应用贝叶斯岭回归模型辨识水下航行器水动力系数研究  

Research of applying Bayesian ridge regression model in identification of hydrodynamic coefficients of underwater vehicles

作　　者：向国 欧勇鹏[1] 陈君杰 王威[1] XIANG Guo;OU Yongpeng;CHEN Junjie;WANG Wei(College of Naval Architecture and Ocean Engineering,Naval Univ.of Engineering,Wuhan 430033,China)

机构地区：[1]海军工程大学舰船与海洋学院,武汉430033

出　　处：《海军工程大学学报》2025年第1期104-112,共9页Journal of Naval University of Engineering

基　　金：国家部委基金资助项目(41407020602)。

摘　　要：针对水下航行器水动力系数求取问题,研究了应用贝叶斯岭回归模型进行水动力系数辨识的方法。首先,建立水下航行器六自由度运动模型并令其以指定方式运动,经白噪声调和、数据采样得到航行器空间运动仿真试验数据;然后,利用贝叶斯岭回归模型辨识获得一组新的水动力系数,并据此预报了航行器水平面Z形、回转运动;最后,通过与原模型的水动力系数、运动轨迹的对比分析发现:除敏感度不高的少数耦合项外,应用贝叶斯岭回归模型辨识水动力系数均具有良好的精度;相较于原模型,辨识所得一阶水动力系数偏差小于5%、二阶系数偏差小于20%,Z形、回转运动参数预报偏差小于6%。In order to obtain the hydrodynamic coefficients of underwater vehicles,a method using Bayesian ridge regression model for coefficient identification was explored.Firstly,a six-degree-of-freedom motion model of the underwater vehicle was established and it was set to move in a specified way.The simulator test data of the vehicle′s spatial motion was obtained through white noise harmonization and data sampling.And then,a new set of hydrodynamic coefficients was identified using the Bayesian ridge regression model,and the vehicle′s zigzag and rotational motion were predicted on that basis.Through comparative analysis with the original model′s hydrodynamic coefficients and motion trajectory,it is found that except for a few insensitive coupling terms,the application of the Bayesian ridge regression model for identifying hydrodynamic coefficients has good accuracy.Compared with the original model,the identified first-order hydrodynamic coefficients has a deviation of less than 5%,the second-order coefficients has a deviation of less than 20%,and the predicted deviations of the zigzag and turning motion parameters are less than 6%.

关 键 词：水动力系数 参数辨识 贝叶斯岭回归模型 操纵运动模型 

分 类 号：U674.76[交通运输工程—船舶及航道工程]