A Deep Learning-Based Approach for Train Arrival Time Prediction

Bas Buijse; Vahideh Reshadat; Oscar  Enzing

doi:10.1007/978-3-030-91608-4_22

A Deep Learning-Based Approach for Train Arrival Time Prediction

Bas Buijse, Vahideh Reshadat, Oscar Enzing

Information Systems IE&IS

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

3 Citations (Scopus)

203 Downloads (Pure)

Abstract

Level crossings have a function to let the traffic cross the railroad from one side to the other. In the Netherlands, 2300 level crossings are spread out over the country, playing a significant role in daily traffic. Currently, there isn’t an accurate estimation of the arrival time of trains at level crossings while it plays an important role in traffic flow management in intelligent transport systems. This paper presents a state-of-the-art deep learning model for predicting the arrival time of trains at level crossings using spatial and temporal aspects, external attributes, and multi-task learning. The spatial and temporal aspects incorporate geographical and historical travel data and the attributes provide specific information about a train route. Using multi-task learning all the information is combined and an arrival time prediction is made both for the entire route as for sub-parts of that route. Experimental results show that on average, the error is only 281 s with an average trip time of one hour. The model is able to accurately predict the arrival time at level crossings for various time steps in advance. The source code is available at https://github.com/basbuijse/train-arrival-time-estimator.

Original language	English
Title of host publication	Intelligent Data Engineering and Automated Learning - 22nd International Conference, IDEAL 2021, Proceedings
Subtitle of host publication	22nd International Conference, IDEAL 2021, Manchester, UK, November 25–27, 2021, Proceedings
Editors	David Camacho, Peter Tino, Richard Allmendinger, Hujun Yin, Antonio J. Tallón-Ballesteros, Ke Tang, Sung-Bae Cho, Paulo Novais, Susana Nascimento
Place of Publication	Cham
Publisher	Springer
Chapter	22
Pages	213-222
Number of pages	10
ISBN (Electronic)	978-3-030-91608-4
ISBN (Print)	978-3-030-91607-7
DOIs	https://doi.org/10.1007/978-3-030-91608-4_22
Publication status	Published - 2021
Event	22nd International Conference on Intelligent Data Engineering and Automated Learning, IDEAL 2021 - Manchester, United Kingdom Duration: 25 Nov 2021 → 27 Nov 2021 Conference number: 22 https://ideal-conf.com/ideal2021

Publication series

Name	Lecture Notes in Computer Science (LNCS)
Publisher	Springer
Volume	13113
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Name	Information Systems and Applications, incl. Internet/Web, and HCI (LNISA)
Volume	13113

Conference

Conference	22nd International Conference on Intelligent Data Engineering and Automated Learning, IDEAL 2021
Abbreviated title	IDEAL 2021
Country/Territory	United Kingdom
City	Manchester
Period	25/11/21 → 27/11/21
Internet address	https://ideal-conf.com/ideal2021

Keywords

Deep learning
Multi-task learning
Spatial-temporal neural networks
Train arrival time prediction

Access to Document

10.1007/978-3-030-91608-4_22

FINAL AUTHOR VERSIONFinal author version , 1.26 MB

Cite this

Buijse, B., Reshadat, V., & Enzing, O. (2021). A Deep Learning-Based Approach for Train Arrival Time Prediction. In D. Camacho, P. Tino, R. Allmendinger, H. Yin, A. J. Tallón-Ballesteros, K. Tang, S.-B. Cho, P. Novais, & S. Nascimento (Eds.), Intelligent Data Engineering and Automated Learning - 22nd International Conference, IDEAL 2021, Proceedings: 22nd International Conference, IDEAL 2021, Manchester, UK, November 25–27, 2021, Proceedings (pp. 213-222). (Lecture Notes in Computer Science (LNCS); Vol. 13113), (Information Systems and Applications, incl. Internet/Web, and HCI (LNISA); Vol. 13113). Springer. https://doi.org/10.1007/978-3-030-91608-4_22

Buijse, Bas ; Reshadat, Vahideh ; Enzing, Oscar . / A Deep Learning-Based Approach for Train Arrival Time Prediction. Intelligent Data Engineering and Automated Learning - 22nd International Conference, IDEAL 2021, Proceedings: 22nd International Conference, IDEAL 2021, Manchester, UK, November 25–27, 2021, Proceedings. editor / David Camacho ; Peter Tino ; Richard Allmendinger ; Hujun Yin ; Antonio J. Tallón-Ballesteros ; Ke Tang ; Sung-Bae Cho ; Paulo Novais ; Susana Nascimento. Cham : Springer, 2021. pp. 213-222 (Lecture Notes in Computer Science (LNCS)). (Information Systems and Applications, incl. Internet/Web, and HCI (LNISA)).

