Identification Modeling of Ship Maneuvering Motion Based on Echo State Gaussian Process

The mathematical model of ship maneuvering motion is a crucial foundation for both prediction of ship navigation state and automatic control of ship motion. This paper proposes a practical and robust nonparametric identification method based on echo state Gaussian process (ESGP) to establish the nonparametric model of ship maneuvering motion. The traditional echo state networks (ESNs) and basic Gaussian processes (GPs) have found successfully applications in nonparametric modeling and time-series forecasting. The proposed method combines the strengths of both ESN and GP approaches. On the one hand, it offers a more robust alternative to conventional reservoir computing networks; on the other hand, it can directly generate confidence intervals for prediction results. The KVLCC2 ship model is taken as the object of this research. The datasets collected from several standard and nonstandard zigzag maneuvers of free-running model, which are provided by SIMMAN 2008, are used as training and testing data. To assess the robustness and generalization ability of the established nonparametric model, the prediction results of the ESGP method are compared with the experimental data. It is shown that the ESGP method proposed in this paper can achieve high prediction accuracy and provide a measure of credibility for the output results, which makes it more practical and applicable for modeling and prediction of ship maneuvering motion.