中文摘要：

为准确高效地预测船舶在波浪中的航行状态以保证人员、货物和船舶的安全,提出一种基于减聚类的自适应神经模糊推理系统（Subtractive Clustering based Adaptive Neural-Fuzzy Inference System,SC-ANFIS）模型.SC-ANFIS模型使用减聚类算法对输入样本进行聚类分析,得到模糊规则数,并建立神经模糊推理系统,再使用结合BP算法与最小二乘估计算法的混合算法对建立的预测系统进行优化训练,得到最优的预测系统模型,并使用自相关分析确定预测系统模型的输入.运用该模型对大连海事大学科研教学船“育鲲”号的横摇运动进行实时预测,结果验证了该方法可行、有效,并具有较高的预测精度.

英文摘要：

In order to predict navigation states of ships in wind and waves accurately and efficiently and ensure the safety of personnel, cargoes and ships, a Subtractive Clustering based Adaptive Neural-Fuzzy Inference System （SC-ANFIS） model is proposed. In the model, the subtractive clustering algorithm is used to carry out clustering analysis on input samples so as to obtain the fuzzy rule number, and then the neural-fuzzy inference system is built. The hybrid algorithm combining BP algorithm with the least squares estimation algorithm is used to train and optimize the system so as to establish an optimal predic-tion system model. The input of the prediction system model is determined by auto-correlation analysis.The rolling motion of the ship Yukun （a scientific research and training ship of Dalian Maritime Universi-ty） is predicted in real time by the model. The prediction results demonstrate that the proposed method is feasible, effective and of high prediction accuracy.