@inproceedings{545ea317263a4612b153f0084a72dae8,

title = "A Deep Learning-Based Approach for Train Arrival Time Prediction",

abstract = "Level crossings have a function to let the traffic cross the railroad from one side to the other. In the Netherlands, 2300 level crossings are spread out over the country, playing a significant role in daily traffic. Currently, there isn{\textquoteright}t an accurate estimation of the arrival time of trains at level crossings while it plays an important role in traffic flow management in intelligent transport systems. This paper presents a state-of-the-art deep learning model for predicting the arrival time of trains at level crossings using spatial and temporal aspects, external attributes, and multi-task learning. The spatial and temporal aspects incorporate geographical and historical travel data and the attributes provide specific information about a train route. Using multi-task learning all the information is combined and an arrival time prediction is made both for the entire route as for sub-parts of that route. Experimental results show that on average, the error is only 281 s with an average trip time of one hour. The model is able to accurately predict the arrival time at level crossings for various time steps in advance. The source code is available at https://github.com/basbuijse/train-arrival-time-estimator.",

keywords = "Deep learning, Multi-task learning, Spatial-temporal neural networks, Train arrival time prediction",

author = "Bas Buijse and Vahideh Reshadat and Oscar Enzing",

year = "2021",

doi = "10.1007/978-3-030-91608-4_22",

language = "English",

isbn = "978-3-030-91607-7",

series = "Lecture Notes in Computer Science (LNCS)",

publisher = "Springer",

pages = "213--222",

editor = "David Camacho and Peter Tino and Richard Allmendinger and Hujun Yin and Tall{\'o}n-Ballesteros, {Antonio J.} and Ke Tang and Sung-Bae Cho and Paulo Novais and Susana Nascimento",

booktitle = "Intelligent Data Engineering and Automated Learning - 22nd International Conference, IDEAL 2021, Proceedings",

note = "22nd International Conference on Intelligent Data Engineering and Automated Learning, IDEAL 2021, IDEAL 2021 ; Conference date: 25-11-2021 Through 27-11-2021",

url = "https://ideal-conf.com/ideal2021",

}

Buijse, B, Reshadat, V & Enzing, O 2021, A Deep Learning-Based Approach for Train Arrival Time Prediction. in D Camacho, P Tino, R Allmendinger, H Yin, AJ Tallón-Ballesteros, K Tang, S-B Cho, P Novais & S Nascimento (eds), Intelligent Data Engineering and Automated Learning - 22nd International Conference, IDEAL 2021, Proceedings: 22nd International Conference, IDEAL 2021, Manchester, UK, November 25–27, 2021, Proceedings. Lecture Notes in Computer Science (LNCS), vol. 13113, Information Systems and Applications, incl. Internet/Web, and HCI (LNISA), vol. 13113, Springer, Cham, pp. 213-222, 22nd International Conference on Intelligent Data Engineering and Automated Learning, IDEAL 2021, Manchester, United Kingdom, 25/11/21. https://doi.org/10.1007/978-3-030-91608-4_22

A Deep Learning-Based Approach for Train Arrival Time Prediction. / Buijse, Bas; Reshadat, Vahideh; Enzing, Oscar .
Intelligent Data Engineering and Automated Learning - 22nd International Conference, IDEAL 2021, Proceedings: 22nd International Conference, IDEAL 2021, Manchester, UK, November 25–27, 2021, Proceedings. ed. / David Camacho; Peter Tino; Richard Allmendinger; Hujun Yin; Antonio J. Tallón-Ballesteros; Ke Tang; Sung-Bae Cho; Paulo Novais; Susana Nascimento. Cham: Springer, 2021. p. 213-222 (Lecture Notes in Computer Science (LNCS); Vol. 13113), (Information Systems and Applications, incl. Internet/Web, and HCI (LNISA); Vol. 13113).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

