Freeway Travel Time Information From Input- Output Vehicle Counts: A Drift Correction Method Based on AVI Data

Abstract

Input-output cumulative count curves are a powerful tool to forecast freeway travel times in the very short term, as they rely on the predictive information given by the vehicles' accumulation in the target section. Therefore, they could represent a very appealing method to feed real-time traffic information systems. However, traffic detectors' count drift implies very poor accuracy in the estimation of the vehicles' accumulation, leading to completely unreliable travel time predictions. A drift correction method is necessary. In contrast, several technologies allow accurate direct travel time measurements, like automatic vehicle identification (AVI) or vehicle tracking systems. In all cases, direct travel time measurements are obtained once the vehicle has crossed the target section. This means that the information is representative of the near past traffic conditions, while the objective of real-time information systems is to transmit information about traffic conditions in the near future. In this context, the present paper aims to fuse the information provided by input-output diagrams, obtained from loop detectors, with AVI direct travel time measurements. This fusion allows exploiting the accuracy of the direct measurements to correct the count drift in loop detectors. Then, corrected input-output curves can be used to obtain reliable short-term predictions of travel times. The proposed data fusion method has been applied to a test site in the AP7 freeway near Barcelona, obtaining significantly better results than with the common practices of simply disseminating direct measurements or using spot-speed methods.