Hierarchical Fuzzy Control with Applications to Seesaw Systems

• Main Authors: Ming-Lin Lee, 李明霖
• Other Authors: Hung-Yuan Chung
• Format: Others
• Language: en_US
• Published: 2000
• Online Access: http://ndltd.ncl.edu.tw/handle/38433356674054036292

碩士 === 國立中央大學 === 電機工程研究所 === 88 === The design of fuzzy controllers is commonly a time-consuming activity involving knowledge acquisition, definition of the controller structure, definition of rules, and other controller parameters. At present, one of the important issues in fuzzy logic systems is how to reduce the number of involved rules and their corresponding computation requirements. The idea of hierarchical fuzzy systems (HFSs) has been reported. But, the involved fuzzy rules in the middle of the hierarchical structure have little physical meaning and consequently are hard to design. This phenomenon becomes prominent as the number of layers grows larger in an HFS. To overcome the problem that intermediate outputs have nothing to do with the physical variables, this thesis propose a new kind of mapping rule base scheme to get the rule base of HFS without the physical meaning. As a result, all of the rule bases of fuzzy logic units (FLUs) don’t design again and we can reduce the number of involved rules. In many layers, the mapping rule is useful, equivalently. The several simulations on computer are given to confirm the correctness and to illustrate design procedures. Moreover, Experiments on a practical system, such as an inverted wedge system, assisted with genetic algorithm, verify the effectiveness of the proposed methods. Judging from simulative and experimental results, the methods described in this provide efficient approaches to design HFSs.