An automatized signal timing detection at signalized intersections with detailed vehicle trajectory data

Abstract

Signalized intersections are a fundamental part of urban networks. Their understanding is crucial to identify congestion patterns, queues, delays, and safety issues in local and network level. In this work, we analyze multimodal vehicle trajectories and propose a methodology to extract the signal timing schedule of an intersection using the pNEUMA dataset. In addition, we combine the available information from OpenStreetMap (OSM) to map match the trajectories to the underlying network and to identify more accurately the location of traffic signals. Then the methodology to extract the signal timing schedule of an intersection consists of the following steps: i) critical movements identification, ii) computation of crossing times at the traffic signals, iii) cycle length detection and iv) phase length of each critical movement. Results show that by using the OSM data, the methodology can then be applied to any intersection in the network and provide critical information in macroscopic level.