Methods on dynamic delineation of intersections based on floating car data

Estimation of travel time in the analysis of network performance has been suffering from low accuracy due to errors of vehicle speed evaluation, which is affected by the way that intersection areas in the network are delineated. In this paper, two methods, the density-based and the deceleration point-based, were proposed for the purpose of dynamically delineating the intersection based on Floating Car Data (FCD), which is utilized to identify the speed inflection of vehicles passing through intersections and make analysis of traffic delay at intersections. Analysis results of archived FCD from Nanjing showed that the proposed methods could effectively reveal the time-dependent law of traffic flow changes at intersections and thus could achieve higher accuracy in estimating control delay, as well as identifying traffic condition compared to the static delineation method.