| 局部递归神经网络控制器及其应用 |
| |
| 引用本文： | 姬北辰,丛爽,吴汉生.局部递归神经网络控制器及其应用[J].计算机工程与设计,2004,25(12):2170-2172. |
| |
| 作者姓名： | 姬北辰 丛爽 吴汉生 |
| |
| 作者单位： | 中国科学技术大学,自动化系,安徽,合肥,230027 |
| |
| 基金项目： | 安徽省自然科学基金项目(03042301)
中国科技大学高等教育和教学研究计划基金项目(2003-2004) |
| |
| 摘 要： | 基于人工神经网络提出了一种局部递归神经网络控制器。在描述了带有输出反馈和激活反馈的网络控制器的结构组成并定义了作为设计目标的误差函数后，采用带有弹性的梯度下降法，获得适用于实时在线调整权值的修正公式，给出了所提的网络控制器的设计步骤及其控制策略。将所提出的网络控制器应用到典型的单级倒立摆的实验系统中，将实验所获得的结果与LQY方法的实验结果进行了对比。
|
| 关 键 词： | 网络控制器 递归神经网络 倒立摆 实时 输出反馈 梯度下降法 权值 局部 激活 实验结果 |
| 文章编号： | 1000-7024(2004)12-2170-03 |
Locally recurrent neural network controller and its application |
| |
| JI Bei-chen,CONG Shuang,WU Han-sheng.Locally recurrent neural network controller and its application[J].Computer Engineering and Design,2004,25(12):2170-2172. |
| |
| Authors: | JI Bei-chen CONG Shuang WU Han-sheng |
| |
| Abstract: | A locally recurrent neural network controller based on neural network is proposed. After describing the structure of the neural controller with output feedback and activation feedback and defining the cost function as a designed goal, the method of resilient back-propagation is used to obtain the learning rules which are used to update weights on-line. Then the design procedures and control strategy of proposed neural controller are given. Finally, the proposed neural controller is applied to a typical practical system of single inverted pendulum, and the results of experiment are compared with those of LQY control. |
| |
| Keywords: | artificial neural network controller stability learning rate |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
