A contribution for elaboration of self-organizing fuzzy controllers

The fuzzy controller considered in this work is a rule-based type, where the linguistic rules correspond to the actions a human operator would take when controlling a given process. These rules are expressed through the concept of fuzzy sets. A study of a self-organizing controller (SOC) is presented in which rules can be created and modified according to a performance measurement. The SOC structure is analyzed and a model reference adaptive self-organizing controller (MRAC) is developed. The rules setting in SOC-MRAC is made indirectly by improvement of a look-up table that contains the control actions. Table positions are established in a way that response to the controlled system follows the desired response given by the reference model. The system is simulated and applied to DC motor speed control in order to verify the actual capability of the controller. Results for distinct values of set point are presented