TY - GEN

T1 - A Deep Learning-Based Approach for Train Arrival Time Prediction

AU - Buijse, Bas

AU - Reshadat, Vahideh

AU - Enzing, Oscar

N1 - Conference code: 22

PY - 2021

Y1 - 2021

N2 - Level crossings have a function to let the traffic cross the railroad from one side to the other. In the Netherlands, 2300 level crossings are spread out over the country, playing a significant role in daily traffic. Currently, there isn’t an accurate estimation of the arrival time of trains at level crossings while it plays an important role in traffic flow management in intelligent transport systems. This paper presents a state-of-the-art deep learning model for predicting the arrival time of trains at level crossings using spatial and temporal aspects, external attributes, and multi-task learning. The spatial and temporal aspects incorporate geographical and historical travel data and the attributes provide specific information about a train route. Using multi-task learning all the information is combined and an arrival time prediction is made both for the entire route as for sub-parts of that route. Experimental results show that on average, the error is only 281 s with an average trip time of one hour. The model is able to accurately predict the arrival time at level crossings for various time steps in advance. The source code is available at https://github.com/basbuijse/train-arrival-time-estimator.

AB - Level crossings have a function to let the traffic cross the railroad from one side to the other. In the Netherlands, 2300 level crossings are spread out over the country, playing a significant role in daily traffic. Currently, there isn’t an accurate estimation of the arrival time of trains at level crossings while it plays an important role in traffic flow management in intelligent transport systems. This paper presents a state-of-the-art deep learning model for predicting the arrival time of trains at level crossings using spatial and temporal aspects, external attributes, and multi-task learning. The spatial and temporal aspects incorporate geographical and historical travel data and the attributes provide specific information about a train route. Using multi-task learning all the information is combined and an arrival time prediction is made both for the entire route as for sub-parts of that route. Experimental results show that on average, the error is only 281 s with an average trip time of one hour. The model is able to accurately predict the arrival time at level crossings for various time steps in advance. The source code is available at https://github.com/basbuijse/train-arrival-time-estimator.

KW - Deep learning

KW - Multi-task learning

KW - Spatial-temporal neural networks

KW - Train arrival time prediction

UR - http://www.scopus.com/inward/record.url?scp=85126443027&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-91608-4_22

DO - 10.1007/978-3-030-91608-4_22

M3 - Conference contribution

SN - 978-3-030-91607-7

T3 - Lecture Notes in Computer Science (LNCS)

SP - 213

EP - 222

BT - Intelligent Data Engineering and Automated Learning - 22nd International Conference, IDEAL 2021, Proceedings

A2 - Camacho, David

A2 - Tino, Peter

A2 - Allmendinger, Richard

A2 - Yin, Hujun

A2 - Tallón-Ballesteros, Antonio J.

A2 - Tang, Ke

A2 - Cho, Sung-Bae

A2 - Novais, Paulo

A2 - Nascimento, Susana

PB - Springer

CY - Cham

T2 - 22nd International Conference on Intelligent Data Engineering and Automated Learning, IDEAL 2021

Y2 - 25 November 2021 through 27 November 2021

ER -

Buijse B, Reshadat V, Enzing O. A Deep Learning-Based Approach for Train Arrival Time Prediction. In Camacho D, Tino P, Allmendinger R, Yin H, Tallón-Ballesteros AJ, Tang K, Cho SB, Novais P, Nascimento S, editors, Intelligent Data Engineering and Automated Learning - 22nd International Conference, IDEAL 2021, Proceedings: 22nd International Conference, IDEAL 2021, Manchester, UK, November 25–27, 2021, Proceedings. Cham: Springer. 2021. p. 213-222. (Lecture Notes in Computer Science (LNCS)). (Information Systems and Applications, incl. Internet/Web, and HCI (LNISA)). doi: 10.1007/978-3-030-91608-4_22

A Deep Learning-Based Approach for Train Arrival Time Prediction

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this