Traffic4cast at NeurIPS 2022 – Predict Dynamics along Graph Edges from Sparse Node Data: Whole City Traffic and ETA from Stationary Vehicle Detectors

Moritz Neun, Christian Eichenberger, Henry Martin, Markus Spanring, Rahul Siripurapu, Daniel Springer, Leyan Deng, Chenwang Wu, Defu Lian, Min Zhou, Martin Lumiste, Andrei Ilie, Xinhua Wu, Cheng Lyu, Qing-Long Lu, Vishal Mahajan, Yichao Lu, Jiezhang Li, Junjun Li, Yue-Jiao Gong, Florian Grötschla, Joël Mathys, Ye Wei, He Haitao, Hui Fang, Kevin Malm, Fei Tang, Michael Kopp, David Kreil, Sepp Hochreiter
Proceedings of the NeurIPS 2022 Competitions Track, PMLR 220:251-278, 2022.

Abstract

The global trends of urbanization and increased personal mobility force us to rethink the way we live and use urban space. The Traffic4cast competition series tackles this problem in a data-driven way, advancing the latest methods in machine learning for modeling complex spatial systems over time. In this edition, our dynamic road graph data combine information from road maps, $10^{12}$ probe data points, and stationary vehicle detectors in three cities over the span of two years. While stationary vehicle detectors are the most accurate way to capture traffic volume, they are only available in few locations. Traffic4cast 2022 explores models that have the ability to generalize loosely related temporal vertex data on just a few nodes to predict dynamic future traffic states on the edges of the entire road graph. In the core challenge, participants are invited to predict the likelihoods of three congestion classes derived from the speed levels in the GPS data for the entire road graph in three cities 15 min into the future. We only provide vehicle count data from spatially sparse stationary vehicle detectors in these three cities as model input for this task. The data are aggregated in 15 min time bins for one hour prior to the prediction time. For the extended challenge, participants are tasked to predict the average travel times on super-segments 15 min into the future – super-segments are longer sequences of road segments in the graph. The competition results provide an important advance in the prediction of complex city-wide traffic states just from publicly available sparse vehicle data and without the need for large amounts of real-time floating vehicle data.

