Forecasting train arrival time using modular artificial neural networks

Hanseon Cho, Laurence R Rilett

Research output: Contribution to conferencePaper

Abstract

A modular Artificial Neural Network was used to forecast the train arrival time at a highway-railroad grade crossing (HRGC). A modular approach was used because the trains often have different characteristics depending on their cargo and the operational rules in effect at the time they are detected. The arrival time forecast was performed 100 seconds after the train was first detected. About 190 trains were used for training the ANN and 76 trains were used for testing. A modular architecture gave superior results to that of a simple ANN model, standard regression techniques, and simple forecasting methods.

Original languageEnglish (US)
Pages852-860
Number of pages9
StatePublished - Jan 1 2002
Externally publishedYes
EventProceedings of the seventh International Conference on: Applications of Advanced Technology in Transportation - Cambridge, MA, United States
Duration: Aug 5 2002Aug 7 2002

Other

OtherProceedings of the seventh International Conference on: Applications of Advanced Technology in Transportation
CountryUnited States
CityCambridge, MA
Period8/5/028/7/02

Fingerprint

Crossings (pipe and cable)
Railroads
Neural networks
Testing

ASJC Scopus subject areas

  • Engineering(all)

Cite this

Cho, H., & Rilett, L. R. (2002). Forecasting train arrival time using modular artificial neural networks. 852-860. Paper presented at Proceedings of the seventh International Conference on: Applications of Advanced Technology in Transportation, Cambridge, MA, United States.

Forecasting train arrival time using modular artificial neural networks. / Cho, Hanseon; Rilett, Laurence R.

2002. 852-860 Paper presented at Proceedings of the seventh International Conference on: Applications of Advanced Technology in Transportation, Cambridge, MA, United States.

Research output: Contribution to conferencePaper

Cho, H & Rilett, LR 2002, 'Forecasting train arrival time using modular artificial neural networks' Paper presented at Proceedings of the seventh International Conference on: Applications of Advanced Technology in Transportation, Cambridge, MA, United States, 8/5/02 - 8/7/02, pp. 852-860.
Cho H, Rilett LR. Forecasting train arrival time using modular artificial neural networks. 2002. Paper presented at Proceedings of the seventh International Conference on: Applications of Advanced Technology in Transportation, Cambridge, MA, United States.
Cho, Hanseon ; Rilett, Laurence R. / Forecasting train arrival time using modular artificial neural networks. Paper presented at Proceedings of the seventh International Conference on: Applications of Advanced Technology in Transportation, Cambridge, MA, United States.9 p.
@conference{5035303d3dd7467f940e1ec2419861a2,
title = "Forecasting train arrival time using modular artificial neural networks",
abstract = "A modular Artificial Neural Network was used to forecast the train arrival time at a highway-railroad grade crossing (HRGC). A modular approach was used because the trains often have different characteristics depending on their cargo and the operational rules in effect at the time they are detected. The arrival time forecast was performed 100 seconds after the train was first detected. About 190 trains were used for training the ANN and 76 trains were used for testing. A modular architecture gave superior results to that of a simple ANN model, standard regression techniques, and simple forecasting methods.",
author = "Hanseon Cho and Rilett, {Laurence R}",
year = "2002",
month = "1",
day = "1",
language = "English (US)",
pages = "852--860",
note = "Proceedings of the seventh International Conference on: Applications of Advanced Technology in Transportation ; Conference date: 05-08-2002 Through 07-08-2002",

}

TY - CONF

T1 - Forecasting train arrival time using modular artificial neural networks

AU - Cho, Hanseon

AU - Rilett, Laurence R

PY - 2002/1/1

Y1 - 2002/1/1

N2 - A modular Artificial Neural Network was used to forecast the train arrival time at a highway-railroad grade crossing (HRGC). A modular approach was used because the trains often have different characteristics depending on their cargo and the operational rules in effect at the time they are detected. The arrival time forecast was performed 100 seconds after the train was first detected. About 190 trains were used for training the ANN and 76 trains were used for testing. A modular architecture gave superior results to that of a simple ANN model, standard regression techniques, and simple forecasting methods.

AB - A modular Artificial Neural Network was used to forecast the train arrival time at a highway-railroad grade crossing (HRGC). A modular approach was used because the trains often have different characteristics depending on their cargo and the operational rules in effect at the time they are detected. The arrival time forecast was performed 100 seconds after the train was first detected. About 190 trains were used for training the ANN and 76 trains were used for testing. A modular architecture gave superior results to that of a simple ANN model, standard regression techniques, and simple forecasting methods.

UR - http://www.scopus.com/inward/record.url?scp=0036045922&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0036045922&partnerID=8YFLogxK

M3 - Paper

SP - 852

EP - 860

ER -