Field Evaluation of Traffic Performance Measures for Two-Lane Highways in Spain

Abstract

[EN] Two-lane highways operation has been extensively studied. Many of these studies state that the current Highway Capacity Manual (HCM) procedure is difficult to measure in the field. Several promising alternative measures have been proposed which are easy to measure in field, such as the follower density, percent impeded or freedom of flow. Nevertheless, some of these measures are based on hypotheses that may only be applicable to local driver behavior. Moreover, none of the field studies has compared the correlation between all the proposed performance measures and traffic variables. The present field study calibrates and evaluates ten performance measures in Spanish two-lane highways. The data was collected using video recordings in 10 sites on two-lane rural highways in Spain. Observed two-way traffic volumes ranged from 120 to 1,000 veh/h and traffic flows were mainly balanced. From this data, time headways, average travel speed and platooning variables were calculated. The studied performance measures included: average travel speed; average travel speed of passenger cars; percent free-flow speed; percent free-flow speed of passenger cars; percent followers; follower density; percent impeded; average platoon length; traffic intensity; and, freedom of flow. The results indicated that the follower density had the strongest correlation with traffic variables, with a coefficient of determination of 94 %. The estimations were compared with previous models and they were alike within their observation range. The second best performance measure was the percent followers and the estimates were very similar to the models in Finland. The 2010 HCM overestimated the percent followers at low traffic flows, which could indicate that the extrapolation of medium-high traffic volume driver behavior was not too accurate at our observation range. The other platooning-related variables had lower correlations, while the speed-related measures presented the weakest correlation with traffic variables.