非线性自适应逆控制方法研究
文献类型:学位论文
作者 | 柳晓菁 |
学位类别 | 工学博士 |
答辩日期 | 2006-05-27 |
授予单位 | 中国科学院研究生院 |
授予地点 | 中国科学院自动化研究所 |
导师 | 易建强 |
关键词 | 自适应逆 径向基函数网络 模糊神经网络 支持向量机 桥式吊车 adaptive inverse Radial based function networks Fuzzy neural networks Support vector machine Overhead crane |
其他题名 | Research on the Approach of Nonlinear Adaptive Inverse control |
学位专业 | 控制理论与控制工程 |
中文摘要 | 自适应逆控制是用自适应滤波器辨识出被控对象的逆模型,并将其串联到对象的输入端作为控制器来控制对象的动态特性。这种开环控制避免了因为不恰当的反馈引起的系统不稳定现象,而且分别处理系统的动态特性控制和对象扰动的控制问题,从而最大限度地获得好的控制性能。本文以普通非线性对象为研究对象,结合中科院百人计划“智能控制方法及应用研究”,展开非线性自适应逆控制方法研究。所完成的工作主要包括以下七个方面: 首先,介绍了自适应逆控制的概念、选题意义,综述了国内外对自适应逆控制方法和桥式吊车控制的研究进展,并给出了论文的主要内容。 其次,介绍了径向基函数(RBF)网络、模糊神经网络(FNN)和支持向量机(SVM)这三种通用的逼近器的模型、网络结构和学习算法,为后面非线性自适应逆控制的建模和控制器的设计以及扰动消除器的设计做了充分的准备。 第三,提出了一种基于RBF网络的 -滤波非线性自适应逆控制系统,即用一种RBF网络的非线性自适应滤波器为非线性系统进行建模、逆建模、控制器及自适应扰动消除器的设计。 第四,提出了一种基于FNN的非线性自适应逆控制系统,将三个基于FNN的非线性滤波器用在非线性对象建模、控制器的设计和自适应扰动消除器中。 第五,讨论了基于贝叶斯证据框架下最小二乘支持向量机回归(LS-SVMR)的建模和逆建模方法,提出了一种基于LS-SVMR的自适应逆扰动消除控制系统和一种基于LS-SVMR的混合自适应逆控制系统,在此基础上构建了完整的自适应逆控制系统。 第六,将非线性自适应逆控制应用到桥式吊车中,提出了一种桥式吊车的自适应逆控制方法。首先设计了一个角度PD反馈控制环节来保证摆角的迅速衰减,然后又设计了一个PD反馈控制环节来使吊车对象稳定。最后针对这个双闭环被控对象设计了一种基于RBF网络的 -滤波自适应逆控制系统。 最后,对取得的研究成果进行了总结,并展望了需要进一步研究的工作。 |
英文摘要 | Adaptive inverse control uses adaptive filter to identify the inverse model of the plant to be controlled and puts the inverse model in front of the plant as a controller to control plant dynamics. This open-loop control not only avoids the instability of control system caused by unsuitable feedback loop, but also separately controls plant dynamics and plant disturbance. So it can obtain good control performance. In this dissertation, the adaptive inverse control of a class of common nonlinear plants is studied with the support of the Hundred Talents Program of Chinese Academy of Sciences “Researches of Intelligence Control Methods and Their Applications”. The content of the research includes the following seven parts: Firstly, the conception and research purpose are introduced. The research status of control of adaptive inverse control and overhead crane control are comprehensively surveyed. The main work of this dissertation is gieven. Secondly, the model, structure and learning algorithm of the three universal approach method: radial basis function (RBF) networks, fuzzy neural networks (FNN) and support vector machines (SVM) are introduced. These are fully prepared for the design of model, controller and disturbance canceller. Thirdly, a filtered- nonlinear adaptive inverse control system based on RBF networks is proposed, where a nonlinear adaptive filter based on RBF networks is used for modelling, inverse modelling of nonlinear plant, designing of controller and adaptive disturbance canceller. Fourthly, a nonlinear adaptive inverse control system base on FNN is presented, where three nonlinear filters based on FNN are used for nonlinear plant modeling, designing of controller and adaptive disturbance canceller. Fifthly, an approach of modeling and inverse modeling based on least squares support vector machine regression (LS-SVMR) within the Bayesian evidence framework is discussed. An adaptive disturbance canceling system based on LS-SVMR and a hybrid adaptive inverse control system are proposed. Based on this two systems, an integrated adaptive inverse control system is established. |
语种 | 中文 |
其他标识符 | 200318014602978 |
源URL | [http://ir.ia.ac.cn/handle/173211/5910] ![]() |
专题 | 毕业生_博士学位论文 |
推荐引用方式 GB/T 7714 | 柳晓菁. 非线性自适应逆控制方法研究[D]. 中国科学院自动化研究所. 中国科学院研究生院. 2006. |
入库方式: OAI收割
来源:自动化研究所
浏览0
下载0
收藏0
其他版本
除非特别说明,本系统中所有内容都受版权保护,并保留所有权利。