Fuzzy identification and control of a class of nonlinear systems

Two different methods of fuzzy identification of a class of nonlinear systems are discussed in this paper. This is applicable to systems with unknown and partially known mathematical models. The class of systems considered are nonlinear in output but linear in input. In the first method, a gray box model is considered. The nominal values of parameters of the nonlinear system are assumed to be known. The unknown nonlinear function is identified offline by choosing a suitable fuzzy relational model and the parameters of the nonlinear system are updated on-line using recursive least square (RLS) algorithm. In the second method, a block box model is considered. The nonlinear plant is identified on-line by choosing a suitable linear model using RLS in stage-1 and the residual nonlinear part is identified in stage-2 using fuzzy identification. The control input is then calculated based on the identified nonlinear model using weighted one step ahead control method. Numerical examples are given to validate the proposed methods.