节点文献
前馈网络控制器的研究
The Study of Feedforward Neural Networks for Controller
【摘要】 控制器用前馈网络有其特殊性,本文对控制器用前馈网络的结构、神经元特性函数进行了分析,并通过前馈网络控制器闭环实验验证了分析的正确性。确定了不同情况下控制器用前馈网络结构和神经元特性函数,给出了实验结论。研究结果对前馈网络控制系统设计有指导意义。
【Abstract】 The feedforward neural networks for controller had special features. This paper did detailed analyses to the structure of the neural networks and the activation of neurons for controller. According to the closed loop control experiments, that the above analyses was proved correct. The controller with both good structure of networks and activation of neurons was brought too. At last, it turned out that the research results are of guiding significance to neural networks control system designs.
【关键词】 控制器;
闭环控制;
神经网络;
特性函数;
【Key words】 controller; Closed loop control; Neural networks; Activation of neurons;
【Key words】 controller; Closed loop control; Neural networks; Activation of neurons;
【基金】 中国轻工业科技发展基金!轻科技97032号
- 【文献出处】 电子测量与仪器学报 ,JOURNAL OF ELECTRONIC MEASUREMENT AND INSTROMENT , 编辑部邮箱 ,1999年04期
- 【分类号】TP273
- 【被引频次】5
- 【下载频次】29