@inproceedings{PB-VRVis-2019-051,
author = {Graser, Anita and Schmidt, Johanna and Dragaschnig, Melitta and Widhalm, G.},
title = {Data-driven Trajectory Prediction and Spatial Variability of Prediction Performance in Mari- time Location Based Services},
year = {2019},
booktitle = {Adjunct Proceedings of the 15th International Conference on Location Based Services (LBS 2019)},
editor = {Georg Gartner and Haosheng Huang},
url = {https://www.vrvis.at/publications/PB-VRVis-2019-051},
abstract = {Location-based services in the maritime domain aim to improve efficiency and safety of vessel operations. Predictive functionality can in- crease the value of these services beyond ordinary visualizations of the cur- rent operational picture. Trajectory prediction aims to forecast the future path of vessels and can thus help improve logistics as well as help predict potentially dangerous situations. This paper presents ongoing work on da- ta-driven trajectory prediction that leverages information of past vessel movements to improve prediction results. Preliminary results show that data-driven prediction outperforms baseline approaches, particularly in complex situations. However, results also show a large spatial variability in prediction performance. This indicates that it is impossible to compare the performance of different prediction methods by relying solely on the error statistics reported in publications since every research group uses different data samples from different geographic regions},
keywords = {Computational movement analysis, trajectory prediction, Au- tomatic Identification System},
}