Font Size: a A A

Study On Fuzzy Control Of Flexible Manipulator

Posted on:2007-02-04Degree:MasterType:Thesis
Country:ChinaCandidate:X Q CengFull Text:PDF
GTID:2178360182498106Subject:Power electronics and electric drive
Abstract/Summary:PDF Full Text Request
The interpretability of the fuzzy model is the important feature that distinguishes fuzzy systems from many other artificial intelligent techniques. Research on the interpretability has not got enough attention. This paper proposes a method of interpretable modeling based on fuzzy clustering in order to improve the interpretability of fuzzy model. An entropy-based fuzzy partition validity index is proposed and used to improve the interpretability of fuzzy models. The fuzzy clustering technique associated with the proposed validity index is used to find out the optimal number of fuzzy rules and the membership functions in antecedent part, and the parameters in consequent part are found out by means of least square algorithm, then gradient descent algorithm is used to precisely adjust parameters of fuzzy model. Finally, the method is successfully applied to identify the well-known Box-Jenkins data set, it is verified that the method has higher accuracy and better linguistical interpretability.To track expectation locus accurately as the load and link length of two-link flexible manipulator varies, the model of flexible manipulation was decomposed into two subsystems, slowly-changing subsystem and fast-changing subsystem based on the singular perturbation model. A hybrid control method, consisting of adaptive sliding mode control which was applied to slowly-changing subsystem and fuzzy control which was applied to fast-changing subsystem, was presented. The adaptive algorithm can estimate unknown parameters on-line, and fuzzy control can damp the vibration caused by sliding mode control. Simulation result shows that the proposed method not only can accurately and fast track the flexible manipulator locus, but also can effectively damp the elastic vibration, and is robust to the variation of system parameters. The characteristics of the proposed method is that each equivalent vibration parameters can be adaptively estimated and compensated respectively, so it is simple and easy to implement.
Keywords/Search Tags:Fuzzy modeling, Interpretability, Fuzzy clustering, Entropy, Two-link, Flexible manipulator, Singular perturbation, Sliding mode control
PDF Full Text Request
Related items