Summary of Short-term Inland Vessel Trajectory Prediction with Encoder-decoder Models, by Kathrin Donandt et al.
Short-term Inland Vessel Trajectory Prediction with Encoder-Decoder Models
by Kathrin Donandt, Karim Böttger, Dirk Söffker
First submitted to arxiv on: 4 Jun 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: None
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary Deep learning-based vessel trajectory prediction is crucial for safe and efficient inland navigation. Traditional approaches from the maritime domain cannot be directly applied due to unique driving behavior factors. This study compares different encoder-decoder architectures, including transformer encoder-decoders, for predicting next positions of inland vessels given AIS data and river-specific features. The results show that reformulating the regression task as a classification problem and incorporating river-specific features yield the lowest displacement errors. A standard LSTM encoder-decoder outperforms the transformer-based model for the considered data but is computationally more expensive. This study introduces the application of a transformer-based encoder-decoder to ship trajectory prediction, establishing a feature vector using river-specific navigation parameters. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Predicting where boats will go is important for safe and efficient boat travel on rivers. Current methods from ocean navigation don’t work well because rivers have different rules. This study compares different ways to do this predicting using computer programs. The results show that one way of doing it, called “classification,” works better than others when you add special information about the river. A more common method, called LSTM, does slightly better but takes longer to compute. This study is the first to use a new type of program, called a transformer, for predicting boat routes. |
Keywords
» Artificial intelligence » Classification » Deep learning » Encoder decoder » Lstm » Regression » Transformer
