Identification modeling of ship nonlinear motion based on nonlinear innovation

In order to further improve the accuracy of parameter identification of ship nonlinear motion mathematical model, grey wolf optimizer-support vector regression (GWO-SVR) is embedded into the nonlinear innovation processed by hyperbolic tangent function. A novel identification scheme called nonlinear innovation GWO-SVR (NGWO-SVR) is proposed for identification of ship nonlinear motion mathematical model. Based on the full-scale trial data of vessel YUKUN, the NGWO-SVR is used to identify the parameters of the nonlinear motion mathe-matical model. Using full-scale trial data different from the identified samples, the generalization of the pa-rameters obtained from SVR, GWO-SVR and NGWO-SVR are verified. From the evaluation indicator of the prediction results, the parameters identified by the proposed algorithm are more generalized under the condition of effectiveness of certain accuracy. In addition, the indirect sensitivity analysis method is used to analyze the parameters of the nonlinear motion mathematical model. The research results can provide a priori characteristics for intelligent ship motion control and path planning and a reference scheme for the optimization of ship motion mathematical model in marine simulator.

Automated summary

In order to improve the accuracy of parameter identification of ship nonlinear motion mathematical model, a novel identification scheme called nonlinear innovation GWO-SVR (NGWO-SVR) is proposed by embedding grey wolf optimizer-support vector regression (GWO-SVR) into the nonlinear innovation. The NGWO-SVR is used to identify parameters based on full-scale trial data of vessel YUKUN and the generalization of the identified parameters is verified using different trial data. The research results also include indirect sensitivity analysis of the parameters and provide reference for intelligent ship motion control and path planning.