| The dissertation research shows the possibility of applying adaptive scheme to fuzzy control systems, especially to fuzzy controllers for nonlinear systems with unknown nonlinearities. Three approaches to adaptive fuzzy control are presented. The research is focused on the representation capability of fuzzy modelling in nonlinear functions approximation and design of Adaptive Fuzzy Controller (AFC) via feedback linearization and nonlinear PID approaches. Our results show that these approaches to fuzzy control are competitive to other fuzzy logic controllers. The detailed design will be presented, accompanied with simulation results of a number of typical examples. In this paper special attention is given to the relationship between nonlinearities of the system and the assignment of essential elements in fuzzy control systems, such as the number of reference fuzzy subsets, the shape of membership functions, and overlapping conditions between fuzzy subsets. The introduction of adaptive scheme to fuzzy logic control makes the controller more robust in the sense of compensating the unknown uncertainties. Two membership function adaptation algorithms are presented to relax the difficulties of rules extraction and acquisition. Algorithms of combining and separating membership functions are also presented to provide another point of view toward the relationship between membership functions in fuzzy systems. |
