A platform for research: civil engineering, architecture and urbanism
A platform for research: civil engineering, architecture and urbanism
Ship Trajectory Prediction Model Based on Improved Bi-LSTM
Ship trajectory prediction plays an important role in ensuring ship safety; through accurate ship positioning, the future trajectory of ships and their encounter time and location can be obtained, which facilitates the maritime regulatory authorities to assess the risks of ship encounters and implement effective traffic control. Meanwhile, with the rapid development of the shipping industry, the increasingly complex maritime traffic poses potential risks, which may cause serious traffic accidents and huge economic losses. To improve the accuracy of ship navigation risk prediction and ensure the safety of ship navigation, automatic identification system (AIS) data and deep learning models are used to extract the ship trajectory change feature pattern and apply it to ship trajectory prediction. This study builds the improved bidirectional long short-term memory network (Bi-LSTM) model based on rectified adaptive moment estimation (Radam) and lookahead, respectively. The AIS data of the Port of Tianjin area were selected for model training, and the results of comparison experiments show that the improved Bi-LSTM model has a stronger generalization ability, which further improves the trajectory prediction accuracy, and shows excellent predictive performance. The prediction model is feasible for the prediction of ship navigation trajectory.
Ship Trajectory Prediction Model Based on Improved Bi-LSTM
Ship trajectory prediction plays an important role in ensuring ship safety; through accurate ship positioning, the future trajectory of ships and their encounter time and location can be obtained, which facilitates the maritime regulatory authorities to assess the risks of ship encounters and implement effective traffic control. Meanwhile, with the rapid development of the shipping industry, the increasingly complex maritime traffic poses potential risks, which may cause serious traffic accidents and huge economic losses. To improve the accuracy of ship navigation risk prediction and ensure the safety of ship navigation, automatic identification system (AIS) data and deep learning models are used to extract the ship trajectory change feature pattern and apply it to ship trajectory prediction. This study builds the improved bidirectional long short-term memory network (Bi-LSTM) model based on rectified adaptive moment estimation (Radam) and lookahead, respectively. The AIS data of the Port of Tianjin area were selected for model training, and the results of comparison experiments show that the improved Bi-LSTM model has a stronger generalization ability, which further improves the trajectory prediction accuracy, and shows excellent predictive performance. The prediction model is feasible for the prediction of ship navigation trajectory.
Ship Trajectory Prediction Model Based on Improved Bi-LSTM
ASCE-ASME J. Risk Uncertainty Eng. Syst., Part A: Civ. Eng.
Li, Weifeng (author) / Lian, Yifan (author) / Liu, Yaochen (author) / Shi, Guoyou (author)
2024-09-01
Article (Journal)
Electronic Resource
English
Short-term ship motion attitude prediction based on LSTM and GPR
| Elsevier | 2021
Prediction of Ship Trajectory in Nearby Port Waters Based on Attention Mechanism Model
| DOAJ | 2023