Uncertainty reasoning based on cloud models in controllers

The methodology of fuzzy reasoning has been shown to be very useful technology for modeling complex nonlinear systems. However, the most commonly used method for reasoning with fuzzy systems models, the Mamdani-Zadeh paradigm, faces many criticisms, particularly from the probability community. A new mathematical representation of linguistic concepts is presented in this paper. With the new model of normal compatibility clouds and a virtual rule engine, a novel uncertainty reasoning technology is proposed. It not only serves as a foundation of linguistic control, but also integrating fuzziness and randomness in an inseparable way. A case study is given to clean up many doubts raised in the debate between fuzzy theory and probability theory researchers, and to give a good interpretation of the Mamdani-Zadeh operations for the defuzzification strategy as well. The architecture of such a controller shows the advantages in hardware implementations.