A Multi-decoder Recurrent Network for Vessel Trajectory Prediction Using Multi-memory LSTMs

Meng Chen; Chi Zhang; Tengteng Qu; Bo Chen; Chengqi Cheng; Haojiang Deng

doi:10.1109/ICUS55513.2022.9986985

A Multi-decoder Recurrent Network for Vessel Trajectory Prediction Using Multi-memory LSTMs

Meng Chen
, Chi Zhang
, Tengteng Qu^*
, Bo Chen
, Chengqi Cheng
, Haojiang Deng

^*Corresponding author for this work

Peking University
Harbin Institute of Technology Shenzhen
Peng Cheng Laboratory

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

2 Scopus citations

Abstract

Vessel trajectory prediction is the foundation for constructing the marine intelligent traffic management system. Deep learning algorithms driven by massive AIS data provide more possibilities for accurately predicting vessel trajectories. Currently, LSTM models widely used for this task have limited ability to extract temporal features, and prediction errors accumulate rapidly over time. To address the above issues, we design a multi-memory state LSTM (MM-LSTM) unit by introducing global memory states and use a multi-decoder structure to capture features for different prediction periods. Experimental results on the real AIS dataset provided by NOAA show that: first, the MM-LSTM unit can effectively capture the temporal information and reduce the prediction error of vessel trajectory by 1.93km (11.5%) than the standard LSTM; second, the multi-decoder structure can universally improve the trajectory prediction accuracy for each prediction period, which has promising applications in trajectory prediction tasks.

Original language	English
Title of host publication	Proceedings of 2022 IEEE International Conference on Unmanned Systems, ICUS 2022
Editors	Rong Song
Publisher	Institute of Electrical and Electronics Engineers Inc.
Pages	944-949
Number of pages	6
ISBN (Electronic)	9781665484565
DOIs	https://doi.org/10.1109/ICUS55513.2022.9986985
State	Published - 2022
Externally published	Yes
Event	2022 IEEE International Conference on Unmanned Systems, ICUS 2022 - Guangzhou, China Duration: 28 Oct 2022 → 30 Oct 2022

Publication series

Name	Proceedings of 2022 IEEE International Conference on Unmanned Systems, ICUS 2022

Conference

Conference	2022 IEEE International Conference on Unmanned Systems, ICUS 2022
Country/Territory	China
City	Guangzhou
Period	28/10/22 → 30/10/22

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

SDG 14 Life Below Water

Keywords

AIS
LSTM
encoder-decoder model
vessel trajectory prediction

Access to Document

10.1109/ICUS55513.2022.9986985

Cite this

Chen, M., Zhang, C., Qu, T., Chen, B., Cheng, C., & Deng, H. (2022). A Multi-decoder Recurrent Network for Vessel Trajectory Prediction Using Multi-memory LSTMs. In R. Song (Ed.), Proceedings of 2022 IEEE International Conference on Unmanned Systems, ICUS 2022 (pp. 944-949). (Proceedings of 2022 IEEE International Conference on Unmanned Systems, ICUS 2022). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/ICUS55513.2022.9986985

Chen, Meng ; Zhang, Chi ; Qu, Tengteng et al. / A Multi-decoder Recurrent Network for Vessel Trajectory Prediction Using Multi-memory LSTMs. Proceedings of 2022 IEEE International Conference on Unmanned Systems, ICUS 2022. editor / Rong Song. Institute of Electrical and Electronics Engineers Inc., 2022. pp. 944-949 (Proceedings of 2022 IEEE International Conference on Unmanned Systems, ICUS 2022).

@inproceedings{089f6dd9b52844798165cad47029b4d3,

title = "A Multi-decoder Recurrent Network for Vessel Trajectory Prediction Using Multi-memory LSTMs",

abstract = "Vessel trajectory prediction is the foundation for constructing the marine intelligent traffic management system. Deep learning algorithms driven by massive AIS data provide more possibilities for accurately predicting vessel trajectories. Currently, LSTM models widely used for this task have limited ability to extract temporal features, and prediction errors accumulate rapidly over time. To address the above issues, we design a multi-memory state LSTM (MM-LSTM) unit by introducing global memory states and use a multi-decoder structure to capture features for different prediction periods. Experimental results on the real AIS dataset provided by NOAA show that: first, the MM-LSTM unit can effectively capture the temporal information and reduce the prediction error of vessel trajectory by 1.93km (11.5\%) than the standard LSTM; second, the multi-decoder structure can universally improve the trajectory prediction accuracy for each prediction period, which has promising applications in trajectory prediction tasks.",

keywords = "AIS, LSTM, encoder-decoder model, vessel trajectory prediction",

author = "Meng Chen and Chi Zhang and Tengteng Qu and Bo Chen and Chengqi Cheng and Haojiang Deng",

note = "Publisher Copyright: {\textcopyright} 2022 IEEE.; 2022 IEEE International Conference on Unmanned Systems, ICUS 2022 ; Conference date: 28-10-2022 Through 30-10-2022",

