Physics-Anchored Residual Learning Beyond Mean-Bias Correction for Coastal Person-in-Water Drift Prediction
Journal of Marine Science and Engineering · 07.11.2026 · 14 (21) · 2113
DOI: 10.3390/jmse14212113 ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of Marine Science and Engineering, online erschienen am 07.11.2026. Verfasst von Jiaming Li und 4 weiteren Autor:innen. Die Arbeit nennt 21 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Natural Science Foundation of Liaoning Province, Department of Science and Technology of Liaoning Province, Dalian City Science and Technology Plan (Key) Project.
Abstract (Originalsprache)
Coastal person-in-water searches require accurate target trajectories, but sparse field releases constrain corrections that transfer across environmental conditions. We develop a physics-anchored method that uses Tabular Prior-Data Fitted Network (TabPFN) to predict along-track and cross-track velocity residuals from simulated motion and wind–current inputs and then integrates them with the physical trajectory. Evaluation is retrospective, using complete-event holdout across three Dalian releases previously examined during method development, each with five recurring manikin configurations. The physical baseline and constant residual transfer provide matched controls for the conditional correction. At 180 min, average displacement error (ADE) falls from 884.4 to 566.8 m relative to constant residual correction, a 35.9% reduction, while endpoint error is the lowest among the tested methods at 914.3 m. At 350 min, ADE remains 18.8% lower than constant correction, although the latter gives the closer endpoint. Matched geographic-coordinate controls yield substantially higher path errors than the moving representation. The macro path-accuracy gain persists under joint observation-gap exclusion, alternative physical selection and ten inference seeds. Conditional correction therefore extracts predictive information beyond mean-bias transfer, and the moving representation enables that information to improve trajectory placement, with the strongest joint path-and-endpoint evidence at 180 min.
Förderung
- Natural Science Foundation of Liaoning Province
- Department of Science and Technology of Liaoning Province
- Dalian City Science and Technology Plan (Key) Project
- Fundamental Research Funds for the Central Universities
- Maritime Visual Augmented Reality Navigation System Development
Zitieren
APA
Li, J., Jing, Q., Xiao, F., Shen, H., & Yin, Y. (2026). Physics-Anchored Residual Learning Beyond Mean-Bias Correction for Coastal Person-in-Water Drift Prediction. Journal of Marine Science and Engineering, 14(21), 2113. https://doi.org/10.3390/jmse14212113
Vancouver
Li J, Jing Q, Xiao F, Shen H, Yin Y. Physics-Anchored Residual Learning Beyond Mean-Bias Correction for Coastal Person-in-Water Drift Prediction. Journal of Marine Science and Engineering. 2026;14(21):2113. doi:10.3390/jmse14212113
BibTeX
@article{Li2026,
author = {Li, Jiaming and Jing, Qianfeng and Xiao, Fangbing and Shen, Helong and Yin, Yong},
title = {Physics-Anchored Residual Learning Beyond Mean-Bias Correction for Coastal Person-in-Water Drift Prediction},
journal = {Journal of Marine Science and Engineering},
year = {2026},
volume = {14},
number = {21},
pages = {2113},
doi = {10.3390/jmse14212113}
}
EUVIMED