Parameter Identification of Ship Manoeuvring Model Based on Particle Swarm Optimization and Support Vector Machines

The method of system based manoeuvring simulation provides an effective way to predict ship manoeuvrability. Accurate determination of the hydrodynamic derivatives in the mathematical model of ship manoeuvring motion is vital to the prediction accuracy. A support vector machines (SVM) based approach is proposed in this paper. By analyzing the data from free-running model tests of KVLCC2 ship, the hydrodynamic derivatives in an Abkowitz model are identified. To diminish the parameter drift in the identification, a difference method is adopted to reconstruct the sample for identification. To obtain the optimized structural parameters in SVM, particle swarm optimization (PSO) method is incorporated into SVM. Predictions of manoeuvring motion are presented based on the regression model. Comparisons between the predicted results and the test results demonstrate the validity of the proposed methods.