• CN:11-2187/TH
  • ISSN:0577-6686

›› 2008, Vol. 44 ›› Issue (4): 113-117.

• 论文 • 上一篇    下一篇

惩罚函数的构造及多模态平稳过渡策略

胡旭晓;潘晓弘;何卫;陈罡   

  1. 浙江大学现代制造工程研究所;杭州机床集团有限公司;浙江纺织服装学院
  • 发布日期:2008-04-15

Construction of Penalty Function and Multi-modal Smooth Transition Strategy

HU Xuxiao;PAN Xiaohong;HE Wei;CHEN Gang   

  1. Institute of Modern Manufacturing Engineering, Zhejiang University Hangzhou Machine Tool Group Co., Ltd. Zhejiang Texile & Fashion College
  • Published:2008-04-15

摘要: 为了使智能控制中多模态之间能够实现平稳过渡,结合一种多模态的划分方法,构造一类新的惩罚函数,并提出一种基于惩罚函数的多模态平稳过渡策略。通常,不同的控制策略针对不同的模态,而基于惩罚函数的多模态平稳过渡策略把不同的控制策略推广到所有模态,控制量是所有控制策略的加权,惩罚函数的作用是使与自己模态相对应的控制策略所含权值较大,另外模态的控制策略在此模态中所含权值都受到惩罚。最后,应用基于惩罚函数的多模态平稳过渡策略,在具有大进给力的纳米级驱动部件上进行试验,证明了所提出的平稳过渡策略的有效性。

关键词: 惩罚函数, 构造, 平稳过渡, 智能控制

Abstract: In order to realize steady transition between the modes in intelligent control, a new penalty function is found combined with a kind of multi-mode divided method, and a multi-mode steady transition strategy is proposed based on the penalty function. Different from the usual way that each control strategy in view of each mode, different control strategies are promoted to all modes in the multi-mode steady transition strategy, the control quantity is the sum of all control strategies by weight, the function of the penalty function is made the weight of corresponding control strategy bigger, and that of other control strategies is punished in this mode. Finally, the steady transition strategy is confirmed by the experiment on the nano precision drive parts which have large feeding force.

Key words: Construct, Steady transition, Intelligent control, Penalty function

中图分类号: