Similarity based control strategy reconfiguration by fuzzy reasoning and fuzzy automata

In case of control partially known complex systems, a kind of control strategy could be built simply by combining some known, partially valid control strategies. These known strategies are covering only a part of the whole state space of the possible control strategies. In some situation, these strategies could be combined in a hierarchical fuzzy way to form a strategy covering at least the same state space as all the attendant strategies. There are two main questions of this kind of strategy reconfiguration. The first is the decision about the level of necessity and the type of the strategy (rulebase in the case of fuzzy control) needed to use by the actual system. The second is the way of generating the actual strategy, as a combination of the different existing known partially valid strategies. One solution for these questions could be the combination of the fuzzy automata based system state approximation and the hierarchical interpolative fuzzy reasoning-the similarity based strategy reconfiguration. As an example for the practical application of the proposed structure, path tracking a collision avoidance strategy of a simulated AGV is introduced in this paper