Data Driven Dynamic System Modeling Based on Fuzzy Logic
Abstract
Fuzzy logic is a general computation system based on the fuzzy set theory, which is inspired by human
thinking. This system is based on the relations between logical expressions and linguistic variables.One of the most
important advantages in terms of engineering is that it does not need a mathematical model of the system of interest.
The main problem is to determine the most suitable values of its parameters so as to perform the task expected from
it.In this study, it has investigated that the parameters of commonly used Takagi-Sugeno(TS) type fuzzy system are
determined based on input / output data at hand. First, the input-output variables of the fuzzy system are determined
by the method given in this study, and the variables of the input are fuzzificated by the homogeneous distributed
membership functions in the related input space. Thus, the premise/antecedent parameters of the fuzzy system are
determined. Then, the consequent/rule parameters of the fuzzy system are determined based on the input-output
sample data with the least square estimation (LSE) method. In this study, this method is discussed on the modeling of
five different dynamical systems with fuzzy systems of TS type. The results show that the method can be used
effectively if the designer has input-output samples.
Keywords
References
- R. BabuŠka, Fuzzy Modeling for Control, Kluwer Academic Publishers, 1998.
- L. Zadeh, «Outline of a new approach to the analysis of complex systems and decision processes,» pp. 28-44, 1973.
- T. Takagi ve M. Sugeno, «Fuzzy identification of systems and its application to modeling and control,» IEEE Trans. on Systems, Man, and Cybernetics, Cilt %1 / %2SMC-15, no. 1, pp. 116-132, 1985.
- R. Babuska, «Fuzzy Systems, Modeling and Identification,» [Çevrimiçi]. Available: https://www.researchgate.net/profile/Robert_Babuska/publication/228769192_Fuzzy_Systems_Modeling_and_Identification/links/02e7e5223310e79d19000000.pdf?origin=publication_list&ev=srch_pub_xdl. [Erişildi: 22 12 2016].
- K. S. Narendra and K. Parthasarathy, "Identification and control of dynamical systems using neural networks," IEEE Transactions on Neural Networks, vol. 1, no. 1, pp. 4-27, 1990.
- Y. Oussar, I. Rivals ve L. Dreyfus, «Training wavelet networks for nonlinear dynamic input output modeling,» Neurocomputing, cilt 20, pp. 173-188, 1998.
- P. S. Sastry, G. Santharam ve K. P. Unnikrishnan, «Memory neuron networks for identification and control of dynamical systems,» IEEE Transaction on Neural Networks, cilt 5, no. 2, pp. 306 - 319, 1994.
Details
Primary Language
Turkish
Subjects
-
Journal Section
Research Article
Publication Date
June 28, 2017
Submission Date
November 20, 2017
Acceptance Date
-
Published in Issue
Year 2017 Volume: 4 Number: 1