We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:


Current browse context:


Change to browse by:

References & Citations


(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Physics > Physics and Society

Title: Spatial-Temporal Inference of Urban Traffic Emissions Based on Taxi Trajectories and Multi-Source Urban Data

Authors: Jielun Liu, Ke Han, Xiqun (Michael) Chen, Ghim Ping Ong
Abstract: Vehicle trajectory data collected via GPS-enabled devices have played increasingly important roles in estimating network-wide traffic, given their broad spatial-temporal coverage and representativeness of traffic dynamics. This paper exploits taxi GPS data, license plate recognition (LPR) data and geographical information for reconstructing the spatial and temporal patterns of urban traffic emissions. Vehicle emission factor models are employed to estimate emissions based on taxi trajectories. The estimated emissions are then mapped to spatial grids of urban areas to account for spatial heterogeneity. To extrapolate emissions from the taxi fleet to the whole vehicle population, we use Gaussian process regression models supported by geographical features to estimate the spatially heterogeneous traffic volume and fleet composition. Unlike previous studies, this paper utilizes the taxi GPS data and LPR data to disaggregate vehicle and emission characteristics through space and time in a large-scale urban network. The results of a case study in Hangzhou, China, reveal high-resolution spatiotemporal patterns of traffic flows and emissions and identify emission hotspots at different locations. This study provides an accessible means of inferring the environmental impact of urban traffic with multi-source data that are now widely available in urban areas.
Comments: 25 pages, 11 figures
Subjects: Physics and Society (physics.soc-ph); Applications (stat.AP)
Cite as: arXiv:1809.10834 [physics.soc-ph]
  (or arXiv:1809.10834v1 [physics.soc-ph] for this version)

Submission history

From: Xiqun Chen [view email]
[v1] Fri, 28 Sep 2018 03:02:31 GMT (1095kb)

Link back to: arXiv, form interface, contact.