ZHANG Jing, PEI Xue-hong, XING Hai-feng. Improved RBF Neural Network Control of Magnetic Levitation[J]. 2011, 16(1): 48-52.DOI:
磁悬浮改进RBF神经网络控制的仿真研究
摘要
利用RBF神经网络的自学习、自适应能力
在传统PID控制基础上
提出一种改进的RBF在线辨识的自适应PID控制方法.为避免系统启停过程中
短时大偏差引起的超调较大
文中利用RBF网络提供辨识信息实现对参数KP、KD调整
对参数KI不做整定
以满足磁悬浮系统的动态和静态性能要求.设计中采用S-函数建立磁悬浮系统的非线性模型
搭建PD参数整定和RBF网络辨识器
并与常规方法进行比较.仿真结果表明
该方法能够有效控制磁悬浮系统
具有更好的自适应性和抗干扰性.
Abstract
As magnetic levitation system has the characteristics of nonlinearity and open-loop instability
it is difficult to achieve ideal effect with conventional control.In this paper
an improved RBF on-line identification method of adaptive PID control is proposed based on traditional PID control and the self-learning
adaptive capacity of RBF neural networks.In order to satisfy the static and dynamic performance requirements of magnetic levitation system and avoid the large overshoot following the short and large ...