CASIA OpenIR  > 毕业生  > 博士学位论文
Alternative TitleStudy on Type-2 Fuzzy Logic Systems and Type-2 Fuzzy Control Methods
Thesis Advisor易建强
Degree Grantor中国科学院研究生院
Place of Conferral中国科学院自动化研究所
Degree Discipline控制理论与控制工程
Keyword二型模糊系统 神经网络 单输入规则模块 遗传算法 先验知识 词计算 感知推理方法 Type-2 Fuzzy Logic System Neural Network Single Input Rule Module Genetic Algorithm Prior Knowledge Computing With Words Perceptual Reasoning
Abstract作为一型模糊系统及其控制方法的一种改进, 二型模糊系统及其控制方法不但能够有效地处理复杂、非线性、不精确系统, 而且在处理系统不确定性、减少模糊规则数目、抗干扰等方面都具有明显的优越性. 由于在实际应用中, 采样数据的不确定性、外界对系统的各种干扰、未建模动态等广泛存在, 二型模糊系统及其控制方法的研究显得更加必要. 因此, 近年来在国际上, 二型模糊系统及其控制方法正逐渐成为一大研究热点. 但国内外与二型模糊系统及其控制方法相关的研究还不成熟, 还存在很多理论与技术问题需要解决. 因此, 本课题结合863计划项目“实现载荷水平调节的四绳索垂直牵引装卸机器人研制”(2007AA04Z239), 国家自然科学基金项目“智能控制和计算智能的方法及应用(60621001)” 及“基于多源知识的二型模糊系统设计与控制研究(60975060)”, 进一步深入探讨与二型模糊系统设计相关的问题, 研究基于二型模糊系统的新控制方法, 拓展二型模糊系统及其控制方法的应用领域. 本文主要的工作和贡献有: 1、针对二型模糊神经网络参数优化问题, 提出了基于反向传播算法与最小二乘算法的混合算法. 设计了基于二型模糊神经网络的直接自适应控制方案以及基于二型模糊神经网络的逆控制方案, 并采用这两种方案分别实现了多容水箱的液位控制和绳索牵引自动水平调节吊具系统的水平调节控制. 2、为减少模糊规则数、降低二型模糊系统的运算复杂性与模糊规则设计的难度, 提出了基于单输入规则模块的二型模糊系统设计方法. 分析了基于单输入规则模块的二型模糊系统与Takagi-Sugeno (TS)一型模糊系统、TS二型模糊系统的关系; 讨论了基于单输入规则模块的二型模糊控制系统的优化与稳定性问题; 在倒立摆系统、平移振荡器系统、拖车系统上进行了仿真实验, 验证了所提方法的控制性能. 3、为改善所设计的二型模糊系统的性能, 提出了融合先验知识的二型模糊系统设计方法. 证明了在先验知识约束下模糊规则前件与后件参数应满足的约束条件; 给出了基于先验知识与训练数据的多源二型模糊系统的设计过程; 通过仿真实验证实了融合先验知识的二型模糊系统设计方法的有效性及优越性. 4、给出并证明了在二型词计算中所用到的感知推理方法(PerceptualReasoning Method)的一些新性质. 讨论了梯形二型模糊集合斜率一致化问题,并给出了相应的算法流程.
Other AbstractAs an extension of traditional fuzzy logic (type-1 fuzzy logic), type-2 fuzzy logic has obvious advantages for handling different sources of uncertainties, reducing the number of fuzzy rules, etc., as type-2 fuzzy logic utilizes type-2 fuzzy sets which can provide additional degrees of freedom and have more parameters compared with traditional fuzzy sets (type-1 fuzzy sets) in type-1 fuzzy logic. In real-world applications, the fact that uncertainties can not be avoided makes the study of type-2 fuzzy logic necessary. Hence, in recent years, type-2 fuzzy logic and its applications have become an active topic in the area of computation intelligence. However, till now, there are still many theoretical and technical problems to be resolved in type-2 fuzzy logic. Under the support of the 863 Program (2007AA04Z239) and the National Natural Science Foundations of China (60621001, 60975060), in this thesis, novel properties and design methods of type-2 fuzzy logic systems (T2FLSs) are studied and new control schemes are developed. The main contributions of this thesis include the following issues: 1. A hybrid algorithm, which is a combination of the back-propagation (BP) algorithm and the least squares algorithm, is proposed for the parameter learning of the type-2 fuzzy neural network (T2FNN). A T2FNN based direct adaptive control scheme is developed and applied in the liquid level control process of the coupled-tank system. Also, a T2FNN based inverse control strategy is proposed to achieve the level adjustment of the cable-driven self-leveling crane system. 2. To make the design of T2FLSs easier, the single-input-rule-modules (SIRMs) based T2FLSs (SIRM-T2FLSs) are proposed. The properties of the SIRM-T2FLS are investigated. These properties are focused on clarifying its relationships with other fuzzy logic systems (e.g. Takagi-Sugeno type-1 fuzzy logic system, Takagi-Sugeno T2FLS). And, both the parameter optimization and stability analysis problems of the SIRMs based type-2 fuzzy control systems are studied. Also, the SIRM-T2FLSs are utilized to control the inverted pendulum system, the TORA system and the truck-trailer system. 3. To improve the performance of the designed T2FLSs, a new method, which can incorporate prior knowledge into the design process of T2FLSs, is proposed. Sufficient conditions on the antecedent and consequent parameters of T2FLSs are given to ensure that the prior knowledge can be encoded. The design process of the...
Other Identifier200718014628009
Document Type学位论文
Recommended Citation
GB/T 7714
李成栋. 二型模糊系统及其控制方法研究[D]. 中国科学院自动化研究所. 中国科学院研究生院,2010.
Files in This Item:
File Name/Size DocType Version Access License
CASIA_20071801462800(4613KB) 暂不开放CC BY-NC-SAApplication Full Text
Related Services
Recommend this item
Usage statistics
Export to Endnote
Google Scholar
Similar articles in Google Scholar
[李成栋]'s Articles
Baidu academic
Similar articles in Baidu academic
[李成栋]'s Articles
Bing Scholar
Similar articles in Bing Scholar
[李成栋]'s Articles
Terms of Use
No data!
Social Bookmark/Share
All comments (0)
No comment.

Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.