Cite this Paper

BibTeX
@InProceedings{pmlr-v220-neun23a, title = {Traffic4cast at NeurIPS 2022 – Predict Dynamics along Graph Edges from Sparse Node Data: Whole City Traffic and ETA from Stationary Vehicle Detectors}, author = {Neun, Moritz and Eichenberger, Christian and Martin, Henry and Spanring, Markus and Siripurapu, Rahul and Springer, Daniel and Deng, Leyan and Wu, Chenwang and Lian, Defu and Zhou, Min and Lumiste, Martin and Ilie, Andrei and Wu, Xinhua and Lyu, Cheng and Lu, Qing-Long and Mahajan, Vishal and Lu, Yichao and Li, Jiezhang and Li, Junjun and Gong, Yue-Jiao and Gr\"otschla, Florian and Mathys, Jo\"el and Wei, Ye and Haitao, He and Fang, Hui and Malm, Kevin and Tang, Fei and Kopp, Michael and Kreil, David and Hochreiter, Sepp}, booktitle = {Proceedings of the NeurIPS 2022 Competitions Track}, pages = {251--278}, year = {2022}, editor = {Ciccone, Marco and Stolovitzky, Gustavo and Albrecht, Jacob}, volume = {220}, series = {Proceedings of Machine Learning Research}, month = {28 Nov--09 Dec}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v220/neun23a/neun23a.pdf}, url = {https://proceedings.mlr.press/v220/neun23a.html}, abstract = {The global trends of urbanization and increased personal mobility force us to rethink the way we live and use urban space. The Traffic4cast competition series tackles this problem in a data-driven way, advancing the latest methods in machine learning for modeling complex spatial systems over time. In this edition, our dynamic road graph data combine information from road maps, $10^{12}$ probe data points, and stationary vehicle detectors in three cities over the span of two years. While stationary vehicle detectors are the most accurate way to capture traffic volume, they are only available in few locations. Traffic4cast 2022 explores models that have the ability to generalize loosely related temporal vertex data on just a few nodes to predict dynamic future traffic states on the edges of the entire road graph. In the core challenge, participants are invited to predict the likelihoods of three congestion classes derived from the speed levels in the GPS data for the entire road graph in three cities 15 min into the future. We only provide vehicle count data from spatially sparse stationary vehicle detectors in these three cities as model input for this task. The data are aggregated in 15 min time bins for one hour prior to the prediction time. For the extended challenge, participants are tasked to predict the average travel times on super-segments 15 min into the future – super-segments are longer sequences of road segments in the graph. The competition results provide an important advance in the prediction of complex city-wide traffic states just from publicly available sparse vehicle data and without the need for large amounts of real-time floating vehicle data.} }
Endnote
%0 Conference Paper %T Traffic4cast at NeurIPS 2022 – Predict Dynamics along Graph Edges from Sparse Node Data: Whole City Traffic and ETA from Stationary Vehicle Detectors %A Moritz Neun %A Christian Eichenberger %A Henry Martin %A Markus Spanring %A Rahul Siripurapu %A Daniel Springer %A Leyan Deng %A Chenwang Wu %A Defu Lian %A Min Zhou %A Martin Lumiste %A Andrei Ilie %A Xinhua Wu %A Cheng Lyu %A Qing-Long Lu %A Vishal Mahajan %A Yichao Lu %A Jiezhang Li %A Junjun Li %A Yue-Jiao Gong %A Florian Grötschla %A Joël Mathys %A Ye Wei %A He Haitao %A Hui Fang %A Kevin Malm %A Fei Tang %A Michael Kopp %A David Kreil %A Sepp Hochreiter %B Proceedings of the NeurIPS 2022 Competitions Track %C Proceedings of Machine Learning Research %D 2022 %E Marco Ciccone %E Gustavo Stolovitzky %E Jacob Albrecht %F pmlr-v220-neun23a %I PMLR %P 251--278 %U https://proceedings.mlr.press/v220/neun23a.html %V 220 %X The global trends of urbanization and increased personal mobility force us to rethink the way we live and use urban space. The Traffic4cast competition series tackles this problem in a data-driven way, advancing the latest methods in machine learning for modeling complex spatial systems over time. In this edition, our dynamic road graph data combine information from road maps, $10^{12}$ probe data points, and stationary vehicle detectors in three cities over the span of two years. While stationary vehicle detectors are the most accurate way to capture traffic volume, they are only available in few locations. Traffic4cast 2022 explores models that have the ability to generalize loosely related temporal vertex data on just a few nodes to predict dynamic future traffic states on the edges of the entire road graph. In the core challenge, participants are invited to predict the likelihoods of three congestion classes derived from the speed levels in the GPS data for the entire road graph in three cities 15 min into the future. We only provide vehicle count data from spatially sparse stationary vehicle detectors in these three cities as model input for this task. The data are aggregated in 15 min time bins for one hour prior to the prediction time. For the extended challenge, participants are tasked to predict the average travel times on super-segments 15 min into the future – super-segments are longer sequences of road segments in the graph. The competition results provide an important advance in the prediction of complex city-wide traffic states just from publicly available sparse vehicle data and without the need for large amounts of real-time floating vehicle data.
APA
Neun, M., Eichenberger, C., Martin, H., Spanring, M., Siripurapu, R., Springer, D., Deng, L., Wu, C., Lian, D., Zhou, M., Lumiste, M., Ilie, A., Wu, X., Lyu, C., Lu, Q., Mahajan, V., Lu, Y., Li, J., Li, J., Gong, Y., Grötschla, F., Mathys, J., Wei, Y., Haitao, H., Fang, H., Malm, K., Tang, F., Kopp, M., Kreil, D. & Hochreiter, S.. (2022). Traffic4cast at NeurIPS 2022 – Predict Dynamics along Graph Edges from Sparse Node Data: Whole City Traffic and ETA from Stationary Vehicle Detectors. Proceedings of the NeurIPS 2022 Competitions Track, in Proceedings of Machine Learning Research 220:251-278 Available from https://proceedings.mlr.press/v220/neun23a.html.

Related Material