An Algorithm For Online Self-Organization Of Fuzzy Controllers

This work plesents a fuzzy controller capable of designing its own strut:hue online, based on the data obtained during the normal system (mei atm', The algorithm clocs not use plevions information about the differential equations that define the plant's behaviour The controller may be initially empty. as the inetlitpd is able to distinguish the important input, variables by assigning them more membership lunetwits With this awl; the method works in two stages the adaptation ol the consequents for every tested topology anti the online addition of new membership functions To show as capabilities. simulation results with a liquid tank system ale analyzed