Ship Trajectory Prediction Using ACoAtt–LSTM and AIS Data

Mingze Li,Bing Li,Zhigang Qi,Jiashuai Li,Jiawei Wu

DOI: https://doi.org/10.3390/ijgi13030085

IF: 3.4

2024-03-09

ISPRS International Journal of Geo-Information

Abstract:Predicting ship trajectories plays a vital role in ensuring navigational safety, preventing collision incidents, and enhancing vessel management efficiency. The integration of advanced machine learning technology for precise trajectory prediction is emerging as a new trend in sophisticated geospatial applications. However, the complexity of the marine environment and data quality issues pose significant challenges to accurate ship trajectory forecasting. This study introduces an innovative trajectory prediction method, combining data encoding representation, attribute correlation attention module, and long short-term memory network. Initially, we process AIS data using data encoding conversion technology to improve representation efficiency and reduce complexity. This encoding not only preserves key information from the original data but also provides a more efficient input format for deep learning models. Subsequently, we incorporate the attribute correlation attention module, utilizing a multi-head attention mechanism to capture complex relationships between dynamic ship attributes, such as speed and direction, thereby enhancing the model's understanding of implicit time series patterns in the data. Finally, leveraging the long short-term memory network's capability for processing time series data, our approach effectively predicts future ship trajectories. In our experiments, we trained and tested our model using a historical AIS dataset. The results demonstrate that our model surpasses other classic intelligent models and advanced models with attention mechanisms in terms of trajectory prediction accuracy and stability.

geography, physical,remote sensing,computer science, information systems

What problem does this paper attempt to address?

The paper primarily addresses the issue of ship trajectory prediction in maritime navigation safety by proposing a new solution. Specifically, the research aims to enhance maritime navigation safety, prevent collision accidents, and improve ship management efficiency by increasing the accuracy of ship trajectory predictions.

To tackle this problem, the authors propose an innovative trajectory prediction method that combines data encoding representation, an Attribute Correlation Attention (ACoAtt) module, and Long Short-Term Memory (LSTM) networks. The main contributions of this study are as follows:

1. **Novel Data Representation Method**: The paper first encodes Automatic Identification System (AIS) data to improve data representation efficiency and reduce complexity. This encoding not only retains the key information of the original data but also provides a more efficient data format for deep learning models.

2. **Enhanced Correlation of Dynamic Features**: The introduction of the ACoAtt module utilizes a multi-head attention mechanism to capture the complex relationships between dynamic features of ships (such as speed and direction), thereby enhancing the model's understanding of implicit time series patterns.

3. **Optimized Time Series Data Processing**: A multi-layer LSTM network is used to model the temporal dependencies in AIS data, leveraging LSTM's capability to handle spatiotemporal semantics to improve the accuracy of short-term ship trajectory predictions.

Through these technical means, the researchers have constructed a model capable of effectively predicting future ship trajectories. Experimental results on historical AIS datasets indicate that this model surpasses other classic intelligent models and advanced models with attention mechanisms in terms of prediction accuracy and stability. This method holds significant importance for improving maritime traffic safety, optimizing navigation decisions, and reducing collision risks.