Evaluating Measures of Effectiveness for Quality of Service Estimation on Two-Lane Rural Highways

The level of service on two-lane rural highways can be estimated using the percent time spent following (PTSF) and average travel speed (ATS). Because PTSF cannot be directly observed in the field, the literature suggests alternative measures of service (MOS) that can be obtained from traffic stream parameters. The objective of this paper was to analyze measures of service that could adequately describe the quality of service on two-lane rural highways. Traffic data collected on 16 road segments in the state of São Paulo were used to calibrate and validate the traffic simulation software CORSIM by means of a genetic algorithm. The recalibrated CORSIM was used to create a synthetic set of traffic data that comprise an extensive range of traffic flow and road geometry patterns. Nine alternative measures of service were analyzed, and models that relate these measures of service to directional traffic flow were fitted using a synthetic traffic data set. Comparisons between the values obtained from these models and the values obtained from the field indicated that two measures of service, density and follower density, could be used to create level of service criteria for two-lane rural highways in Brazil.