计算机工程与应用 ›› 2012, Vol. 48 ›› Issue (16): 233-236.

• 工程与应用 • 上一篇    下一篇

基于RBF神经网络与Smith预估补偿的智能PID控制

王菲菲,陈  玮   

  1. 上海理工大学 光电信息与计算机工程学院,上海 200093
  • 出版日期:2012-06-01 发布日期:2012-06-01

Intelligent PID control based on RBF neural network and Smith predictive compensation

WANG Feifei, CHEN Wei   

  1. Department of Photoelectric Information & Computer Engineering, Shanghai Ligong University, Shanghai 200093, China
  • Online:2012-06-01 Published:2012-06-01

摘要: 针对工业控制中普遍存在的大滞后现象,提出了一种将RBF神经网络算法和Smith预估补偿算法与传统的PID控制器相结合的智能RBF-Smith-PID控制策略。该方法利用RBF神经网络的在线学习、控制参数自整定能力,和Smith预估补偿对纯滞后系统的良好控制,有效地克服了常规PID控制的缺陷,提高了系统的鲁棒性和自适应性,对纯滞后系统起到了良好的控制。

关键词: 纯滞后, 神经网络, Smith预估控制, PID控制

Abstract: Aiming at the phenomena of big time delay are normally exist in industry control, this paper proposes an intelligent RBF-Smith-PID control based on RBF neural network algorithm and Smith predictive compensation algorithm and traditional PID controller. This method uses the ability of online-study, a self-turning control strategy of RBF neural network, and better control of Smith predictive compensation to deal with the big time delay, overcome the limitation of traditional PID control effectively, improve the system’s robustness and self-adaptability, get satisfactory control to deal with the big time delay system.

Key words: time-delay, neural network, Smith predictive control, PID control