What does this research mean for the field?

The Non-Markovian Memory-Augmented Environment-Perceived and Physics-Informed Network (NMA-EPIN) provides highly accurate and robust ship motion prediction, maintaining an R2 of approximately 0.91 even under extreme sensor noise, packet loss, and delays by enforcing kinematic consistency and modeling hydrodynamic memory. Novelty: ClaimNovelty.METHODOLOGICAL. Consensus alignment: ConsensusAlignment.NEUTRAL.

What question did this study set out to answer?

This research aims to develop a robust model for ship motion prediction that can handle non-ideal observations and capture hydrodynamic memory effects.

June 9, 2026Open Access

A Physics-Informed Non-Markovian Deep Learning Model for Robust Ship Motion Prediction Under Non-Ideal Observations

Key Points

This research aims to develop a robust model for ship motion prediction that can handle non-ideal observations and capture hydrodynamic memory effects.
Proposes the Non-Markovian Memory-Augmented Environment-Perceived and Physics-Informed Network (NMA-EPIN) for prediction.
Utilizes a Control-Physics-Informed Neural Network (CPINN) to ensure kinematic and control constraints.
Evaluates the model's performance against various conditions, including extreme sensor noise and delays.
Achieves an average coefficient of determination R2=0.977 under nominal conditions.
Retains R2≈0.91 under extreme perturbations with 50% sensor noise and delays.
Significantly outperforms baseline models in terms of motion prediction accuracy.

Abstract

High-fidelity ship dynamics models are essential for the reliable operation of maritime autonomous systems. However, existing Markov-based maneuvering models and purely data-driven predictors struggle to capture hydrodynamic memory and degrade under non-ideal sensing. To address these challenges, this paper proposes a novel approach for robust ship motion prediction, the Non-Markovian Memory-Augmented Environment-Perceived and Physics-Informed Network (NMA-EPIN). This method explicitly models long-term hydrodynamic dependencies through a memory-augmented architecture. Within NMA-EPIN, a Control-Physics-Informed Neural Network (CPINN) paradigm enforces velocity–position kinematic consistency and control-logic alignment as soft constraints, suppressing cumulative drift under degraded observations. Experiments on a high-fidelity simulated dataset show that NMA-EPIN attains an average coefficient of determination R2=0.977 under nominal conditions, effectively eliminating the position drift observed in baselines. Under extreme compound perturbations (50% sensor noise, packet loss, and delays), NMA-EPIN retains R2≈0.91, which significantly outperforms the baselines.

Read Full Paperexternally

KI fragen

Bookmark

View Full Paper