| 水下机器人运动系统的神经网络辨识 |
| |
| 引用本文： | 刘云霞,林孝工,王智学.水下机器人运动系统的神经网络辨识[J].机械工程与自动化,2006(2):81-83,89. |
| |
| 作者姓名： | 刘云霞 林孝工 王智学 |
| |
| 作者单位： | 哈尔滨工程大学,动力与核能工程学院,黑龙江,哈尔滨,150001 |
| |
| 摘 要： | 建立了水下机器人的动力模型，分析了辨识该模型的神经网络结构，采用带自反馈的Elman网络来获得更精确的结果。针对BP算法即误差反传算法的缺陷，提出了用混合优化算法——误差反传算法和遗传算法的混合算法（又称：GA＆BP算法）修正网络权值。最后，将改进的Elman网络应用于水下机器人的非线性辨识。通过仿真证明了该方法用于高阶非线性系统的实用性。
|
| 关 键 词： | 系统辨识 水下机器人 神经网络 |
| 文章编号： | 1672-6413（2006）02-0081-03 |
| 收稿时间： | 2005-10-19 |
| 修稿时间： | 2005-10-19 |
Identification of Dynamic Model of Underwater Vehicle Based on Dynamic Recurrent Neural Network |
| |
| LIU Yun-xia,LIN Xiao-gong,WANG Zhi-xue.Identification of Dynamic Model of Underwater Vehicle Based on Dynamic Recurrent Neural Network[J].Mechanical Engineering & Automation,2006(2):81-83,89. |
| |
| Authors: | LIU Yun-xia LIN Xiao-gong WANG Zhi-xue |
| |
| Affiliation: | College of Power and Nuclear Energy,Harbin Engineering University,Harbin 150001 ,China |
| |
| Abstract: | The hydrodynamic model of an underwater vehicle(UV) is established in this paper,and the neural network(NN) structure is introduced to identify this hydrodynamic model.To obtain better accuracy,this paper adopted part-recursive Elman NN which has a layer with a self-feedback modulus to remember the former output of the hidden-layer.To overcome the slow convergence of the BP algorithm,BP algorithm and genetic algorithm(BP&GA) is proposed,which can train the weight of the network..In the end the motion simulation is carried out to verify the reliability of the nonlinear hydrodynamic model. |
| |
| Keywords: | underwater vehicle system identification neural network(NN) |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
