A fuzzy automaton for control applications

A fuzzy automaton which handles real-valued data has been of interest to control society as a mean to tackle identification problem of dynamical systems. A firm basis to apply fuzzy automata (FA) with fuzzy alphabet is investigated, in a systematic way, to control and modeling applications within real-valued environments. In this way, a novel definition of transducer FA and its operating environment will be devised. The new definition yields a close connection between FA and rule-based fuzzy inference systems which provides an induction algorithm for FA. Despite of similarity in design phase, FA and rule-based inference engines have totally different behavior due to the effect of state dynamics. The latter concept will be discussed through a case study on modeling a time-dynamic process.