基于整体型船舶操纵运动数学模型,结合船舶操纵运动中横倾运动的特点,建立了应用支持向量机的四自由度船舶操纵运动黑箱预报模型。对某集装箱船进行了10°/10°、15°/15°、20°/20°Z形试验和5°、15°、35°回转试验的数值仿真,获得了操纵运动仿真数据;使用1%的15°/15°、20°/20°Z形试验和5°、15°回转试验仿真数据来训练支持向量,并用训练好的支持向量机预报10°/10°、20°/20°Z形试验和35°回转试验;将预报结果与仿真试验数据进行对比,验证了所提出黑箱预报模型的有效性及良好的泛化性能。
Based on a whole ship model and heeling motion characteristics in ship maneuvering motion, a black-box prediction model for ship maneuvering motion in 4 degrees of freedom is established by using support vector machines. A container ship is taken as study object; while 10°/10°, 15°/15°, 20°/20° zigzag tests and 5°, 15°, 35° turning circle maneuvers are simulated. One percent of the simulation data in 15°/15°, 20°/20° zigzag tests and 5°, 15° turning circle maneuvers are used to train the support vectors. The trained support vector machines is used to predict the 10°/10°, 20°/20° zigzag tests and 35° turning circle maneuver. The predicted results are compared with those of simulation tests to demonstrate the good prediction capability and generalization performance of the proposed model.