On the development of a fuzzy model for nonlinear systems

The authors propose a fuzzy algorithm for modeling nonlinear physical systems. Each of the nonlinear coefficients in the system dynamic equation is modeled by a set of fuzzy rules. An identification algorithm incorporating a recursive least-squares method and an optimum search process is then used to optimize the parameters of the fuzzy rules. This ensures that all unknown fuzzy parameters can be predicted systematically. The feasibility of such an algorithm is demonstrated by two examples, a 2-link manipulator and a servovalve-controlled pneumatic chamber. Both computer simulation and experimental results show that the proposed fuzzy algorithm is very useful for modeling nonlinear systems, as the predicted system responses match the actual ones very well