Continuous-time analog defuzzifier for product-sum based implementations

An analog defuzzifier using arithmetic product and sum is proposed. Instead of the separate computation required in MIN/MAX based implementations, conclusion and defuzzification can be simultaneously realized by a simple aggregation of analog multipliers in which the relative positions of the output membership functions can also be continuously varied, thus allowing its use in adaptive fuzzy and neuro-fuzzy controllers. Electrical simulation results of a CMOS version of the defuzzifier are provided to illustrate how the centre-of-area is efficiently approximated with the proposed implementation.