Offor, Kennedy John
Wang, Peng
Mihaylova, Lyudmila S
2019-10-02
<p>In the commonly used Kriging approaches, the covariance function depends only on the separation distance irrespective of the traffic at the considered locations. A key limitation of such an approach is its inability to capture well the traffic dynamics and transitions between different states. This paper proposes a Bayesian Kriging approach for the prediction of urban traffic. The approach can capture these dynamics and model changes via the covariance matrix. The main novelty consists in representing both stationary and nonstationary changes in traffic flows by a discriminative covariance function conditioned on the observation at each location. An advantage of the approach is that it can represent congested regions and interactions in both upstream and downstream areas. Experiment carried out with real data from Santander, Spain shows that RMSE of our method outperforms traditional Kriging method by 8.4%.<br>
</p>
Funder: Tertiary Education Trust Fund (TETFund), Nigeria
https://doi.org/10.5281/zenodo.3470240
oai:zenodo.org:3470240
eng
Zenodo
https://zenodo.org/communities/tetfund
https://zenodo.org/communities/eu
https://doi.org/10.5281/zenodo.3470239
info:eu-repo/semantics/openAccess
Creative Commons Attribution 4.0 International
https://creativecommons.org/licenses/by/4.0/legalcode
SDF Symposium, 13th Sensor Data Fusion Symposium: Trends, Solutions, and Applications, Bonn, Germany, 15-17 October 2019
particle filter, traffic prediction, Kriging, Bayesian inference, Gaussian Process
Multi-Model Bayesian Kriging for Urban Traffic State Prediction
info:eu-repo/semantics/lecture