An advanced inference strategy for fuzzy control based on a new fuzzy implication function

In fuzzy controllers of Mamdani´s type the processing is performed in the steps of fuzzification, inference and defuzzification. In this paper, an advanced inference and defuzzification strategy is proposed which yields a higher level of transparency as compared with the commonly used methods. First, some aspects of the chronological development of fuzzy inference are discussed. This illustrates the reason for using the minimum as fuzzy implication function in most applications, which is not generalization of the binary implication function. In the case of the first fundamental inference strategy, the step of inference is performed by an AND-operation and the combination of the different rules is performed by an OR-operation. In the case of the second fundamental inference strategy, the step of inference is performed by applying a generalization of the binary implication function and the combination of the different rules is performed by an AND-operation. In this paper, an advanced new fuzzy implication function is introduced which is a generalization of the binary implication function. This new fuzzy implication function can be used in the second fundamental inference strategy