DSformer for Ship Motion Prediction: A Statistics-Driven Framework with Environment-Adaptive Hyperparameter Tuning

Ge, Haowen; Li, Ying; Mao, Yuntao; Li, Jian; Chen, Ziwei; Bai, Pengying; Peng, Xueming

doi:10.3390/jmse14030244

This is an early access version, the complete PDF, HTML, and XML versions will be available soon.

Open AccessArticle

DSformer for Ship Motion Prediction: A Statistics-Driven Framework with Environment-Adaptive Hyperparameter Tuning

by

Haowen Ge

^1,†,

Ying Li

^2,†,

Yuntao Mao

^1,†

,

Jian Li

²,

Ziwei Chen

^1,*,

Pengying Bai

^2,* and

Xueming Peng

^2,*

¹

School of Electronic and Information Engineering, Beijing Jiaotong University, Beijing 100044, China

²

Beijing Institute of Mechanical Engineering, Beijing 100854, China

^*

Authors to whom correspondence should be addressed.

^†

These authors contributed equally to this work.

J. Mar. Sci. Eng. 2026, 14(3), 244; https://doi.org/10.3390/jmse14030244

Submission received: 18 December 2025 / Revised: 15 January 2026 / Accepted: 22 January 2026 / Published: 23 January 2026

(This article belongs to the Section Ocean Engineering)

Download Versions Notes

Abstract

Given the central importance of maritime logistics to global trade, accurate and efficient vessel motion forecasting is essential for strengthening supply chain resilience and improving operational efficiency. However, traditional physical and statistical models often fail to effectively capture the multivariate, noisy, and strongly coupled nature of maritime dynamics. In this manuscript, we adapt the DSformer architecture for ship motion forecasting, leveraging its dual sampling and dual-attention design to address the multi-scale and cross-variable dependencies inherent in maritime data. Across three real-world datasets, the adapted DSformer reduces prediction error by 23% and training time by 70% compared with 13 state-of-the-art (SOTA) baselines. Moreover, we identify a consistent relationship between sampling strategies and sea states, where dense sampling performs best under stable conditions, whereas moderately sparse sampling with multi-head attention improves robustness under turbulent environments. These results apply the algorithm’s new capabilities to the daily management of maritime logistics. By adapting the architecture to real-world operational settings and optimizing its key parameters, the approach enables efficient, real-time vessel forecasting and decision support across global supply chains.

Keywords: ship motion prediction; DSformer; multi-scale temporal analysis; deep learning; maritime safety

Share and Cite

MDPI and ACS Style

Ge, H.; Li, Y.; Mao, Y.; Li, J.; Chen, Z.; Bai, P.; Peng, X. DSformer for Ship Motion Prediction: A Statistics-Driven Framework with Environment-Adaptive Hyperparameter Tuning. J. Mar. Sci. Eng. 2026, 14, 244. https://doi.org/10.3390/jmse14030244

AMA Style

Ge H, Li Y, Mao Y, Li J, Chen Z, Bai P, Peng X. DSformer for Ship Motion Prediction: A Statistics-Driven Framework with Environment-Adaptive Hyperparameter Tuning. Journal of Marine Science and Engineering. 2026; 14(3):244. https://doi.org/10.3390/jmse14030244

Chicago/Turabian Style

Ge, Haowen, Ying Li, Yuntao Mao, Jian Li, Ziwei Chen, Pengying Bai, and Xueming Peng. 2026. "DSformer for Ship Motion Prediction: A Statistics-Driven Framework with Environment-Adaptive Hyperparameter Tuning" Journal of Marine Science and Engineering 14, no. 3: 244. https://doi.org/10.3390/jmse14030244

APA Style

Ge, H., Li, Y., Mao, Y., Li, J., Chen, Z., Bai, P., & Peng, X. (2026). DSformer for Ship Motion Prediction: A Statistics-Driven Framework with Environment-Adaptive Hyperparameter Tuning. Journal of Marine Science and Engineering, 14(3), 244. https://doi.org/10.3390/jmse14030244

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Article metric data becomes available approximately 24 hours after publication online.

Article Menu

DSformer for Ship Motion Prediction: A Statistics-Driven Framework with Environment-Adaptive Hyperparameter Tuning

Abstract

Share and Cite

Article Metrics

Further Information

Guidelines

MDPI Initiatives

Follow MDPI