How to effectively identify traffic bottlenecks on freeway and accordingly implement targeted countermeasures remains a critical issue for traffic management and control. With the increasing development of GPS-embedded smartphone navigation, vehicle trajectory data collected in a crowdsourcing way provides a means of constructing spatiotemporal diagrams to support the decision-making process for traffic authorities. To this end, this study presents a comparative case study by comparing two technical methods, i.e., wavelet transform and image processing, which enable to facilitate the identification of recurrent traffic bottlenecks on freeways using vehicle trajectory data. The data utilized were 45-day probe vehicle data collected on urban expressways of Beijing during January and February, 2015. The validation results by referring to field bottlenecks show that image processing method outperforms wavelet transform in terms of estimation accuracy.

CICTP