ShipSeer: Pushing accuracy-performance boundaries in ship motion prediction with spectral and multi-component analysis

Accurate ship motion prediction is vital for ensuring maritime safety, cargo stability, and operational efficiency. However, nonlinear ship dynamics, long-term dependency challenges, multi-step error accumulation, and limited onboard computational resources make long-term prediction difficult. This paper introduces ShipSeer, a resource-efficient MLP-based multi-input multi-output forecasting framework that jointly predicts eight ship motion states over long horizons. By combining spectral and multi-component analysis, ShipSeer can address the above problems while efficiently extracting key frequencies and parallelly modeling different temporal patterns. Evaluated on four real-world datasets, ShipSeer outperforms nine state-of-the-art models, achieving 19.67% higher accuracy and 49.25% faster inference. Robustness and ablation studies confirm the effectiveness of each module, demonstrating ShipSeer's practical value in advancing the intelligence of modern marine systems under complex and noisy sea conditions. Furthermore, experiments demonstrate how environmental conditions are reflected in the spectral characteristics of ship motion data, providing physically consistent explanations for the observed prediction behavior.

• Record URL:
  • https://doi.org/10.1016/j.oceaneng.2026.124611
• Record URL:
  • http://www.sciencedirect.com/science/article/pii/S0029801826004452
• Availability:
  • Find a library where document is available. Order URL: http://worldcat.org/issn/00298018
• Supplemental Notes:
  • © 2026 Elsevier Ltd. All rights are reserved, including those for text and data mining, AI training, and similar technologies. Abstract reprinted with permission of Elsevier.
• Authors:
  • Yu, Chunyang
  • 0009-0006-2510-0018
  • Mao, Yuntao
  • 0009-0001-4706-8809
  • Liu, Dengke
  • 0009-0002-9570-0925
  • Chen, Ziwei
• Publication Date: 2026-4-15

Language

• English

Media Info

• Media Type: Web
• Features: References;
• Pagination: 124611
• Serial:
  • Ocean Engineering
  • Volume: 352
  • Issue Number: 0
  • Publisher: Pergamon
  • ISSN: 0029-8018
  • EISSN: 1873-5258
  • Serial URL: http://www.sciencedirect.com/science/journal/00298018

Subject/Index Terms

• TRT Terms: Marine engineering; Predictive models; Ship motion; Spectrum analysis
• Subject Areas: Marine Transportation; Vehicles and Equipment;

Filing Info

• Accession Number: 01981037
• Record Type: Publication
• Files: TRIS
• Created Date: Feb 24 2026 9:02AM