Desenvolvimento de método para estimar o nível de serviço em interseções do tipo rotatória alongada não vazada localizadas em rodovias de pista simples

Abstract

In Brazil, the elaboration of intersection designs takes place from the precepts contained in manuals prepared by official government agencies. However, some types of intersections do not include these documents, called in the literature as alternative intersections, the elongated roundabout (ER) being one of these cases. The main world reference regarding road capacity assessment and level of service, the Highway Capacity Manual (HCM), widely used in Brazil, does not provide a method to determine the level of service of this type of intersection, as well as in the national and international literature. This makes it impossible to evaluate an operational evaluation of this type of intersection and its correct dimensioning to meet the specific conditions of traffic flow streams. In this context, the present work proposed a method to estimate the level of service of this type of intersection on two-lane highways. For that, traffic data on Brazilian roads were collected to obtain a diversity of traffic flows and to characterize the desired speeds in the weaving segment and the adjacent segments of the intersection. After performing a sensitivity analysis, the need to calibrate the behavioral parameters of the Aimsun Next traffic simulation software was not verified. Thus, Aimsun Next was used to create several hypothetical scenarios, in terms of traffic and road geometry, which allowed the extraction of models to obtain the chosen performance measures: the average delay (di) and the time of extra distance travel (EDTT). After verifying the correlations between these performance measures and potential explanatory variables, some models were generated, using linear regression, to estimate di and EDTT. The validation of the proposed method was carried out, comparing the results of the performance measures obtained by the developed models and by the data collected in the field, having observed that these two sets of data were very close.