A Multitask Learning Model for Traffic Flow and Speed Forecasting
Intelligent Transportation Systems (ITS) research and applications benefit from accurate short-term traffic state forecasting. To improve the forecasting accuracy, this paper proposes a deep learning based multitask learning Gated Recurrent Units (MTL-GRU) with residual mappings. To enhance the perf...
Main Authors: | , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
IEEE
2020-01-01
|
Series: | IEEE Access |
Subjects: | |
Online Access: | https://ieeexplore.ieee.org/document/9080108/ |
id |
doaj-4535c3ad8a544b19aad157a527b06a55 |
---|---|
record_format |
Article |
spelling |
doaj-4535c3ad8a544b19aad157a527b06a552021-03-30T02:43:20ZengIEEEIEEE Access2169-35362020-01-018807078071510.1109/ACCESS.2020.29909589080108A Multitask Learning Model for Traffic Flow and Speed ForecastingKunpeng Zhang0https://orcid.org/0000-0003-1979-4261Lan Wu1https://orcid.org/0000-0002-2497-6556Zhaoju Zhu2Jiang Deng3Key Laboratory of Grain Information Processing and Control, Ministry of Education, Henan University of Technology, Zhengzhou, ChinaKey Laboratory of Grain Information Processing and Control, Ministry of Education, Henan University of Technology, Zhengzhou, ChinaSchool of Mechanical Engineering and Automation, Fuzhou University, Fuzhou, ChinaFInSight AI Lab, Qingdao Fantaike Bearing Company, Ltd., Qingdao, ChinaIntelligent Transportation Systems (ITS) research and applications benefit from accurate short-term traffic state forecasting. To improve the forecasting accuracy, this paper proposes a deep learning based multitask learning Gated Recurrent Units (MTL-GRU) with residual mappings. To enhance the performance of the MTL-GRU, feature engineering is introduced to select the most informative features for the forecasting. Then, based on real-world datasets, numerical results show that the MTL-GRU can well estimate traffic flow and speed simultaneously, and performs better than other counterparts. Experiments also show that the deep learning based MTL-GRU model can overpower the bottleneck caused by enlarging training datasets and continue to gain benefits. The results suggest the proposed MTL-GRU model with residual mappings is promising to forecast short-term traffic state.https://ieeexplore.ieee.org/document/9080108/Short-term traffic forecastingdeep learningmultitask learningfeature engineering |
collection |
DOAJ |
language |
English |
format |
Article |
sources |
DOAJ |
author |
Kunpeng Zhang Lan Wu Zhaoju Zhu Jiang Deng |
spellingShingle |
Kunpeng Zhang Lan Wu Zhaoju Zhu Jiang Deng A Multitask Learning Model for Traffic Flow and Speed Forecasting IEEE Access Short-term traffic forecasting deep learning multitask learning feature engineering |
author_facet |
Kunpeng Zhang Lan Wu Zhaoju Zhu Jiang Deng |
author_sort |
Kunpeng Zhang |
title |
A Multitask Learning Model for Traffic Flow and Speed Forecasting |
title_short |
A Multitask Learning Model for Traffic Flow and Speed Forecasting |
title_full |
A Multitask Learning Model for Traffic Flow and Speed Forecasting |
title_fullStr |
A Multitask Learning Model for Traffic Flow and Speed Forecasting |
title_full_unstemmed |
A Multitask Learning Model for Traffic Flow and Speed Forecasting |
title_sort |
multitask learning model for traffic flow and speed forecasting |
publisher |
IEEE |
series |
IEEE Access |
issn |
2169-3536 |
publishDate |
2020-01-01 |
description |
Intelligent Transportation Systems (ITS) research and applications benefit from accurate short-term traffic state forecasting. To improve the forecasting accuracy, this paper proposes a deep learning based multitask learning Gated Recurrent Units (MTL-GRU) with residual mappings. To enhance the performance of the MTL-GRU, feature engineering is introduced to select the most informative features for the forecasting. Then, based on real-world datasets, numerical results show that the MTL-GRU can well estimate traffic flow and speed simultaneously, and performs better than other counterparts. Experiments also show that the deep learning based MTL-GRU model can overpower the bottleneck caused by enlarging training datasets and continue to gain benefits. The results suggest the proposed MTL-GRU model with residual mappings is promising to forecast short-term traffic state. |
topic |
Short-term traffic forecasting deep learning multitask learning feature engineering |
url |
https://ieeexplore.ieee.org/document/9080108/ |
work_keys_str_mv |
AT kunpengzhang amultitasklearningmodelfortrafficflowandspeedforecasting AT lanwu amultitasklearningmodelfortrafficflowandspeedforecasting AT zhaojuzhu amultitasklearningmodelfortrafficflowandspeedforecasting AT jiangdeng amultitasklearningmodelfortrafficflowandspeedforecasting AT kunpengzhang multitasklearningmodelfortrafficflowandspeedforecasting AT lanwu multitasklearningmodelfortrafficflowandspeedforecasting AT zhaojuzhu multitasklearningmodelfortrafficflowandspeedforecasting AT jiangdeng multitasklearningmodelfortrafficflowandspeedforecasting |
_version_ |
1724184619089133568 |