Scholarlink

The prediction of a moving object’s future position or trajectory is crucial in navigation and maritime logistics, as it directly impacts route optimization, traffic management, and collision avoidance. Precise predictions enhance operational efficiency, safety, and the ability to navigate obstacles and environmental constraints. While technologies like the Automatic Identification System (AIS) offer real-time tracking, they often overlook the influence of weather conditions on navigation. This study proposes a bidirectional deep learning model that integrates AIS data with meteorological information, accounting for both historical and future vessel trajectories. The approach improves prediction accuracy, especially under weather conditions. Evaluation of the model using metrics such as Root Mean Square Error (RMSE), Mean Squared Error (MSE), and Mean Absolute Error (MAE) highlights its potential to enhance safer navigation and more efficient maritime logistics. The results show that combining AIS and meteorological data significantly improves prediction accuracy, leading to more adaptive and reliable vessel trajectory forecasting models.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Machine Learning-Based Predictive Approach for the Vessel’s Trajectory Using Automatic Identification System and Meteorological Data

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Houyem Mjadri,
  • Wided Oueslati,
  • Afef Bahri

摘要

The prediction of a moving object’s future position or trajectory is crucial in navigation and maritime logistics, as it directly impacts route optimization, traffic management, and collision avoidance. Precise predictions enhance operational efficiency, safety, and the ability to navigate obstacles and environmental constraints. While technologies like the Automatic Identification System (AIS) offer real-time tracking, they often overlook the influence of weather conditions on navigation. This study proposes a bidirectional deep learning model that integrates AIS data with meteorological information, accounting for both historical and future vessel trajectories. The approach improves prediction accuracy, especially under weather conditions. Evaluation of the model using metrics such as Root Mean Square Error (RMSE), Mean Squared Error (MSE), and Mean Absolute Error (MAE) highlights its potential to enhance safer navigation and more efficient maritime logistics. The results show that combining AIS and meteorological data significantly improves prediction accuracy, leading to more adaptive and reliable vessel trajectory forecasting models.