Font Size: a A A

Study On Fuzzy Modeling Methods And Its Application In Chaotic Time Series Prediction

Posted on:2008-06-18Degree:MasterType:Thesis
Country:ChinaCandidate:L Y MaFull Text:PDF
GTID:2178360212995241Subject:Control theory and control engineering
Abstract/Summary:PDF Full Text Request
Complex and uncertain systems are often poorly modeled with conventional approaches that attempt to find a global function or analytical structure for a nonlinear system. A new approach is outlined by L.A.Zadeh that"provides an approximate and yet effective means of describing the behavior of systems which are too complex or too ill-defined to admit use of precise mathematical analysis."But due to the nonlinear systems are too complex and the fuzzy system is immature research domain, there exist many issues should be improved to be solved. This dissertation surrounds fuzzy modeling methods for nonlinear systems and its application in chaotic time series prediction to discuss and to research.First, it offers an overview of the developing process of the theory and methods of fuzzy modeling and its current situation. Then it gives the introduction of the basic define and principle in fuzzy identification methods, so it paves way for the later research.Consider the problems that there are very few researches on nonlinear T-S fuzzy systems'universal approximation. So in this paper, we will study the nonlinear approximation of a kind of nonlinear T-S fuzzy system based on Stone-Weirstrass theorem, and prove that when the fuzzy sets are Gaussian membership functions, it has universal approximation.In accordance with the problems that the common fuzzy clustering algorithm has the process of iterative calculate, so it is quite time-consuming. This paper proposed a new algorithm of fuzzy identification based on fuzzy clustering for nonlinear systems. It is based on the fuzzy clustering membership function and equalized universe method (EUM). This algorithm needn't the iteration process for researching cluster centers, and computation time required to partition a data set into c classes is significantly reduced, and it is suitable foron-line modeling and control. The simulation results of chaotic time series prediction demonstrate the effectiveness and the applicability of the proposed method.In accordance with the problems that ordinary clustering algorithm just attained a partial superior result when determine clustering clusters arbitrarily. The paper introduces a new method for fuzzy modeling based on a nearest neighbor clustering and FCMV clustering which overcomes the limitation. The methods consists of a sequence of steps aiming towards developing a Takagi-Sugeno (TS) fuzzy model of optimal structure and realize the modeling and forecasting of the nonlinear system. Finally, the effectiveness of proposed algorithms are demonstrated by performance data of Mackey-Glass chaotic time sequence.
Keywords/Search Tags:Fuzzy identification, Fuzzy clustering, Universal approximation, Equalized universe method, Chaotic time series
PDF Full Text Request
Related items