Fuzzy logic approximation to unknown dynamic systems via input-output measurements

An online fuzzy logic approximator for unknown, single-input-single-output, nonlinear dynamic systems is proposed in this paper. The proposed approximator consists of finite numbers of fuzzy variables which are continuous and well-defined in the system input and output domain, and a small size of rule base which contains unknown parameters. Using the measurable input-output data pairs, all the unknown parameters in rule base will be tuned to the "best" values, and thus, the fuzzy model is guaranteed to have an identical transfer function with which the original unknown nonlinear dynamic system has.