Blind prediction of ship manoeuvring motion in 4 degrees of freedom based on support vector machines

Based on the whole ship model, a blind prediction method by using support vector machines is proposed for the 4 degrees of freedom ship manoeuvring motion including roll motion. The blind prediction is achieved relying only on the ship motion state parameters. 10°/10°, 20°/20° zigzag tests and 35° turning circle manoeuvre are simulated for a container ship. One percent of the simulation data of 20°/20° zigzag test are used to train the support vectors; and the trained support vector machines is used to predict the whole 20°/20° zigzag test. Besides, 10°/10° zigzag test and 35° turning circle manoeuvre are also predicted. The predicted results are compared with those of simulation tests to verify the prediction method and to demonstrate its good generalization performance.