基于支持向量机的船舶操纵运动在线建模(英文)  

On-line Modeling of Ship Maneuvering Motion based on Support Vector Machines

作　　者：徐锋[1] 邹早建[1,2] 尹建川[1] 

机构地区：[1]上海交通大学船舶海洋与建筑工程学院,上海200240 [2]上海交通大学海洋工程国家重点实验室,上海200240

出　　处：《船舶力学》2012年第3期218-225,共8页Journal of Ship Mechanics

基　　金：Supported by the National Natural Science Foundation of China (50979060,51079031)

摘　　要：支持向量机算法以结构风险最小化为准则,具有良好的泛化性,是一种先进的人工智能算法。文章根据最小二乘支持向量机对其增量式算法进行推导,并应用该算法对船舶操纵运动进行在线建模。通过仿真试验对粘性力和舵力水动力导数进行在线辨识,通过自航模试验对操纵性K、T指数进行在线辨识,并利用K、T指数对自航模的转首角速度进行了预报,所得结果都同试验值非常接近,证明了增量式最小二乘支持向量机应用于船舶操纵运动在线建模的有效性。Support Vector Machines （SVM） is an advanced artificial intelligence （AI） algorithm. It is based on the criteria of structural risk minimization （SRM） and has high generalization performance. In this paper, incremental least square Support Vector Machines （LS-SVM） is deduced and applied to on-line modeling of ship maneuvering motion. The viscous force derivatives and rudder force derivatives in the linear Abkowitz model are identified on-line based on the simulation test. K and T indexes in the first-order linear response model are identified on-line based on the free running model test, and yaw rate is predicted by the identified response model. The results of identification and prediction are all in good agreement with the model test data, which demonstrates the validity of incremental LS-SVM in on-line modeling of ship maneuvering motion.

关 键 词：支持向量机 船舶操纵 仿真试验 自航模试验 

分 类 号：U661.31[交通运输工程—船舶及航道工程]