基于梯形隶属函数的模糊控制器的逼近精度分析

Analysis on the Approximation Accuracy of Fuzzy Controllers with Trapezoid Membership Functions

讨论了一类基于梯形隶属函数、单值模糊器、乘积型推理和中心平均解模糊的模糊控制器结构。利用泰勒公式,给出了一个在预定逼近精度的条件下如何划分模糊论域的公式,并且用仿真例子验证了该公式的正确性。仿真结果表明,该文所提出的计算公式是一个充分条件。

A structure is discussed for a type of fuzzy controllers, which are based on trapezoid membership functions singleton fuzzifier, product inference and center average defuzzifier. And a formula is given to partition fuzzy universe of discourse on the condition of given approximation accuracy. Finally the correctness of this formula is verified by a simulation example and the result of which shows the formula is a sufficient condition.