论文部分内容阅读
模糊神经网络(f-NN)为从控制操作的传感器数值数据中获取“if-then”控制规则提供新的工具,这种方法把模糊控制规则的抽取过程转化成f-NN网络结构及参数的生成与学习问题。控制规则是模糊控制系统设计的基础,控制规则的可靠程度是一个不容忽视的问题。本文提出一种对所获取的控制规则进行置信度估计的分析方法,并给出便于工程上使用的近似公式。
The fuzzy neural network (f-NN) provides a new tool for obtaining “if-then” control rules from control-operated sensor numerical data. This method transforms the extraction of fuzzy control rules into f-NN network structures and Parameter Generation and Learning Problems. Control rules are the basis of fuzzy control system design, the reliability of control rules is a problem that can not be ignored. In this paper, a method of estimating the confidence of the obtained control rules is proposed and an approximate formula for the convenience of engineering is given.