year = "2022",

doi = "10.1109/ICUS55513.2022.9986985",

language = "英语",

series = "Proceedings of 2022 IEEE International Conference on Unmanned Systems, ICUS 2022",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

pages = "944--949",

editor = "Rong Song",

booktitle = "Proceedings of 2022 IEEE International Conference on Unmanned Systems, ICUS 2022",

address = "美国",

}

Chen, M, Zhang, C, Qu, T, Chen, B, Cheng, C & Deng, H 2022, A Multi-decoder Recurrent Network for Vessel Trajectory Prediction Using Multi-memory LSTMs. in R Song (ed.), Proceedings of 2022 IEEE International Conference on Unmanned Systems, ICUS 2022. Proceedings of 2022 IEEE International Conference on Unmanned Systems, ICUS 2022, Institute of Electrical and Electronics Engineers Inc., pp. 944-949, 2022 IEEE International Conference on Unmanned Systems, ICUS 2022, Guangzhou, China, 28/10/22. https://doi.org/10.1109/ICUS55513.2022.9986985

A Multi-decoder Recurrent Network for Vessel Trajectory Prediction Using Multi-memory LSTMs. / Chen, Meng; Zhang, Chi; Qu, Tengteng et al.
Proceedings of 2022 IEEE International Conference on Unmanned Systems, ICUS 2022. ed. / Rong Song. Institute of Electrical and Electronics Engineers Inc., 2022. p. 944-949 (Proceedings of 2022 IEEE International Conference on Unmanned Systems, ICUS 2022).

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

TY - GEN

T1 - A Multi-decoder Recurrent Network for Vessel Trajectory Prediction Using Multi-memory LSTMs

AU - Chen, Meng

AU - Zhang, Chi

AU - Qu, Tengteng

AU - Chen, Bo

AU - Cheng, Chengqi

AU - Deng, Haojiang

PY - 2022

Y1 - 2022

N2 - Vessel trajectory prediction is the foundation for constructing the marine intelligent traffic management system. Deep learning algorithms driven by massive AIS data provide more possibilities for accurately predicting vessel trajectories. Currently, LSTM models widely used for this task have limited ability to extract temporal features, and prediction errors accumulate rapidly over time. To address the above issues, we design a multi-memory state LSTM (MM-LSTM) unit by introducing global memory states and use a multi-decoder structure to capture features for different prediction periods. Experimental results on the real AIS dataset provided by NOAA show that: first, the MM-LSTM unit can effectively capture the temporal information and reduce the prediction error of vessel trajectory by 1.93km (11.5%) than the standard LSTM; second, the multi-decoder structure can universally improve the trajectory prediction accuracy for each prediction period, which has promising applications in trajectory prediction tasks.

AB - Vessel trajectory prediction is the foundation for constructing the marine intelligent traffic management system. Deep learning algorithms driven by massive AIS data provide more possibilities for accurately predicting vessel trajectories. Currently, LSTM models widely used for this task have limited ability to extract temporal features, and prediction errors accumulate rapidly over time. To address the above issues, we design a multi-memory state LSTM (MM-LSTM) unit by introducing global memory states and use a multi-decoder structure to capture features for different prediction periods. Experimental results on the real AIS dataset provided by NOAA show that: first, the MM-LSTM unit can effectively capture the temporal information and reduce the prediction error of vessel trajectory by 1.93km (11.5%) than the standard LSTM; second, the multi-decoder structure can universally improve the trajectory prediction accuracy for each prediction period, which has promising applications in trajectory prediction tasks.

KW - AIS

KW - LSTM

KW - encoder-decoder model

KW - vessel trajectory prediction

UR - https://www.scopus.com/pages/publications/85146487168

U2 - 10.1109/ICUS55513.2022.9986985

DO - 10.1109/ICUS55513.2022.9986985

M3 - 会议稿件

AN - SCOPUS:85146487168

T3 - Proceedings of 2022 IEEE International Conference on Unmanned Systems, ICUS 2022

SP - 944

EP - 949

BT - Proceedings of 2022 IEEE International Conference on Unmanned Systems, ICUS 2022

A2 - Song, Rong

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 2022 IEEE International Conference on Unmanned Systems, ICUS 2022

Y2 - 28 October 2022 through 30 October 2022

ER -

Chen M, Zhang C, Qu T, Chen B, Cheng C, Deng H. A Multi-decoder Recurrent Network for Vessel Trajectory Prediction Using Multi-memory LSTMs. In Song R, editor, Proceedings of 2022 IEEE International Conference on Unmanned Systems, ICUS 2022. Institute of Electrical and Electronics Engineers Inc. 2022. p. 944-949. (Proceedings of 2022 IEEE International Conference on Unmanned Systems, ICUS 2022). doi: 10.1109/ICUS55513.2022.9986985

A Multi-decoder Recurrent Network for Vessel Trajectory Prediction Using Multi-memory LSTMs

Abstract

Publication series

Conference

UN SDGs

Keywords

Access to Document

Other files and links

Fingerprint

Cite this