In the design of fuzzy controllers there is a need for standardizing the selection of rule table and the scaling factors. For example, in the design of self-organizing fuzzy controllers or the rule self-regulating fuzzy controllers, the rule table or the performance table is often derived either by trial and error or from the heuristics of an expert. Recently a fuzzy control based on natural control laws is suggested and its relationship to the linear state feedback control is derived. In this work, the natural control laws and the regular fuzzy set are employed to modify the self-organizing fuzzy controller and the rule self-regulating fuzzy controller. A systematic procedure for designing these types of adaptive fuzzy controllers is developed. For illustration a natural law fuzzy controller and an LQR are also implemented for the tracking control of a DC servomotor. The results indicate that the modified rule self-regulating fuzzy controller yields the best performance among the four control designs. From the design procedure adaptive fuzzy controllers can be derived with less number of parameters.
All Science Journal Classification (ASJC) codes
- Artificial Intelligence