摘要
模糊化是将物理空间的实测精确量影射为模糊推理空间上的模糊集合。在模糊控制中 ,不同的模糊推理方法要求不同的模糊化方法 ,不同的模糊化方法对模糊控制性能影响很大。本文首先系统地总结了现有的模糊化方法 ,然后提出了模糊向量真值修正模糊化方法 ,最后 ,针对常用的 CRI法 ,完成了不同模糊化方法的一阶惯性时滞定常系统的模糊控制仿真 ,结果表明 ,该方法能够提高 CRI法的模糊控制性能 ,消除稳态误差。
Fuzzifier is used to convert a real-time sampled crisp data in a physical domain to fuzzy (subset) in fuzzy reasoning domain. In fuzzy control, various methods of the fuzzy logic inference (demand) various methods of the fuzzifying, and various methods of the fuzzifying will influence the (performance) of the fuzzy control greatly. The existing fuzzifiers have been summarized systematically (firstly) in this paper. Then a fuzzifier with fuzzy vector revised by true values has been put forward. (Fuzzy) control simulation has been finished at last for a controlled system with an order, inertia, delay time and constant coefficients by the compositional rule of inference, which is often used in fuzzy (control.) The simulation results show that the fuzzifier with fuzzy vector revised by true values is (suitable) to the method of the fuzzy inference which premise is fuzzy vector and capable of improving the performance and eliminating the stable error of the fuzzy control system. The fuzzifier is not only easy to fuzzify the linear membership functions, but also convenient to fuzzify Gauss and bell (membership) functions.
出处
《模糊系统与数学》
CSCD
2004年第3期62-67,共6页
Fuzzy Systems and Mathematics
基金
国家自然科学青年基金资助项目 (5 990 80 0 1)
哈尔滨工业大学跨学科交叉性研究基金资助项目 (HIT.MD2 0 0 0 30 )
关键词
模糊化
模糊向量
真值修正
模糊关系合成推理
稳态误差
Fuzzifying
Fuzzy Vector
Revised by True Value
Compositional Rule of Inference
(Stable Error)