| 神经网络模型参考自适应控制算法研究 |
| |
| 引用本文： | 姜向龙,程善美,李叶松,万淑芸.神经网络模型参考自适应控制算法研究[J].华中科技大学学报(自然科学版),2003,31(1):4-6. |
| |
| 作者姓名： | 姜向龙 程善美 李叶松 万淑芸 |
| |
| 作者单位： | 华中科技大学控制科学与工程系 |
| |
| 基金项目： | 国家自然科学基金资助项目 (5 0 10 5 0 0 5 ) |
| |
| 摘 要： | 分析了基于BP算法的神经网络模型参考自适应控制器对大惯性环节被控对象的控制效果,发现该算法使控制器存在严重的“过学习”现象,为避免这一现象,设计了一种新的误差函数结构,得到改进的BP算法,针对一个存在大惯性环节的线性时变系统,对比分析了神经网络模型参考自适应控制器在采用传统的BP算法和改进的BP算法时得到的不同控制效果。
|
| 关 键 词： | 神经网络 模型参考自适应控制 BP算法 误差函数 “过学习”现象 线性时变系统 |
| 文章编号： | 1671-4512(2003)01-0004-03 |
| 修稿时间： | 2002年3月11日 |
Algorithm for model reference adaptive control based on neural network |
| |
| Jiang Xianglong,Cheng Shanmei,Li Yesong,Wan Shuyun.Algorithm for model reference adaptive control based on neural network[J].JOURNAL OF HUAZHONG UNIVERSITY OF SCIENCE AND TECHNOLOGY.NATURE SCIENCE,2003,31(1):4-6. |
| |
| Authors: | Jiang Xianglong Cheng Shanmei Li Yesong Wan Shuyun |
| |
| Abstract: | A neural network model reference adaptive controller based on BP algorithm was applied to a plant with a great inertia link, and its control effect was analyzed. It was found that BP algorithm led to the "over learning" phenomenon in the controller. In order to overcome it, a new error function structure and an improved BP algorithm were presented. For the control of a linear time variable system with a great inertia link, the comparison of the neural network model reference adaptive controller based on conventional BP algorithm with that based on advanced one were made. |
| |
| Keywords: | neural network model reference adaptive controller (MRAC) BP algorithm error function |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
