| 基于Elman网络的船舶运动模型辨识 |
| 投稿时间：2013-05-08 修订日期：2013-07-27 点此下载全文 |
| 引用本文：孙洪波,施朝健.基于Elman网络的船舶运动模型辨识[J].上海海事大学学报,2014,35(1):10-13. |
| 摘要点击次数: 2041 |
| 全文下载次数: 770 |
|
|
| 基金项目: 国家自然科学基金（51109090） |
|
| 中文摘要:为寻求一种简便的船舶运动模型辨识方法，根据船舶动力学与运动学基本方程的结构形式,建立一种基于Elman神经网络的辨识模型,给出网络结构的选取和确定方法.以载质量为5万t的散货船为例,利用国际海事组织要求的几个典型的船舶操纵试验数据对网络进行训练,计算权值矩阵,获得该船舶可用于船舶操纵性分析的神经网络模型.将网络计算结果代入船舶运动学方程进行船舶航迹仿真,并与试验航迹数据进行对比, 验证网络模型的精确性.比较仿真验证结果和试验数据可知,该网络模型能基本反映被辨识船舶的动态特性,验证其有效性和准确性. |
| 中文关键词:船舶动力学模型 船舶运动学方程 神经网络 系统辨识 |
| |
| Ship motion model identification based on Elman network |
|
|
| Abstract:To find out a simple ship motion model identification method, an identification model based on Elman neural network is established according to the structural form of ship dynamic and kinematic equations, and the selection and determination methods to network structure are provided. Taking a bulk carrier with dead weight tons of 50 000 t as an example, the network is trained by several typical ship maneuvering trail data required by International Maritime Organization, and the associated weighted matrix is calculated, and finally the neural network model for the ship maneuverability analysis is obtained. The network calculation results are substituted into ship kinematics equations to simulate the ship track, and the simulation result is compared with the trial track to verify the accuracy of the network model. The comparison result shows that the network model can basically reflect ship dynamic characteristics, and its effectiveness and accuracy are verified. |
| keywords:ship dynamic model ship kinematic equation neural network system identification |
| 查看全文 查看/发表评论 下载PDF阅读器 |
| 关闭 |
|
|
|
|
|
