intersections in Fortaleza were: (i) the occurrence of property damage only crashes is more likely in conditions with high vehicular demands; (ii) the frequency of crashes with victims in high and low vehicle demand conditions is similar, indicating that a portion of the crashes happened due to the possibility of speed development and another due to the exposure and the vulnerability of the users; (iii) the conditions that presented combinations of low vehicle demand + long traffic cycles and high vehicular demand + short traffic cycles were the most critical, especially in terms of crashes with victims occurrence.