Method to Estimate the Level of Service in Elongated Roundabouts Located on Two-Lane Rural Highways

Abstract

In Brazil, roadway intersections are designed based on the precepts contained in manuals prepared by official government agencies and documents such as the Highway Capacity Manual (HCM). However, some types of intersections, termed alternative intersections, are not included in these documents, such as elongated roundabouts (ERs). This study proposed a method to estimate the level of service of this type of intersection located on two-lane rural highways. For this purpose, traffic data were collected in the field to obtain diverse traffic patterns and to characterize the desired speeds at sites with ERs. The Aimsun Next traffic simulator was used to create several hypothetical scenarios, which allowed the development of models to estimate the chosen performance measures for ER analysis (the average delay, di, and extra distance travel time, EDTT) as a function of traffic and geometry parameters. The validation of the proposed method was performed, and it was observed that the traffic models produced values of di and EDTT close to those found in the field. The results obtained in this work allow the reader to evaluate the operational quality of an intersection constantly implemented in Brazil, an elongated roundabout (ER). No methods were found in the national and international literature aimed exclusively at this type of intersection, which justifies the need for this work. The multilinear models that make up the method serve to estimate performance measures aimed to evaluate the ERs, the average delay, di, and the extra distance travel time, EDTT. To estimate these two performance measures, the user of the proposed method needs to obtain, around the ERs, simple information, such as the traffic flow of the primary road; the traffic flow of the secondary road; the free flow speed; and the percentage of heavy vehicles. The method can reproduce values of these performance measures close to the data collected in the field, considering the conditions observed on Brazilian highways. This may indicate that the method could also be applied in other countries if the proposed multilinear models are appropriately adjusted based on local traffic data.