基于稀疏高斯过程回归的船舶动力学模型辨识  

Identification of ship dynamics model based on sparse Gaussian process regression

作　　者：陈刚 王威[1] 霍聪[1] CHEN Gang;WANG Wei;HUO Cong(College of Naval Architecture and Ocean Engineering,Naval Univ.of Engineering,Wuhan 430033,China;Low Speed Inst.,China Aerodynamics Research and Development Center,Mianyang 621000,China)

机构地区：[1]海军工程大学舰船与海洋学院,武汉430033 [2]中国空气动力研究与发展中心低速空气动力研究所,四川绵阳621000

出　　处：《海军工程大学学报》2022年第5期84-89,共6页Journal of Naval University of Engineering

摘　　要：船舶动力学模型对于自主航行船舶的运动规划和控制具有十分关键的作用。针对高斯过程回归建模计算复杂度高的问题,提出基于稀疏高斯过程回归的船舶动力学模型辨识算法,利用相似度对大样本进行稀疏,改善高斯过程回归难以应用于大样本数据学习的缺陷,以KVLCC2船舶的试验数据验证所提方法的有效性。结果表明:稀疏高斯过程缩短了模型计算时间,得到了比参数化模型精度更高的船舶运动预报。The ship dynamics model plays a key role in the motion planning and control of autonomous ships. In view of the high computational complexity of Gaussian process regression modeling, a ship dynamics model identification algorithm based on sparse Gaussian process regression was proposed, in which large samples were scattered by similarity so as to alleviate the defect that Gaussian process regression was difficult to apply to large sample data learning. The experimental data of KVLCC2 ship were used to verify the validity of the proposed method. The results show that the sparse Gaussian process shortens the calculation time of the model, and obtains ship motion prediction with higher accuracy than the parameterized model.

关 键 词：船舶动力学 稀疏高斯过程 辨识 相似度 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置]