Please use this identifier to cite or link to this item: http://hdl.handle.net/10397/83752
DC FieldValueLanguage
dc.contributorDepartment of Electrical Engineering-
dc.creatorChan, Ping-tong-
dc.identifier.urihttps://theses.lib.polyu.edu.hk/handle/200/1418-
dc.language.isoEnglish-
dc.titleSymbiosis of fuzzy systems and genetic algorithms for modelling and control-
dc.typeThesis-
dcterms.abstractIn recent years, the integration of Fuzzy Systems (FS) and Genetic Algorithms (GA) has gained a lot of attention and many researchers have advocated exploiting the possibilities of this union. Essentially, FS is to provide GA a platform for the incorporation of linguistic IF-THEN rules and logic reasoning, and GA's role is providing FS with non-linear optimisation. The objectives of the studies undertaken in this thesis are to validate these arguments and to propose modelling and control algorithms for some classes of non-linear systems. The study will focus on problems such as optimisation speed, stability issues before and after optimisation, and high dimension systems. The findings are verified through benchmark simulation and experimental trials. The studies are carried out in three phases. In the first phase, a method to model a plant's data is proposed. An Antecedent Validity Adaptation (AVA) principle is introduced to refine Table Look-up (TL) scheme. The term Table Look-up with Validity (TLV) is coined as a scheme for modelling recorded data. The proposed technique is endowed with the ability of using the dormant and hidden information in the available data. An on-line algorithm is also presented. The design of an Optimised Fuzzy Logic Controller (OFLC) is introduced next. With the aid of genetic algorithms, optimised rules of fuzzy logic controller are designed. The use of symmetric properties to save the number of learning rules is described. The proposed algorithm has been applied to both linear and non-linear systems. This algorithm cuts the vast tangle of the rule-table down to size by incorporating the physical properties of systems and mechanisms of GA. Then OFLC is employed to control a beam-and-ball balance hardware test-bed. A Kalman filer controller (KFC) and a manually tuned fuzzy logic controller (MFLC) are also developed to assess the effectiveness of the OFLC. The proposed controller provides favourite performance in a systematic design manner. In the second phase, a supervisory control technique is applied to an integrated Sliding Mode Control (SMC)/FS. An Indirect Adaptive Fuzzy Sliding Mode Control (IAFSMC) with supervisory controller is proposed which guarantees the boundedness of states of the controlled system and parameter projection to avoid the divergence of the parameters. Along this line, an integrated stable FS+GA system is proposed, in which the FS is optimised by GA. A state supervisory controller, studied in the IAFSMC, is used to guarantee the stability. In the third phase, the use of Hierarchical and Fused FS (HFFS) will be discussed to alleviate the explosion of dimension. The design of an Optimised Hierarchical and Fused Fuzzy System (OHFFS) will be introduced which uses the divide-and-conquer technique to separate a large GA/FS optimisation problem into a series of more manageable problems. The fused FS is applied to the cart pole system and the hierarchical FS is employed to the truck-and-trailer system.-
dcterms.accessRightsopen access-
dcterms.educationLevelPh.D.-
dcterms.extent[15, 170, 31 leaves] : ill. ; 30 cm.-
dcterms.issued2000-
dcterms.LCSHIntelligent control systems.-
dcterms.LCSHFuzzy systems.-
dcterms.LCSHFuzzy logic.-
dcterms.LCSHGenetic algorithms.-
dcterms.LCSHHong Kong Polytechnic University -- Dissertations.-
Appears in Collections:Thesis
Show simple item record

Page views

42
Last Week
0
Last month
Citations as of Mar 24, 2024

Google ScholarTM

Check


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