基于最小二乘支持向量机的船舶操纵运动建模

Modeling of Ship Manoeuvring Motion Using Least Squares Support Vector Machines

基于自由自航船模试验或实船试验的系统辨识方法是一种确定船舶操纵运动水动力导数的有效方法。通过对舵角、漂角、转首角速度等试验数据的分析,用最小二乘支持向量机确定了船舶操纵运动数学模型中的水动力导数及其干扰力系数,其中非线性模型的参数辨识采用了多项式核函数。利用辨识得到的参数进行了操纵运动预报仿真并同自航模试验及实船试验数据对比,数值仿真结果验证了方法的有效性。

System identification combined with free-running model tests or lull-scale trials is one ot the effective methods to determine the hydrodynamic derivatives in the mathematical models of ship manoeuvring motion. By analyzing the available data including rudder angles, drift angles and yaw rates, a method based on least squares support vector machines for determining the hydrodynamic derivatives and disturbance coejjicients was proposed. Polynomial kernel function was adopted while identifying the parameters in the nonlinear model. Predictions of manoeuvring motion were proposed using the parameters identified. The results of identification and simulation demonstrate the validity of the identification algorithm proposed.